Subset (2)

section · 216k rows

Split (3)

train · 203k rows

article stringlengths 0 745k	abstract stringlengths 0 92.7k
we study a problem of estimation of a symmetric kernel @xmath13 defined on a large weighted graph with a vertex set @xmath1 and @xmath14 based on a finite number of noisy linear measurements of @xmath10 . for simplicity , assume that these are the measurements of randomly picked entries of @xmath15 matrix @xmath16 , which is a standard sampling model in matrix completion . more precisely , let @xmath17 be @xmath5 independent copies of a random triple @xmath18 , where @xmath19 are independent random vertices sampled from the uniform distribution @xmath20 in @xmath1 , and @xmath21 is a `` measurement '' of the kernel @xmath10 at a random location @xmath22 in the sense that @xmath23 . in what follows , we assume that , for some constant @xmath24 , @xmath25 a.s . , which implies that @xmath26 . the target kernel @xmath10 is to be estimated based on its i.i.d . measurements @xmath17 . we would like to study this problem in the case when the target kernel @xmath10 is , on the one hand , `` low rank '' [ i.e. , @xmath27 is relatively small compared to @xmath28 , and on the other hand , it is `` smooth '' in the sense that its `` sobolev - type norm '' is not too large . discrete versions of sobolev norms can be defined for functions and kernels on weighted graphs in terms of their graph laplacians . the problem of estimation of smooth low - rank kernels is of importance in a number of applications , such as learning kernels representing and predicting similarities between objects , various classification problems in large complex networks ( e.g. , edge sign prediction ) as well as matrix completion problems in the design of recommender systems ( collaborative filtering ) . our main motivation , however , is mostly theoretical : we would like to explore to which extent taking into account smoothness of the target kernel could improve the existing methods of low rank recovery . we introduce some notation used throughout the paper . let @xmath29 be the linear space of _ symmetric kernels _ @xmath30 , @xmath31 ( or , equivalently , symmetric @xmath15 matrices with real entries ) . given @xmath32 , we use the notation @xmath33 for the rank of @xmath34 and @xmath35 for its trace . for two functions @xmath36 , @xmath37 . suppose that @xmath38 is the spectral representation of @xmath34 with @xmath39 , @xmath40 being nonzero eigenvalues of @xmath34 repeated with their multiplicities and @xmath41 being the corresponding orthonormal eigenfunctions ( obviously , there are multiple choices of @xmath42s in the case of repeated eigenvalues ) . we will define @xmath43 as @xmath44 and the support of @xmath34 as @xmath45 . for @xmath46 , define the schatten @xmath47-norm of @xmath34 as @xmath48 where @xmath49 . for @xmath50 , @xmath51 is also called the nuclear norm and , for @xmath52 , @xmath53 is called the hilbert schmidt or frobenius norm . this norm is induced by the hilbert schmidt inner product which will be denoted by @xmath54 . the operator norm of @xmath34 is defined as @xmath55 . by @xmath54 and the corresponding euclidean norm by @xmath56 . ] let @xmath57 be the distribution of random couple @xmath22 . the @xmath58-norm of kernel @xmath34 , @xmath59 is naturally related to the sampling model studied in the paper , and it will be used to measure the estimation error . denote by @xmath60 the corresponding inner product . since @xmath20 is the uniform distribution in @xmath1 , @xmath61 and @xmath62 . in what follows , it will be often more convenient to use these rescaled versions rather than the actual hilbert schmidt norm or inner product . we will denote by @xmath63 the canonical orthonormal basis of the space @xmath64 . based on this basis , one can construct matrices @xmath65 . if @xmath66 is an arbitrary ordering of the vertices in @xmath1 , then @xmath67 is an orthonormal basis of the space @xmath68 of symmetric matrices with hilbert schmidt inner product . in standard matrix completion problems , @xmath1 is a finite set with no further structure ( i.e. , the set of edges of the graph or the weight matrix are not specified ) . in the noiseless matrix completion problems , the target matrix @xmath10 is to be recovered from the measurements @xmath17 , where @xmath69 . the following method is based on nuclear norm minimization over the space of all matrices that `` agree '' with the data @xmath70 it has been studied in detail in the recent literature ; see @xcite and references therein . clearly , there are low rank matrices @xmath10 that can not be recovered based on a random sample of @xmath5 entries unless @xmath5 is comparable with the total number of the entries of the matrix . for instance , for given @xmath71 , let @xmath72 . then , @xmath73 . however , the probability that the only two nonzero entries of @xmath10 are not present in the sample is @xmath74 , and it is close to @xmath75 when @xmath76 . in this case , the matrix @xmath10 can not be recovered . so - called _ low coherence _ assumptions have been developed to define classes of `` generic '' matrices that are not `` low rank '' and `` sparse '' at the same time and for which noiseless low rank recovery is possible with a relatively small number of measurements . for a linear subspace @xmath77 , let @xmath78 be the orthogonal complement of @xmath79 and let @xmath80 be the orthogonal projector onto the subspace @xmath79 . denote @xmath81 , @xmath82 . a _ coherence coefficient _ is a constant @xmath83 such that @xmath84 \\[-8pt ] \nonumber \bigl \|\bigl\langle\operatorname { sign}(s_{\ast})e_u , e_v\bigr \rangle\bigr\|^2 & \leq&\frac{\nu r}{m^2 } , \qquad u , v\in v\end{aligned}\ ] ] ( it is easy to see that @xmath85 can not be smaller than @xmath75 ) . the following highly nontrivial result is essentially due to candes and tao @xcite ( a version stated here is due to gross @xcite and it is an improvement of the initial result of candes and tao ) . it shows that target matrices of `` low coherence '' ( for which @xmath85 is a relatively small constant ) can be recovered exactly using the nuclear norm minimization algorithm ( [ nucemin ] ) provided that the number of observed entries is of the order @xmath86 ( up to a log factor ) . [ gross ] suppose conditions ( [ cohera ] ) hold for some @xmath83 . then , there exists a numerical constant @xmath87 such that , for all @xmath88 , @xmath89 with probability at least @xmath90 . in the case of noisy matrix completion , a matrix version of lasso is based on a trade - off between fitting the target matrix to the data using least squares and minimizing the nuclear norm @xmath91.\ ] ] this method and its modifications have been studied by a number of authors ; see @xcite . the following low - rank oracle inequality was proved in @xcite ( theorem 4 ) for a `` linearized version '' of the matrix lasso estimator @xmath92 . assume that , for some constant @xmath24 , @xmath25 a.s . let @xmath93 and suppose that @xmath94 then , there exists a constant @xmath87 such that with probability at least @xmath95 @xmath96.\ ] ] in particular , @xmath97 very recently , the last bound was proved in @xcite for the matrix lasso estimator ( [ lsnuce ] ) itself in the case when the domain of optimization problem is @xmath98 , where @xmath99 ; in fact , both @xcite and @xcite dealt with the case of rectangular matrices . in the current paper , we are more interested in the case when the target kernel @xmath10 is defined on the set @xmath1 of vertices of a weighted graph @xmath100 with a symmetric matrix @xmath101 of nonnegative weights . this allows one to define the notion of graph laplacian and to introduce discrete sobolev norms characterizing smoothness of functions on @xmath1 as well as symmetric kernels on @xmath102 . denote @xmath103 it is common in graph theory to call @xmath104 the degree of vertex @xmath105 . let @xmath106 be the diagonal @xmath15 matrix ( kernel ) with the degrees of vertices on the diagonal ( it is assumed that the vertices of the graph have been ordered in an arbitrary , but fixed way ) . the laplacian of the weighted graph @xmath107 is defined as @xmath108 . denote @xmath54 the canonical euclidean inner product in the @xmath109-dimensional space @xmath64 of functions @xmath110 and let @xmath56 be the corresponding norm . it is easy to see that @xmath111 implying that @xmath112 is a symmetric nonnegatively definite linear transformation . in a special case of a usual graph @xmath113 with vertex set @xmath1 and edge set @xmath114 , one defines @xmath115 if and only if @xmath116 ( i.e. , vertices @xmath105 and @xmath117 are connected with an edge ) and @xmath118 otherwise . in this case , @xmath104 is the number of edges incident to the vertex @xmath105 and @xmath119 the notion of graph laplacian allows one to define discrete sobolev norms @xmath120 for functions on the vertex set of the graph and thus to describe their smoothness on the graph . given a symmetric kernel @xmath121 , one can also describe its smoothness in terms of the norms @xmath122 . suppose @xmath34 has the following spectral representation : @xmath123 where @xmath124 are the eigenvalues of @xmath34 ( repeated with their multiplicities ) and @xmath125 are the corresponding orthonormal eigenfunctions in @xmath64 , then @xmath126 basically , it means that the smoothness of the kernel @xmath34 depends on the smoothness of its eigenfunctions . in what follows , we will often use rescaled versions of sobolev norms , @xmath127 it will be convenient for our purposes to fix @xmath128 and to define a nonnegatively definite symmetric kernel @xmath129 . we will characterize the smoothness of a kernel @xmath130 by the squared sobolev - type norm @xmath131 . the kernel @xmath132 will be fixed throughout the paper , and its spectral properties are crucial in our analysis . to the graph and its laplacian will be of little importance allowing , possibly , other interpretations of the problem . ] assume that @xmath132 has the following spectral representation @xmath133 where @xmath134 are the eigenvalues repeated with their multiplicities , and @xmath135 are the corresponding orthonormal eigenfunctions ( of course , there is a multiple choice of @xmath136 in the case of repeated eigenvalues ) . let @xmath137 . we will assume in what follows that , for some constant @xmath138 , @xmath139 for all @xmath140 . it will be also convenient to set @xmath141 . let @xmath142 and @xmath143 . it is easy to show ( see the proof of theorem [ lsupth ] below ) that kernel @xmath10 can be approximated by the following kernel : @xmath144 with the approximation error @xmath145 note that the kernel @xmath146 can be viewed as an @xmath147 matrix ( represented in the basis of eigenfunctions @xmath148 ) and @xmath149 , so , one needs @xmath150 parameters to characterize such matrices . thus , one can expect , that such a kernel can be estimated , based on @xmath5 linear measurements , with the squared @xmath58-error of the order @xmath151 . taking into account the bound on the approximation error ( [ apprerr ] ) and optimizing with respect to @xmath152 , it would be also natural to expect the following error rate in the problem of estimation of the target kernel @xmath153 @xmath154.\ ] ] we will show that such a rate is attained ( up to constants and log factors ) for a version of least squares method with a nonconvex complexity penalty ; see section [ sectlsnonconvex ] . this method is not computationally tractable , so , we also study another method , based on convex penalization with a combination of nuclear norm and squared sobolev type norm , and show that the rates are attained for such a method , too , provided that the target matrix satisfies a version low coherence assumption with respect to the basis of eigenfunctions of @xmath132 . more precisely , we will prove error bounds involving so called _ coherence function _ @xmath155 , that characterizes the relationship between the kernel @xmath132 defining the smoothness and the target kernel @xmath10 ; see section [ sectlsconvex ] for more details ; see also @xcite for similar results in the case of `` linearized least squares '' estimator with double penalization . finally , we prove minimax lower bounds on the error rate that are roughly of the order @xmath156 $ ] ( subject to some extra conditions and with additional terms ; see section [ sectlowerbound ] ) . in typical situations , this expression is , up to a constant , of the same order as the upper bound ( [ upperbd ] ) . for instance , if @xmath157 for some @xmath158 , then the minimax error rate of estimation of the target kernel @xmath10 is of the order @xmath159 ( up to log factors ) . when @xmath109 is sufficiently large , the term @xmath160 will be dropped from the minimum , and we end up with a nonparametric convergence rate controlled by the smoothness parameter @xmath161 and the rank @xmath162 of the target matrix @xmath10 ( the dependence on @xmath109 in the first two terms of the minimum is only in the log factors ) . the focus of the paper is on the matrix completion problems with uniform random design , but it is very straightforward to extend the results of the following sections to sampling models with more general design distributions discussed in the literature on low rank recovery ( such as , e.g. , the models of random linear measurements studied in @xcite ) . it is also not hard to replace the range @xmath163 of the response variable @xmath164 by the standard deviation of the noise in the upper and lower bounds obtained below . this is often done in the literature on low - rank recovery , and it can be easily extended to the framework discussed in the paper by modifying our proofs . we have not discussed this in the paper due to the lack of space . in this section , we derive minimax lower bounds on the @xmath165-error of an arbitrary estimator @xmath92 of the target kernel @xmath10 under the assumptions that the response variable @xmath164 is bounded by a constant @xmath24 , the rank of @xmath10 is bounded by @xmath166 and its sobolev norm @xmath167 is bounded by @xmath168 . more precisely , given @xmath169 and @xmath170 , denote by @xmath171 the set of all symmetric kernels @xmath30 such that @xmath172 and . given @xmath173 and @xmath24 , let @xmath174 be the set of all probability distributions of @xmath18 such that @xmath22 is uniformly distributed in @xmath102 , @xmath25 a.s . and @xmath23 , where @xmath175 . for @xmath176 , denote @xmath177 . recall that @xmath178 are the eigenfunctions of @xmath132 orthonormal in the space @xmath179 . then @xmath180 are orthonormal in @xmath181 . we will obtain minimax lower bounds for classes of distributions @xmath182 in two different cases . define @xmath183 in the first case , we assume that , for some ( relatively large ) value of @xmath184 , the quantity @xmath185 is not too large . roughly , it means that most of the components of vectors @xmath186 are uniformly small , say , @xmath187 in other words , the @xmath15 matrix @xmath188 is `` dense , '' so we refer to this case as a _ `` dense case . '' _ the opposite case is when this matrix is _ `` sparse . '' _ its `` sparsity '' will be characterized by the quantity @xmath189 which , in this case , should be relatively small . a typical example is the case when basis of eigenfunctions @xmath190 coincides with the canonical basis @xmath63 of @xmath64 ( then , @xmath191 ) . denote @xmath192 . in the _ dense case _ , the following theorem holds . [ lowerbound ] define @xmath193.\ ] ] there exist constants @xmath194 such that @xmath195 where the infimum is taken over all the estimators @xmath196 based on @xmath5 i.i.d . copies of @xmath18 . in fact , it will follow from the proof that , if @xmath197 ( i.e. , the smallest nonzero eigenvalue of @xmath132 is not too large ) , then the maximum in the definition of @xmath198 can be extended to all @xmath152 . [ lowerboundcorr ] let @xmath199.\ ] ] there exist constants @xmath194 such that @xmath200 take @xmath201 in the statement of theorem [ lowerbound ] and observe that @xmath202 and @xmath203 . it is easy to check that @xmath204 . it is obvious that one can replace the quantity @xmath205 in theorem [ lowerbound ] [ or the quantity @xmath206 in corollary [ lowerboundcorr ] ] by the following smaller quantity : @xmath207,\ ] ] where @xmath208\wedge m. $ ] moreover , denote @xmath209 it is straightforward to check that @xmath210 = \frac{a^2(r\wedge\bar l)\bar l}{n } \vee\frac{\rho ^2}{\lambda_{\bar l+1}}\ ] ] and , if @xmath211 , then @xmath212 suppose that , for some @xmath158 , @xmath213 ( in particular , it means that @xmath214 and @xmath215 ) . then , an easy computation shows that @xmath216 let @xmath201 and take @xmath217\wedge m. $ ] the condition @xmath211 is satisfied , for instance , when either @xmath218 , or @xmath219 , where @xmath220 is a small enough constant ( this , essentially , means that @xmath5 is sufficiently large ) . under either of these conditions , we get the following expression for a minimax lower bound : @xmath221 we now turn to the _ sparse case_. [ lowerboundsparse ] let @xmath222.\ ] ] there exist constants @xmath194 such that @xmath223 it will be clear from the upper bounds of section [ sectlsnonconvex ] ( see the remark after theorem [ lsupth ] ) that , at least in a special case when @xmath148 coincides with the canonical basis of @xmath64 , the additional term @xmath224 is correct ( up to a log factor ) . at the same time , most likely , the `` third terms '' of the bounds of theorem [ lowerbound ] ( in the dense case ) and theorem [ lowerboundsparse ] ( in the sparse case ) have not reached their final form yet . a more sophisticated construction of `` well separated '' subsets of @xmath174 might be needed to achieve this goal . the main difficulty in the proof given below is related to the fact that we have to impose constraints , on the one hand , on the entries of the target matrix represented in the canonical basis and , on the other hand , on the soblolev type norm @xmath225 ( for which it is convenient to use the representation in the basis of eigenfunctions of @xmath132 ) . due to this fact , we are using the last representation in our construction , and we have to use an argument based on the properties of rademacher sums to ensure that the entries of the matrix represented in the canonical basis are uniformly bounded by @xmath163 . this is the reason why the `` third terms '' occur in the bounds of theorems [ lowerbound ] and [ lowerboundsparse ] . in this case , when the constraints are only on the norm @xmath226 and on the variance of the noise and there are no constraints on @xmath227 , it is much easier to prove the lower bound of the order @xmath228 $ ] without any additional terms . note , however , that the condition @xmath229 is of importance in the following sections to obtain the upper bounds for penalized least squares estimators that match the lower bounds up to log factors . proof of theorem [ lowerbound ] the proof relies on several well - known facts stated below . in what follows , @xmath230 denotes kullback leibler divergence between two probability measures @xmath231 defined on the same space and such that @xmath232 ( i.e. , @xmath85 is absolutely continuous with respect to @xmath233 ) . we will denote by @xmath234 the @xmath5-fold product measure @xmath235 . the following proposition is a version of theorem 2.5 in @xcite . [ kbound ] let @xmath236 be a finite set of distributions of @xmath18 such that the following assumptions hold : there exists @xmath237 such that for all @xmath238 , @xmath239 ; there exists @xmath240 such that @xmath241 for all @xmath242 , @xmath243 . then , there exists a constant @xmath244 such that @xmath245 we will also use varshamov gilbert bound ( see @xcite , lemma 2.9 , page 104 ) , sauer s lemma ( see @xcite , page 39 ) and the following elementary bound for rademacher sums ( @xcite , page 21 ) : for all @xmath184 , @xmath246 where @xmath247 are i.i.d . rademacher random variables ( i.e. , @xmath248 with probability @xmath249 and @xmath250 with the same probability ) . we will start the proof with constructing a `` well separated '' subset @xmath236 of the class of distributions @xmath251 that will allow us to use proposition [ kbound ] . fix @xmath252 , @xmath253 and @xmath254 . denote @xmath255 , l''=l - l'$ ] . first assume that @xmath256 . denote @xmath257 , where @xmath258 or @xmath259 . let @xmath260 ( so , @xmath261 is the class of all @xmath262 matrices with entries @xmath263 or @xmath264 ) . given @xmath265 , let @xmath266 be the @xmath267 matrix that consists of @xmath268 $ ] blocks @xmath269 and the last block @xmath270 , where @xmath271r$ ] and @xmath272 is the @xmath273 zero matrix . finally , define the following symmetric @xmath15 matrix : @xmath274 now , given @xmath275 , define a symmetric kernel @xmath276 , @xmath277 it is easy to see that @xmath278 \\[-8pt ] \nonumber k_{\sigma}^{\prime } ( u , v)&= & \kappa\sum _ { i=1}^{l'}\sum_{j=1}^r \sigma_{ij}\phi_i(u)\sum_{k=0}^{[l''/r]-1 } \phi_{l'+rk+j}(v).\end{aligned}\ ] ] let @xmath279 we will show that , if @xmath280 is sufficiently small ( its precise value to be specified later ) , then the set @xmath281 contains at least three quarters of the points of the combinatorial cube @xmath282 . to this end , define @xmath283 where @xmath284 is a random vector with i.i.d . rademacher components . assume , in addition , that @xmath285 and @xmath22 are independent . it is enough to show that @xmath286 with probability at least @xmath287 . we have @xmath288 we will use bound ( [ rademach ] ) to control @xmath289 [ recall that @xmath290 is a rademacher sum ] . denote @xmath291 - 1}\phi_{l'+rk+j}(v ) \biggr)^2 . \ ] ] observe that @xmath292 where @xmath293 and we used the bound @xmath294 - 1 } \phi_{l'+rk+j}(v ) \biggr)^2\leq \frac{l''}{r } \sum _ { k=0}^{[l''/r]-1}\phi_{l'+rk+j}^2(v).\ ] ] thus , applying ( [ rademach ] ) to the rademacher sum @xmath295 , we get @xmath296 given @xmath297 $ ] , denote @xmath298 for @xmath152 . this yields @xmath299 substituting the last bound into ( [ ogogo ] ) , we get @xmath300 now , to get @xmath301 , it is enough to take @xmath302 next observe that @xmath303}\pmatrix{l'r \cr k}.\ ] ] it follows from sauer s lemma that there exists a subset @xmath304 with @xmath305 + 1 $ ] and such that @xmath306 , where @xmath307 , @xmath308 since @xmath253 , we have @xmath309 and @xmath310 . we can now apply varshamov gilbert bound to the combinatorial cube @xmath311 to prove that there exists a subset @xmath312 such that @xmath313 and , for all @xmath314 , @xmath315 it is now possible to choose a subset @xmath316 of @xmath281 such that @xmath317 and @xmath318 . then , we have @xmath319 and @xmath320 for all @xmath321 . we are now in a position to define the set of distributions @xmath236 . for @xmath322 , denote by @xmath323 the distribution of @xmath18 such that @xmath22 is uniform in @xmath102 and the conditional distribution of @xmath164 given @xmath22 is defined as follows : @xmath324 since @xmath325 for all @xmath322 , we have @xmath326,\sigma\in\lambda$ ] . denote @xmath327 . for @xmath328 , we have @xmath329 note that @xmath330 ; see the definitions of @xmath331 and @xmath332 . moreover , we have @xmath333 and @xmath334 - 1 } \phi_i\otimes\phi_{l'+rk+j}\\ & & \hspace{2pt } { } + \kappa\sum _ { i=1}^{r}\sum_{j=1}^{l ' } \sigma_{ji}\sum_{k=0}^{[l''/r]-1 } \phi_{l'+rk+i}\otimes\phi_j \biggr\\|_2 ^ 2 \\ & \leq & 2 \kappa^2 l'r \bigl[l''/r \bigr]\leq\kappa^2 l^2.\end{aligned}\ ] ] therefore , @xmath335 so , we have @xmath336 provided that @xmath337 we can conclude that , for all @xmath338 , @xmath339 provided that @xmath280 satisfies conditions ( [ kappa1 ] ) and ( [ kappa3 ] ) . since also @xmath25 , we have that @xmath340 . next we check that @xmath236 satisfies the conditions of proposition [ kbound ] . it is easy to see that , for all @xmath341 and @xmath342 using the elementary inequality @xmath343 and the fact that @xmath326,\sigma\in\lambda$ ] , we get that @xmath344 a simple computation based on the definition of @xmath345 easily yields that @xmath346\leq8\kappa ^2 l ' l '' \leq4 \kappa^2 l^2.\ ] ] thus , for the @xmath5-fold product - measures @xmath347 , we get @xmath348 for a fixed @xmath322 , this yields @xmath349 \\[-8pt ] \nonumber & \leq&\frac{1}{10 } \log\bigl(\operatorname { card}\bigl(\lambda ' \bigr)-1\bigr),\end{aligned}\ ] ] provided that @xmath350 it remains to use ( [ hemming ] ) and the definition of kernels @xmath331 to bound from below the squared distance @xmath351 for @xmath352 , @xmath353 \geq\frac{1}{64 } \kappa^2 \frac{l^2}{m^2}.\ ] ] since @xmath354 , this implies that @xmath355 in view of ( [ kappa1 ] ) , ( [ kappa2 ] ) and ( [ kappa3 ] ) , we now take @xmath356 with this choice of @xmath280 , @xmath357 . in view of ( [ l2below ] ) and ( [ klabove ] ) , we can use proposition [ kbound ] to get @xmath358 \\[-8pt ] \nonumber & & \qquad\geq \inf_{\hat s } \sup_{p\in{\cal p}}{\mathbb p}_p \bigl\ { \\|\hat s - s_p\\|_{l_2(\pi^2)}^2\geq c_1 \delta_n \bigr\ } \geq c_2,\end{aligned}\ ] ] where @xmath359 and @xmath194 are constants . in the case when @xmath360 , bound ( [ knc ] ) still holds with @xmath361 the proof is an easy modification of the argument in the case when @xmath256 . for @xmath360 , the construction becomes simpler : namely , we define @xmath362 where @xmath363 , and , based on this , redefine kernels @xmath364 . the proof then goes through with minor simplifications . thus , in both cases @xmath365 and @xmath256 , ( [ knc ] ) holds with @xmath366 this is true under the assumption that @xmath253 . note also that @xmath367 . thus , we can replace @xmath368 by the upper bound @xmath369 in the definition of @xmath370 . we can now choose @xmath371 that maximizes @xmath370 to get bound ( [ knc ] ) with @xmath372 . this completes the proof in the case when @xmath373 and @xmath374 . if @xmath375 , it is easy to use the condition @xmath376 and to show that @xmath377 where @xmath378 is a constant depending only on @xmath379 . this completes the proof in the remaining case . proof of theorem [ lowerboundsparse ] the only modification of the previous proof is to replace bound ( [ lr - d ] ) by @xmath380 - 1}\phi_{l'+rk+j}(v ) ) ^2\leq d\sum_{k=0}^{[l''/r]-1}\phi_{l'+rk+j}^2(v ) . $ ] then , the outcome of the next several lines of the proof is that @xmath301 provided that [ instead of ( [ kappa1 ] ) ] @xmath381 as a result , at the end of the proof , we get that ( [ knc ] ) holds with @xmath382 it remains to observe that @xmath383 , which follows from the fact that @xmath384 and to take @xmath201 to complete the proof . in this section , we derive upper bounds on the squared @xmath165-error of the following least squares estimator of the target matrix @xmath10 : @xmath385 where @xmath386 , @xmath387 here @xmath388 denotes a truncation of kernel @xmath389 if @xmath390 , @xmath391 if @xmath392 and @xmath393 if @xmath394 . note that the kernels in the class @xmath395 are symmetric and @xmath396 . note also that the sets @xmath395 , @xmath397 and optimization problem ( [ leastsquare ] ) are not convex . we will prove the following result under the assumption that @xmath25 a.s . recall the definition of the class of kernels @xmath171 in section [ sectlowerbound ] . [ lsupth ] there exist constants @xmath398 such that , for all @xmath93 , with probability at least @xmath95 , @xmath399 \\[-8pt ] \nonumber & & { } + c \biggl(\frac{a^2 ( r\wedge l)l}{n}\log \biggl(\frac{a n m}{(r\wedge l)l } \biggr)+ \frac{a^2 t}{n } \biggr).\end{aligned}\ ] ] in particular , for some constants @xmath400 , for @xmath175 and for all @xmath93 , with probability at least @xmath95 , @xmath401.\ ] ] without loss of generality , assume that @xmath402 this would imply the general case by a simple rescaling of the problem . we will use a version of well - known bounds for least squares estimators over uniformly bounded function classes in terms of rademacher complexities . specifically , consider the following least squares estimator : @xmath403 where @xmath404 are i.i.d . copies of a random couple @xmath405 in @xmath406 , @xmath407 being a measurable space , @xmath408 a.s . , @xmath409 being a class of measurable functions on @xmath410 uniformly bounded by @xmath75 . the goal is to estimate the regression function @xmath411 . define localized rademacher complexity @xmath412 where @xmath20 is the distribution of @xmath413 and @xmath414 is the rademacher process , @xmath415 being a sequence of i.i.d . rademacher random variables independent of @xmath416 . denote @xmath417 and @xmath418 the next result easily follows from theorem 5.2 in @xcite : [ regr ] there exist constants @xmath194 such that , for all @xmath93 , with probability at least @xmath95 , @xmath419 we will apply this proposition to prove theorem [ lsupth ] . in what follows in the proof , denote @xmath420 . in our case , @xmath421 , @xmath22 plays the role of @xmath413 , and @xmath422 plays the role of @xmath20 . let @xmath423 , @xmath424 and @xmath425 . first , we need to upper bound the rademacher complexity @xmath426 for the class @xmath409 . let @xmath427 be the set of all symmetric @xmath15 matrices @xmath34 with @xmath172 and @xmath428 . the @xmath285-covering number @xmath429 of the set @xmath427 with respect to the hilbert schmidt distance ( i.e. , the minimal number of balls of radius @xmath285 needed to cover this set ) can be bounded as follows : @xmath430 such bounds are well known ( see , e.g. , @xcite , lemma 9.3 and references therein ; the proof of this lemma can be easily modified to obtain ( [ cover2 ] ) ) . bound ( [ cover2 ] ) will be used to control the covering numbers of the set of kernels @xmath431 . this set can be easily identified with a subset of the set @xmath432 [ since kernels @xmath433 can be viewed as symmetric @xmath147 matrices of rank at most @xmath434 with @xmath435 and @xmath436 . therefore , we get the following bound : @xmath437 since @xmath438 ( truncation of the entries reduces the hilbert schmidt distance ) , we also have @xmath439 since @xmath440 , @xmath441 therefore , we get the following bound on the @xmath442-covering numbers of the set @xmath443 here @xmath444 denotes the empirical distribution based on observations @xmath445 . the last bound allows us to use inequality ( 3.17 ) in @xcite to control the localized rademacher complexity @xmath426 of the class @xmath409 as follows : @xmath446 \\[-8pt ] \nonumber & \leq & c_1 \biggl[\sqrt{\frac{\delta l(r\wedge l)}{n}}\sqrt{\log \biggl ( \frac{am}{\sqrt{\delta } } \biggr ) } \vee\frac{l(r\wedge l)}{n}\log \biggl(\frac{am}{\sqrt{\delta } } \biggr ) \biggr]\end{aligned}\ ] ] with some constant @xmath447 . this easily yields @xmath448 with some constants @xmath449 . proposition [ regr ] now implies bound ( [ lsoracul ] ) . to prove bound ( [ lsup ] ) , it is enough to observe that , for @xmath175 , @xmath450 indeed , since @xmath175 , we can approximate this kernel by @xmath451 for the error of this approximation , we have @xmath452 & & \quad= m^{-2 } \\|s_l - s_{\ast}\\|_2 ^ 2= m^{-2}\sum_{i\vee j > l } \langle s_{\ast } \phi_i,\phi_j\rangle^2 \\[-2pt ] & & \quad\leq m^{-2}\frac{1}{\lambda_{l+1 } } \sum_{i > l}\sum _ { j=1}^m \lambda _ i\langle s_{\ast}\phi_i,\phi_j\rangle^2 + m^{-2}\frac{1}{\lambda_{l+1 } } \sum_{i=1}^m \sum_{j > l } \lambda _ j\langle s_{\ast}\phi_i,\phi_j\rangle^2\leq \frac{2\rho ^2}{\lambda_{l+1}},\end{aligned}\ ] ] which implies @xmath453 ( since the entries of matrix @xmath10 are bounded by @xmath75 and truncation of the entries reduces the hilbert schmidt distance ) . we also have @xmath454 and @xmath455 therefore , @xmath456 and bound ( [ appro ] ) follows . bound ( [ lsup ] ) is a consequence of ( [ lsoracul ] ) and ( [ appro ] ) . note that , in the case when the basis of eigenfunctions @xmath148 coincides with the canonical basis of space @xmath64 , the following bound holds trivially : @xmath457 this follows from the fact that the entries of both matrices @xmath458 and @xmath459 are bounded by @xmath163 , and their nonzero entries are only in the first @xmath460 rows and the first @xmath460 columns , so , @xmath461 combining this with ( [ lsup ] ) and minimizing the resulting bound with respect to @xmath460 yields the following upper bound ( up to a constant ) that holds for the optimal choice of @xmath460 : @xmath462\vee\frac{a^2 t}{n}.\ ] ] it is not hard to check that , typically , this expression is of the same order ( up to log factors ) as the lower bound of theorem [ lowerboundsparse ] for @xmath191 . next we consider a penalized version of least squares estimator which is adaptive to unknown parameters of the problem ( such as the rank of the target matrix and the optimal value of parameter @xmath460 which minimizes the error bound of theorem [ lsupth ] ) . we still assume that @xmath25 a.s . for some known constant @xmath24 . define @xmath463 \\[-8pt ] \nonumber & & \hspace{45pt}{}+k\frac{a^2 ( r\wedge l)l}{n}\log \biggl(\frac{a n m}{(r\wedge l)l } \biggr ) \biggr\}\end{aligned}\ ] ] and let @xmath464 . here @xmath465 and @xmath466 are fixed constants . the following theorem provides an oracle inequality for the estimator @xmath92 . [ regradapt ] there exists a choice of constants @xmath465 , @xmath466 in ( [ penls ] ) and @xmath87 in the inequality below such that for all @xmath93 with probability at least @xmath95 @xmath467.\nonumber\end{aligned}\ ] ] as in the proof of the previous theorem , we can assume that @xmath402 the general case follows by rescaling . we will use oracle inequalities in abstract penalized empirical risk minimization problems ; see @xcite , theorem 6.5 . we only sketch the proof here skipping the details that are standard . as in the proof of theorem [ lsupth ] , first consider i.i.d . copies @xmath404 of a random couple @xmath405 in @xmath406 , where @xmath407 is a measurable space and @xmath408 a.s . let @xmath468 be a finite family of classes of measurable functions from @xmath410 into @xmath469 $ ] . consider the corresponding family of least squares estimators @xmath470 suppose the following upper bounds on localized rademacher complexities for classes @xmath471 hold : @xmath472 where @xmath473 are nondecreasing functions of @xmath474 that do not depend on the distribution of @xmath405 . let @xmath475,\ ] ] and @xmath476 are constants and @xmath477 are positive numbers . define the following penalized least squares estimator of the regression function . the next result is well known ; it can be deduced , for instance , from theorem 6.5 in @xcite . [ regror ] there exists constants @xmath478 in the definition ( [ regpe ] ) of @xmath479 and a constant @xmath480 such that , for all @xmath481 , with probability at least @xmath482 @xmath483.\ ] ] we apply this result to the estimator @xmath484 , where @xmath485 is defined by ( [ penls ] ) ( with @xmath486 ) . in this case , @xmath421 , @xmath22 plays the role of @xmath413 , @xmath487 , @xmath488 , @xmath489 . in view of ( [ radem ] ) , we can use the following bounds on localized rademacher complexities for these function classes : @xmath490\ ] ] with some constant @xmath491 , and we have @xmath492 with some constant @xmath493 . define @xmath494 . this yields the bound @xmath495 these considerations and proposition [ regror ] imply the claim of the theorem . it follows from theorem [ regradapt ] that , for some constant @xmath87 and for all @xmath93 , @xmath496 where @xmath497 . $ ] denoting @xmath498 it is easy to see that @xmath499 suppose that , for some @xmath158 , @xmath213 . under this assumption , it is easy to show that the upper bound on the squared @xmath58-error of the estimator @xmath92 is of the order @xmath500 ( in fact , the log factors can be written in a slightly better , but more complicated way ) . up to the log factors , this is the same error rate as in the lower bounds of section [ sectlowerbound ] ; see ( [ minmaxlower ] ) . our main goal in this section is to study the following penalized least squares estimator with a combination of two convex penalties : @xmath501,\hspace{-35pt}\ ] ] where @xmath502 is a closed convex set of symmetric kernels such that , for all @xmath503 , @xmath504 and @xmath505 are regularization parameters . the first penalty involved in ( [ lsnucsob ] ) is based on the nuclear norm @xmath506 , and it is used to `` promote '' low - rank solutions . the second penalty is based on a `` sobolev type norm '' @xmath507 . it is used to `` promote '' the smoothness of the solution on the graph . we will derive an upper bound on the error @xmath508 of estimator @xmath509 in terms of spectral characteristics of the target kernel @xmath10 and matrix @xmath132 . as before , @xmath132 is a nonnegatively definite symmetric kernel with spectral representation @xmath133 where @xmath134 are the eigenvalues of @xmath132 repeated with their multiplicities and @xmath135 are the corresponding orthonormal eigenfunctions . we will also use the decomposition of identity associated with @xmath132 : @xmath510 clearly , @xmath511 is a nondecreasing projector - valued function . despite the fact that the eigenfunctions @xmath512 are not uniquely defined in the case when @xmath132 has multiple eigenvalues , the decomposition of identity @xmath513 is uniquely defined ( in fact , it can be rewritten in terms of spectral projectors of @xmath132 ) . the distribution of the eigenvalues of @xmath132 is characterized by the following _ spectral function _ : @xmath514 denote @xmath515 ( in other words , @xmath516 is the smallest @xmath517 such that @xmath518 ) . it was assumed in the that there exists a constant @xmath519 such that @xmath520 for all @xmath140 . in what follows , we use a regularized majorant of spectral function @xmath521 . let @xmath522 be a nondecreasing function such that @xmath523 , the function @xmath524 is nonincreasing and , for some @xmath525 , @xmath526 without loss of generality , we assume in what follows that @xmath527 [ otherwise , one can take the function @xmath528 instead ] . the conditions on @xmath529 are satisfied if for some @xmath525 , the function @xmath530 is nonincreasing : in this case , @xmath531 is also nonincreasing and @xmath532 consider a kernel @xmath130 ( an oracle ) with spectral representation : @xmath533 , where @xmath534 , @xmath535 are nonzero eigenvalues of @xmath34 ( possibly repeated ) and @xmath536 are the corresponding orthonormal eigenfunctions . denote @xmath537 . the following _ coherence function _ will be used to characterize the relationship between the kernels @xmath34 and @xmath132 : @xmath538 it is immediate from this definition that @xmath539 note also that @xmath540 is a nondecreasing function of @xmath541 and @xmath542 [ for @xmath543 , @xmath540 can be interpreted as a `` partial rank '' of @xmath34 ] . as in the case of spectral function @xmath34 , we need a regularized majorant for the coherence function @xmath540 . denote by @xmath545 the set of all nondecreasing functions @xmath546 such that @xmath547 is nonincreasing and @xmath548 . it is easy to see that the class of functions @xmath549 contains the smallest function ( uniformly in @xmath550 ) that will be denoted by @xmath551 and it is given by the following expression : @xmath552 it easily follows from this definition that @xmath553 . note that since the function @xmath554 is nonincreasing and it is equal to @xmath555 for @xmath556 , we have @xmath557 [ main ] there exists constants @xmath562 depending only on @xmath563 such that , for all @xmath564 $ ] with probability at least @xmath95 , @xmath565 \\ & & \qquad\quad{}+ c\frac{a^2 t_{n , m}}{n}.\nonumber\end{aligned}\ ] ] \(2 ) note also that theorem [ main ] holds in the case when @xmath569 . in this case , our method coincides with nuclear norm penalized least squares ( matrix lasso ) and @xmath570 , so the bound of theorem [ main ] becomes @xmath571 + c\frac{a^2 t_{n , m}}{n}.\hspace{-35pt}\ ] ] similar oracle inequalities were proved in @xcite for a linearized least squares method with nuclear norm penalty . using simple aggregation techniques , it is easy to construct an adaptive estimator for which the oracle inequality of theorem [ main ] holds with the optimal value of @xmath572 that minimizes the right - hand side of the bound . to this end , divide the sample @xmath573 into two parts , @xmath574 where @xmath575 + 1 $ ] . the first part of the sample will be used to compute the estimators @xmath576 , @xmath577 , @xmath578 [ they are defined by ( [ lsnucsob ] ) , but they are based only on the first @xmath579 observations ] . the second part of the sample is used for model selection @xmath580 finally , let @xmath581 . [ mainadapt ] under the assumptions and notation of theorem [ main ] , with probability at least @xmath95 , @xmath582 } \bigl(m^2 \eps^2 \bar\varphi\bigl(s;\bar \eps^{-1}\bigr ) + \bar\eps\bigl\\|w^{1/2}s\bigr\\|_{l_2(\pi^2)}^2 \bigr ) \bigr ] \\ & & { } + c\frac{a^2 ( \log(m+1)+t_{n , m})}{n}.\nonumber\end{aligned}\ ] ] the idea of aggregation result behind this theorem is rather well known ; see @xcite , chapter 8 . the proof can be deduced , for instance , from proposition [ regr ] used in section [ sectlsnonconvex ] . specifically , this proposition has to be applied in the case when @xmath409 is a finite class of functions bounded by @xmath75 . let @xmath583 . then , for some numerical constant @xmath584 @xmath585\ ] ] ( see , e.g. , @xcite , theorem 3.5 ) , and proposition [ regr ] easily implies that , for all @xmath93 , with probability at least @xmath95 @xmath586 where @xmath493 is a constant . we will assume that @xmath486 ( in the general case , the result would follow by rescaling ) and use bound ( [ finiteaggregate ] ) , conditionally on the first part of the sample , in the case when @xmath587 . then , given @xmath588 , with probability at least @xmath95 , @xmath589 by theorem [ main ] [ with @xmath590 replaced by @xmath591 and the union bound , we get that , with probability at least @xmath95 , for all @xmath578 , @xmath592\\ & & \qquad\quad { } + c_3\frac{\log(m+1)+t_{n , m}}{n}\nonumber\end{aligned}\ ] ] with some constant @xmath593 . therefore , the minimal error of estimators @xmath458 , @xmath594 , can be bounded with the same probability by the minimum over @xmath595 of the expression in the right - hand side of ( [ boundthmain ] ) . moreover , using monotonicity of the function @xmath596 and the condition that @xmath597 , it is easy to replace the minimum over @xmath460 by the infimum over @xmath572 . combining the resulting bound with ( [ qua - qua ] ) and adjusting the constants yields the claim . using more sophisticated aggregation methods ( e.g. , such as the methods studied in @xcite ) it is possible to construct an estimator @xmath92 for which the oracle inequality similar to ( [ boundthmainadapt ] ) holds with constant @xmath75 in front of the approximation error term @xmath598 . to understand better the meaning of function @xmath599 involved in the statements of theorems [ main ] and [ mainadapt ] , it makes sense to relate it to the low coherence assumptions discussed in the . indeed , suppose that , for some @xmath600 , @xmath601 this is a part of standard low coherence assumptions on matrix @xmath34 with respect to the orthonormal basis @xmath602 ; see ( [ cohera ] ) . clearly , it implies that ) with ( [ lowvarphi ] ) . ] @xmath603 suppose that @xmath604 and @xmath605 if condition ( [ up_varphi ] ) holds for the target kernel @xmath10 with @xmath82 and some @xmath83 , then theorem [ main ] implies that with probability at least @xmath95 , @xmath606 and theorem [ mainadapt ] implies that with the same probability , @xmath607 } \biggl(\frac{a^2\nu r \bar f(\bar\eps^{-1 } ) \log(2m)}{n } + \bar\eps \bigl\\|w^{1/2}s_{\ast}\bigr\\|_{l_2(\pi^2)}^2 \biggr)\\ & & { } + c \frac{a^2(\log(m+1)+ t_{n , m})}{n}.\end{aligned}\ ] ] if @xmath608 for some @xmath609 , then it is easy to check that @xmath610 . under the assumption that @xmath611 , we get the bound @xmath612 under the following slightly modified version of low coherence assumption ( [ up_varphi ] ) , @xmath613 one can almost recover upper bounds of section [ sectlsnonconvex ] , @xmath614 the main difference with what was proved in section [ sectlsnonconvex ] is that now the low coherence constant @xmath85 is involved in the bounds , so the methods discussed in this section yield correct ( up to log factors ) error rates provided that the target kernel @xmath10 has `` low coherence '' with respect to the basis of eigenfunctions of @xmath132 . proof of theorem [ main ] bound ( [ bound_th_main ] ) will be proved for a fixed oracle @xmath503 and an arbitrary function @xmath615 with @xmath616 instead of @xmath617 . it then can be applied to the function @xmath599 ( which is the smallest function in @xmath549 ) . without loss of generality , we assume that @xmath486 ; the general case then follows by a simple rescaling . finally , we will denote @xmath618 throughout the proof . define the following orthogonal projectors @xmath619 in the space @xmath68 with hilbert schmidt inner product : @xmath620 we will use a well known representation of subdifferential of convex function @xmath621 : @xmath622 where @xmath623 ; see @xcite , appendix a.4 and references therein . denote @xmath624 so that @xmath625 . an arbitrary matrix @xmath626 can be represented as @xmath627 where @xmath628 . since @xmath629 is a minimizer of @xmath630 , there exists a matrix @xmath631 such that @xmath632 belongs to the normal cone of @xmath633 at the point @xmath629 ; see @xcite , chapter 2 , corollary 6 . this implies that @xmath634 and , in view of ( [ aravno ] ) , @xmath635 \\[-8pt ] \nonumber & & \qquad{}+ 2\frac{\bar\eps}{m^2 } \langle w\hat{s},\hat{s}-s \rangle\leq0.\end{aligned}\ ] ] here and in what follows @xmath636 denotes the empirical distribution based on the sample @xmath637 . the corresponding true distribution of @xmath18 will be denoted by @xmath638 . it easily follows from ( [ aravno1 ] ) that @xmath639 where @xmath640 we can now rewrite the last bound as @xmath641 and use a simple identity @xmath642 to get the following bound : @xmath643 for an arbitrary @xmath644 , @xmath645 , where @xmath646 is a matrix with @xmath647 . it follows from the trace duality property that there exists an @xmath646 with @xmath648 [ to be specific , @xmath649 such that @xmath650 where the first equality is based on the fact that @xmath651 is a self - adjoint operator and the second equality is based on the fact that @xmath34 has support @xmath79 . using this equation and monotonicity of subdifferentials of convex functions , we get @xmath652 substituting this into the left - hand side of ( [ eqstep1 ] ) , it is easy to get @xmath653 \\[-8pt ] \nonumber & & \qquad\quad{}-2\bar\eps\bigl\langle w^{1/2}s , w^{1/2}(\hat s - s)\bigr\rangle_{l_2(\pi^2 ) } \\ & & \qquad\quad { } + 2 \langle\xi,\hat s - s\rangle + 2(p - p_n ) ( s - s_{\ast } ) ( \hat s - s)\nonumber \\ & & \qquad\quad{}+2(p - p_n ) ( \hat s - s)^2.\nonumber\end{aligned}\ ] ] we need to bound the right - hand side of ( [ basic ] ) . we start with deriving a bound on @xmath654 , expressed in terms of function @xmath655 . note that , for all @xmath656 , @xmath657 which easily implies @xmath658 \\[-8pt ] \nonumber & & \qquad\quad{}+ \biggl(\sum_{\lambda_k>\lambda } \frac{\\|\operatorname { sign}(s)\phi_k\\|^2}{\lambda_k } \biggr)^{1/2 } \biggl(\sum_{\lambda_k>\lambda } \lambda_k\bigl\\|(\hat s - s)\phi_k\bigr\\| ^2 \biggr)^{1/2 } \\ & & \qquad\leq \biggl(\sum_{\lambda_k\leq\lambda } \\|p_l \phi_k\\|^2 \biggr)^{1/2}\\| \hat s - s \\|_2 + \biggl(\sum_{\lambda_k>\lambda } \frac{\\|p_l\phi_k\\|^2}{\lambda_k } \biggr)^{1/2 } \bigl\\|w^{1/2}(\hat s - s ) \bigr\\|_2.\nonumber\hspace*{-35pt}\end{aligned}\ ] ] denote @xmath661 . suppose that @xmath662 $ ] for some @xmath663 . we will use the properties of functions @xmath615 and @xmath529 . in particular , recall that the functions @xmath664 and @xmath531 are nonincreasing . using these properties and the condition that @xmath665 we get @xmath666 which proves the first bound . to prove the second bound , replace in the inequalities above @xmath667 by @xmath75 and @xmath668 by @xmath669 . in the case when @xmath556 , both bounds are trivial since their left - hand sides are equal to zero . it follows from from ( [ odin ] ) and the first bound of lemma [ boundsums ] that @xmath670 this implies the following bound : @xmath671 where we used twice an elementary inequality @xmath672 . we will apply this bound for @xmath673 to get the following inequality : @xmath674 to bound the next term in the right - hand side of ( [ basic ] ) , note that @xmath675 \\[-8pt ] \nonumber & & \qquad\leq \bar\eps\bigl\\|w^{1/2}s \bigr\\|_{l_2(\pi^2)}^2 + \frac{\bar\eps}{4}\bigl\\| w^{1/2}(\hat s - s)\bigr\\|_{l_2(\pi^2)}^2.\end{aligned}\ ] ] the main part of the proof deals with bounding the stochastic term @xmath676 on the right - hand side of ( [ basic ] ) . to this end , define ( for fixed @xmath677 ) @xmath678 and consider the following empirical process : @xmath679 where @xmath680 clearly , we have @xmath681 \\[-8pt ] \nonumber & & \qquad\leq 2\alpha_n \bigl(\\|\hat s - s\\|_{l_2(\pi^2 ) } , \bigl\\|{\cal p}_{l}^{\perp}\hat s\bigr\\|_1,\bigl \\|w^{1/2}(\hat s - s)\bigr\\|_{l_2(\pi^2 ) } \bigr),\end{aligned}\ ] ] and it remains to provide an upper bound on @xmath682 that is uniform in some intervals of the parameters @xmath683 ( such that either the norms @xmath684 belong to these intervals with a high probability , or bound of the theorem trivially holds ) . note that the functions @xmath685 are uniformly bounded by a numerical constant ( under the assumptions that @xmath486 , @xmath25 and all the kernels are also bounded by @xmath163 ) and we have @xmath686 with some numerical constant @xmath687 . using talagrand s concentration inequality for empirical processes we conclude that for fixed @xmath688 with probability at least @xmath95 and with some constant @xmath689 we will make this bound uniform in @xmath690 , \delta_k^-<\delta_k^+ , k=1,2,3 $ ] ( these intervals will be chosen later ) . define @xmath691 + 1 , k=1,2,3 $ ] and let @xmath692 + 2 ) . $ ] by the union bound , with probability at least @xmath95 and for all @xmath693 + 1 , k=1,2,3 $ ] , @xmath694 by monotonicity of @xmath695 and of the right - hand side of the bound with respect to each of the variables @xmath696 , we conclude that with the same probability and with some numerical constant @xmath697 , for all @xmath698 , k=1,2,3 $ ] , @xmath699 to bound the expectation @xmath700 on the right - hand side of ( [ talagr ] ) , note that , by the definition of function @xmath701 , @xmath702 & & \qquad\quad{}+ { \mathbb e}\sup \bigl\ { \bigl\|(p_n - p ) ( a - s)^2 \bigr\| \dvtx a\in{\cal t}(\delta_1,\delta_2 , \delta_3 ) \bigr\}.\nonumber\end{aligned}\ ] ] a standard application of symmetrization inequality followed by contraction inequality for rademacher sums ( see , e.g. , @xcite , chapter 2 ) yields @xmath703 \\[-9pt ] \nonumber & & \qquad \leq 16 { \mathbb e}\sup \bigl\ { \bigl\|r_n(a - s ) \bigr\|\dvtx a\in{\cal t}(\delta_1,\delta_2 , \delta_3 ) \bigr\}.\end{aligned}\ ] ] it easily follows from ( [ expal1 ] ) and ( [ expal2 ] ) that @xmath704 \\[-9pt ] \nonumber & & { } + 16 { \mathbb e}\sup \bigl\{\bigl\|\langle\xi_2 , a - s\rangle\bigr\| \dvtx a\in{\cal t}(\delta_1,\delta_2 , \delta_3 ) \bigr\},\end{aligned}\ ] ] where @xmath705 and @xmath706 being i.i.d . rademacher random variables independent of @xmath637 . we will upper bound the expectations on the right - hand side of ( [ expal3 ] ) , which reduces to bounding @xmath707 for each of the random matrices @xmath708 . for @xmath709 and @xmath710 , we have @xmath711 & \leq&\bigl\|\langle{\cal p}_l\xi_i , a - s\rangle\bigr\|+ \\|\xi_i\\|\bigl\\|{\cal p}_l^{\perp}(a ) \bigr\\|_1\\[-2pt ] & \leq&\bigl\|\langle{\cal p}_l \xi_i , a - s\rangle\bigr\|+ \delta_2 \\|\xi_i\\|.\nonumber\end{aligned}\ ] ] to bound @xmath712 , we use the following simple corollary of a well - known noncommutative bernstein inequality ( see , e.g. , @xcite ) obtained by integrating exponential tails of this inequality : let @xmath713 be a random symmetric matrix with @xmath714 , @xmath715 and @xmath716 for some @xmath717 and let @xmath718 be @xmath5 i.i.d . copies of @xmath713 . then @xmath719 it is applied to i.i.d . random matrices @xmath720 in the case of matrix @xmath721 and to i.i.d . random matrices @xmath722 in the case of matrix @xmath723 . in both cases , @xmath724 and , by a simple computation , @xmath725 ( see , e.g. , @xcite , section 9.4 ) , bound ( [ opber ] ) implies that , for @xmath709 , @xmath726= : \eps^{\ast}.\ ] ] for all symmetric @xmath15 matrices @xmath646 , @xmath730 assuming that @xmath731 \bigl\\|w^{1/2}m\bigr\\|_2 ^ 2 & = & \sum_{k , j=1}^m \lambda_k \bigl\| \langle m , \phi_k\otimes\phi_j\rangle \bigr\|^2 \leq1,\end{aligned}\ ] ] it is easy to conclude that @xmath732 it follows @xmath733 & & \qquad\leq \biggl(\sum_{k , j=1}^m \bigl ( \lambda_k^{-1}\wedge\delta^2\bigr)\bigl\|\langle{\cal p}_l \xi , \phi_k \otimes\phi_j \rangle\bigr\|^2 \biggr)^{1/2 } \biggl(\sum _ { k , j=1}^m \frac{\|\langle m , \phi_k\otimes\phi_j\rangle\|^2}{\lambda _ k^{-1}\wedge\delta^2 } \biggr)^{1/2 } \\ & & \qquad\leq \sqrt{2 } \biggl(\sum_{k , j=1}^m \bigl(\lambda_k^{-1}\wedge\delta^2\bigr)\bigl\| \langle{\cal p}_l \xi , \phi_k \otimes\phi_j \rangle\bigr\|^2 \biggr)^{1/2}.\nonumber\end{aligned}\ ] ] consider the following inner product : @xmath734 and let @xmath735 be the corresponding norm . we will provide an upper bound on @xmath736 recall that @xmath737 where @xmath738 for @xmath739 and @xmath740 for @xmath741 . note that in the first case @xmath742 , and in the second case @xmath743 . therefore , @xmath744 \\[-9pt ] \nonumber & \leq&2\sqrt{\frac{{\mathbb e}\\|{\cal p}_l e_{x , x'}\\|_w^2}{n}}.\end{aligned}\ ] ] it remains to bound @xmath745 , @xmath746 & & \qquad= { \mathbb e}\sum_{k , j=1}^m \bigl(\lambda_k^{-1}\wedge\delta^2\bigr)\bigl\|\bigl \langle { \cal p}_l ( e_{x , x ' } ) , \phi_k \otimes \phi_j\bigr\rangle\bigr\|^2 \nonumber\\[-2pt ] & & \qquad= \sum_{k , j=1}^m \bigl ( \lambda_k^{-1}\wedge\delta^2\bigr ) m^{-2}\sum_{u , v\in v}\bigl\|\bigl\langle e_{u , v } , { \cal p}_l(\phi_k \otimes\phi _ j)\bigr\rangle\bigr\|^2 \nonumber\\[-2pt ] & & \qquad\leq m^{-2}\sum_{k , j=1}^m \bigl(\lambda_k^{-1}\wedge\delta^2\bigr ) \bigl\\| { \cal p}_l(\phi_k \otimes\phi_j ) \bigr\\|_2 ^ 2 \\[-2pt ] & & \qquad\leq 2m^{-2}\sum_{k , j=1}^m \bigl(\lambda_k^{-1}\wedge\delta^2\bigr ) \bigl(\\|p_l\phi_k\\|^2+\\|p_l \phi_j\\|^2\bigr ) \nonumber\\[-2pt ] & & \qquad= 2m^{-1}\sum_{k=1}^m \bigl(\lambda_k^{-1}\wedge\delta^2\bigr ) \\|p_l\phi_k\\|^2 + 2m^{-2}\sum _ { k=1}^m \bigl(\lambda_k^{-1 } \wedge\delta^2\bigr)\\|p_l\\|_2 ^ 2 \nonumber\\[-2pt ] & & \qquad= 2m^{-1}\sum_{k=1}^m \bigl(\lambda_k^{-1}\wedge\delta^2\bigr ) \\|p_l\phi_k\\|^2 + 2m^{-2}r\sum _ { k=1}^m \bigl(\lambda_k^{-1 } \wedge\delta^2\bigr).\nonumber\end{aligned}\ ] ] note that @xmath747 using the first bound of lemma [ boundsums ] , we get from ( [ hi - hi ] ) that @xmath748 \\[-8pt ] \nonumber & = & ( c_{\gamma}+1 ) \delta^2 \varphi\bigl(\delta^{-2}\bigr).\end{aligned}\ ] ] we also have @xmath749 which , by the second bound of lemma [ boundsums ] , implies that @xmath750 using bounds ( [ var ] ) , ( [ au ] ) and ( [ ua ] ) and the fact that @xmath751 , we get @xmath752 \\[-8pt ] \nonumber & \leq & 4m^{-1}(c_{\gamma}+1 ) \delta^2\varphi\bigl(\delta^{-2}\bigr).\end{aligned}\ ] ] let @xmath753 . using lemma [ lemma1 ] , we get @xmath754 in the case when @xmath755 , we get @xmath756 in the opposite case , when @xmath757 , we use the fact that the function @xmath758 is nonincreasing . this implies that @xmath759 , and we get @xmath760 we can conclude that @xmath761 this bound will be combined with ( [ pjat ] ) and ( [ expectxi ] ) to get that , for @xmath709 , @xmath762 in view of ( [ expal3 ] ) , this yields the bound @xmath763 that holds with some constant @xmath764 for all @xmath765 . using ( [ talagr ] ) , we conclude that for some constants @xmath766 and for all @xmath698 , k=1,2,3 $ ] , @xmath767\ ] ] that holds with probability at least @xmath95 . this yields the following upper bound on the stochastic term in ( [ basic ] ) [ see also ( [ stochterm ] ) ] : @xmath768\nonumber\end{aligned}\ ] ] that holds provided that @xmath769,\qquad \bigl\\|{\cal p}_l^{\perp } \hat s \bigr\\|_1\in\bigl[\delta_2 ^ -,\delta_2^+\bigr ] , \nonumber \\[-8pt ] \\[-8pt ] \nonumber \bigl\\|w^{1/2}(\hat s - s)\bigr\\|_{l_2(\pi^2)}&\in&\bigl[\delta_3 ^ - , \delta_3^+\bigr].\end{aligned}\ ] ] we substitute bound ( [ stochtermbound ] ) in ( [ basic ] ) and further bound some of its terms as follows : @xmath770 and @xmath771 we will also use ( [ tri ] ) to control the term @xmath772 in ( [ basic ] ) and ( [ chetyre ] ) to control the term @xmath773 . if condition ( [ condeps ] ) holds with @xmath774 , then @xmath775 . by a simple algebra , it follows from ( [ basic ] ) that @xmath776 with some constant @xmath584 . since , under condition ( [ condeps ] ) with @xmath486 , @xmath777 , we can conclude that @xmath778 \\[-8pt ] \nonumber & & \qquad\leq \\|s - s_{\ast}\\|_{l_2(\pi^2)}^2 + c_2 m^2 \eps^2\varphi\bigl(\bar\eps^{-1}\bigr ) + \bar\eps\bigl\\|w^{1/2}s\bigr\\|_{l_2(\pi^2)}^2 + \frac{\bar t}{n}\end{aligned}\ ] ] with some constant @xmath493 . we still have to choose the values of @xmath779 and to handle the case when conditions ( [ conddel ] ) do not hold . first note that due to the assumption that @xmath780 , we have @xmath781 , @xmath782 and @xmath783 thus , we can set @xmath784 , which guarantees that the upper bounds of ( [ conddel ] ) are satisfied . we will also set @xmath785 in the case when one of the lower bounds of ( [ conddel ] ) does not hold , we can still use inequality ( [ stochtermbound ] ) , but we have to replace each of the norms @xmath786 which are smaller than the corresponding @xmath787 by the quantity @xmath787 . then it is straightforward to check that inequality ( [ finis ] ) still holds for some value of constant @xmath493 . with the above choice of @xmath788 , we have @xmath789 this completes the proof .	let @xmath0 be a weighted graph with a finite vertex set @xmath1 , with a symmetric matrix of nonnegative weights @xmath2 and with laplacian @xmath3 . let @xmath4 be a symmetric kernel defined on the vertex set @xmath1 . consider @xmath5 i.i.d . observations @xmath6 , where @xmath7 are independent random vertices sampled from the uniform distribution in @xmath1 and @xmath8 is a real valued response variable such that @xmath9 . the goal is to estimate the kernel @xmath10 based on the data @xmath11 and under the assumption that @xmath10 is low rank and , at the same time , smooth on the graph ( the smoothness being characterized by discrete sobolev norms defined in terms of the graph laplacian ) . we obtain several results for such problems including minimax lower bounds on the @xmath12-error and upper bounds for penalized least squares estimators both with nonconvex and with convex penalties .
according to the birkgoff theorem @xcite _ the metric of a vacuum spherically symmetric space time is static , singular in the origin , unique and in spherically symmetric coordinates has the well - known schwarzschild form _ ( when @xmath0 signature form is chosen ) @xmath1 where @xmath2 is the schwarzschild radius . we define the regularity of a metric at a point as i ) an indefinite differentiability of all metric components at this point ; ii ) the condition @xmath3 holds true at this point relative to any coordinate system associated with our system by a unique and continuous transformation . petrov @xcite has shown a counterexample of the metric contradicting the birkgoff theorem @xmath4 where @xmath5 . metric ( [ e02 ] ) is not static . here it is necessary to note that ( [ e01 ] ) and ( [ e02 ] ) are defined in the same coordinate system . further , one can find another argument against the uniqueness of ( [ e01 ] ) .the vacuum einstein equations @xmath6 , @xmath7 are nonlinear second order differential equations . therefore , according to the general theory of differential equations@xcite one should have at least two constants of integration in the solution , but schwarzschild metric ( [ e01 ] ) has only one . in this letter we would like to show that it is possible to construct more general metrics in the spherically symmetric and axially symmetric cases . because of the letter format we have to restrict the consideration only to the main results without proving the theorems , which we plan to do in the next paper . the question about the applicability of our results in brane models with noncompact extra dimensions and in astrophysics will also be discussed in subsequent publications . * theorem 1 . * _ the _ _ spherically symmetric , static and asymptotically free metric , which is a solution of vacuum einstein equations @xmath6 , @xmath7 has the following form : _ @xmath8{r^{3 } + b^{3}}}\biggr ) dt^{2 } - \frac { r^{4 } dr^{2}}{\biggl(\sqrt[3]{r^{3 } + b^{3 } } \biggr)^{4 } \biggl(1 - \frac{a}{\sqrt[3 ] { r^{3 } + b^{3 } } } \biggr ) } \nonumber \\ & - & \biggl(\sqrt[3]{r^{3 } + b^{3}}\biggr)^{2}\biggl(d\theta^{2 } + \sin^{2 } \theta d\varphi^{2 } \biggr ) , \label{e03}\end{aligned}\ ] ] _ where @xmath9 is the schwarzschild radius and @xmath10 is an additional parameter of the metric . this metric is unique to within conformal transformations . _ * conclusion 1 . * _ at @xmath11 we obtain schwarzschild metric ( [ e01 ] ) . _ * conclusion 2 . * _ at @xmath12 we obtain particle- like version ( analogously to ref . @xcite ) of schwarzschild metric @xcite . it was first obtained by k.schwarzschild under the condition @xmath13 , which has violated the general covariance of the theory . _ thus , metric ( [ e03 ] ) has the following properties : 1 . when @xmath11 ( the classical schwarzschild case ) it has the curvature singularity at @xmath14 and the coordinate singularity at @xmath15 ; 2 . when @xmath16 it has only the coordinate singularity at @xmath17 { a^{3}-b^{3}}$ ] ; 3 . when @xmath18 metric ( [ e03 ] ) is regular everywhere . the fulfillment of the last contention can be verified by the transformation to the kruskal - type coordinates @xmath19 where @xmath20{r^{3}+b^{3}}$ ] . accordingly , metric ( [ e03 ] ) takes the form @xmath21 it is necessary to call the reader s attention to the fact that formally ( [ e03 ] ) corresponds to all the metric requirements of the differentiable manifold only in the case @xmath18 . the case @xmath22 requires singular coordinate transformations , and that is not completely correct from the formal mathematical point of view while working with smooth manifolds . * remark . * _ by chainging radial coordinate _ @xmath23 _ to _ @xmath24 _ metric ( [ e03 ] ) may be represented in the schwarzschild form _ @xmath25 it seems that parameter @xmath26 can be cancelled completely , but it would not be correct . in fact parameter @xmath26 is included into the @xmath24 domain of definition , namely @xmath27 for this reason variable @xmath24 is not equal to the radial coordinate in all the domain of definition . therefore , any attempt to neglect the parameter @xmath26 leads to a loss of generality of the results obtained . moreover , if the schwarzschild radius @xmath28 is defined in such a manner that it corresponds to newton s law at @xmath29 , then , the value of @xmath26 only changes the geodesics deviations near the origin . * lemma . * _ if one has spherically symmetric space time with metric ( [ e03 ] ) , the radial null geodesic equation is _ @xmath30 _ at @xmath31 the asymptotic behavior is : _ @xmath32 theorem 1 perhaps improved if one could cancel the condition of staticity . however , the requirement of asymptotical flatness of the metric is essentially necessary due to the petrov example . further , the unexpected behaviour of the newton gravity force near the origin when a particle like version of the metric ( with the appropriate choice of the value of parameter @xmath26 ) occurs . the gravity force vanishes at the origin when @xmath12 and tends to infinity as @xmath33 at @xmath34 . such unusual behaviour may be responsible for a gravitational collapse without finishing in a black hole configuration @xcite . * theorem 2 . * _ the three parametric generalization of reissner - nordstrom metric is given by _ @xmath35 _ where @xmath20{r^{3}+b^{3}}$ ] and @xmath36 is a u(1 ) type charge . this metric is regular at the origin when @xmath37 and everywhere when @xmath38 and @xmath39 . _ * theorem 3 . * _ the three parametric generalization of the kerr metric is given by _ @xmath40 _ where @xmath41 is a specific inertia moment , this metric is regular everywhere when @xmath37 and coordinate singularities are not taken into account . in this letter the metric that generalizes the well - known schwarzchild ( reissner - nordstrom , kerr ) metric is considered . this metric includes ( as a partial case ) the schwarzchild metric ( when an additional parameter @xmath26 vanishes ) , a class of metrics containing only the event horizon without the origin singularity and , finally , a class of singularity free particle - like metrics . the proof of the theorems follows . the applicability of this type of metric rather than the schwarzchild metric to describe the existing astrophysical data ( on accretion , for instance ) and to increase the generality of our knowledge of string black holes are under the investigation . the author would like to thank s.alexeyev and all the participants of the zelmanov memorial seminar ( http://xray.sai.msu.ru/sciwork/zelmanov ) for the useful discussions on the subject of this paper .	regular generalizations of spherically and axially symmetric metrics and their properties are considered . newton gravity law generalizations are reduced for null geodesics .
many x - ray sources ( xrss ) are inside globular clusters ( gcs ; e.g. heinke 2011 ) , but only five of them unambiguosly host a black hole ( bh ; maccarone et al . 2007 , hereafter m07 ; brassington et al . 2010 ; shih et al . 2010 ; maccarone et al . 2011 ; irwin et al . 2010 , hereafter i10 ; see king 2011 for an alternative explanation ) . these five objects were observed as luminous xrss ( @xmath6 ) with strong variability ( so that blending is excluded ; kalogera et al . the spectra of at least two of their host gcs are quite peculiar . one of them , rz 2109 ( in the elliptical ngc 4472 ; zepf et al . 2007 , 2008 , and steele et al . 2011 ; hereafter z07 , z08 , s11 ) displays prominent and broad [ oiii]@xmath75007,4959 emission ( the fwhm of [ oiii]@xmath8 is @xmath9@xmath10 ; its flux is @xmath11 ; at @xmath12 it corresponds to @xmath13 ; see z08 ) , but no h line emission@xmath14 , [ nii]@xmath15 , [ sii]@xmath16 ) is detected , except perhaps [ oiii]@xmath17 . ] . the spectrum of the gc hosting the xrs cxoj033831.8 - 352604 ( in the elliptical ngc 1399 ; see i10 ) is somewhat similar in the absence of h and the presence of [ oiii ] lines ; but widths and luminosities are lower ( fwhm @xmath18 , @xmath19 ) , and [ nii ] @xmath76583,6548 emission is comparable to [ oiii ] . these observations are hard to explain , because of the high [ oiii]/h@xmath20 ratio ( z08 estimate @xmath21 ; i10 finds a 3@xmath22 lower limit of 5 ; both are very uncertain ) and of the large broadening , implying motions much faster than gc velocity dispersions ( @xmath23 ) . optical lines are too wide for a planetary nebula , and too narrow for a supernova remnant . z08 suggest that the [ oiii ] emission from rz 2109 might come from a strong wind originating around the bh , photoionized by the xrs . such scenario strongly favours a bh with mass @xmath24 , rather than an intermediate - mass bh ( imbh ) with mass in the 10@xmath25 - 10@xmath26m@xmath27 range . this was preferred to the hypothesis that the [ oiii]-emitting gas closely orbits ( @xmath28 ) an imbh , on the basis of considerations about the maximum possible emission from such a small volume . such argument is severely weakened if the orbiting gas is very metal - rich , e.g. as the result of the tidal disruption of a white dwarf ( wd ) by an imbh ( see i10 ) : according to porter ( 2010 ) , the cxoj033831.8 - 352604 data are compatible with this scenario , whereas in the case of rz 2109 the conflict remains ; however , clausen & eracleous ( 2011 ) showed that simulations of wd disruption can reproduce the [ oiii ] line profiles observed in rz 2109 . on the other hand , maccarone & warner ( 2011 ) argued that the line emission of cxoj033831.8 - 352604 might be better explained with the photoionization ( by the xrs ) of the wind from an r coronae borealis star in the same gc . in this paper , we examine a further possible scenario , and apply it to the case of rz 2109 : [ oiii ] lines might be associated with a recent nova eruption in the core of the gc . in particular , we discuss the case ( first mentioned by s11 ) where the [ oiii ] emission comes from the photoionization of the nova ejecta by the xrs . x - ray emission from novae is typically fainter ( @xmath29 , even though at least one nova was reported to reach @xmath30 - see henze et al . 2010 ) and shorter ( decay times @xmath31 ) than what is observed for these sources . thus , we assume that the nova eruption is _ not _ associated to the xrs , i.e. that the co - existence of the two phenomena in the same gc is serendipitous . nova eruptions occur in binary systems where a wd accretes from a normal star ( e.g. gehrz et al . 1998 ) . nuclear reactions ignite explosively in the material accumulated on the wd surface , leading to a brief super - eddington phase during which the wd outer layers are ejected ( e.g. starrfield et al . 1972 , prialnik 1986 ) . ejecta velocities are in the 1002500@xmath32 range , and their metal mass fractions can be as high as @xmath33 ( yaron et al . 2005 , shara et al . 2010 ; hereafter y05 , s10 ) . typical ejecta masses are in the @xmath34@xmath35 range ; but y05 and s10 showed that nova eruptions on slowly - accreting , low - temperature wds can eject up to @xmath36 of gas . ejecta mass estimates in x - ray detected novae ( pietsch et al . 2007 ) somewhat confirm theoretical expectations : ejecta masses of @xmath37 are likely in at least 2 out of 18 novae . wd - hosting binaries are common in gcs ( sigurdsson & phinney 1995 ) ; thus , novae should be frequent in gcs . this disagrees with the low number of novae observed in gcs ; but the discrepancy might be explained by observational biases . we parametrize the nova rate in a gc as @xmath38 , where @xmath39 is the nova rate per unit stellar mass in elliptical galaxies ( della valle et al . 1994 ; mannucci et al . 2005 ; henze et al . 2009 ) , @xmath40 is the gc mass , and @xmath41 is the enhancement of the nova rate in gcs ( compared to ellipticals ) . we expect @xmath42 because of the over - abundance of wd - hosting binaries in gcs . on the other hand , observations of the m31 and m87 gc systems can put upper limits on nova occurrence rates : for m87 , shara et al . ( 2004 ) infer @xmath43 , whereas for m31 we have an upper limit @xmath44 ( henze et al . 2009 ) : assuming @xmath45 , these rates imply @xmath46 . we use the photoionization code cloudy ( version 08.00 , see ferland et . al . 1998 ) to model the optical spectrum resulting from the photoionization of the nova ejecta by the xrs . we represent the nova ejecta as a slab of material with thickness @xmath47 , number density @xmath48 , filling factor @xmath49 , and metallicity @xmath50 , at a distance @xmath51 from the xrs , covering a fraction @xmath52 of the solid angle around it . we note that @xmath52 determines two different geometrical configurations : 1 . if @xmath52 is large ( @xmath53 ) , the xrs must be _ within _ the nova ejecta . in such case , we neglect the offset between the xrs and the centre of the ejecta shell , so that the radius of the ejecta shell is @xmath54 ; then , the time elapsed after the nova explosion is @xmath55 , where @xmath56 , and @xmath57 . 2 . if @xmath52 is small , we assume the xrs to be _ outside _ the ejecta , and estimate @xmath58 from simple geometry : @xmath59 , so that @xmath60 . the xrs spectrum was taken from m07 : we use a disk blackbody spectrum with @xmath61 and 0.210 kev luminosity @xmath62 . we note that m07 reported a large variation ( a drop by a factor of @xmath63 ) in the count rate during a single _ xmm - newton _ observation . such variation is consistent with an increase in the intervening column density , from @xmath64 ( consistent with the galactic value ) to @xmath65 @xmath66 . thus , we assume that the intrinsic x - ray spectrum does not change ; but we experiment with different @xmath65 values . lc parameter & range of values + @xmath67 [ cm ] & 1618.5 , with 0.25 steps + @xmath68 [ cm ] & 15.0 , 15.5 , 16 , 16.5 + @xmath52 & 0.003 , 0.01 , 0.03 , 0.1 , 2/3 + @xmath69 [ @xmath70 & -2 to 10 , with 0.1 steps + @xmath49 & 2/3 + @xmath50 & 0.14 , 0.29 , 0.50 + @xmath71 [ kev ] & 0.22 + @xmath72 [ erg s@xmath1 ] & @xmath74 + @xmath75 [ @xmath76 & 19 , 20 , 21 , 21.5 + @xmath77 [ m@xmath27 ] & @xmath79 + table 1 describes the extent of the explored parameter space . we point out that : 1 . we use the cloudy pre - defined nova abundance set . it represents typical nova ejecta , and corresponds to an o / h number ratio of @xmath80 ( @xmath81 times larger than in the solar mixture ) , and to a metal mass fraction @xmath82 . we obtain mixtures with different metallicities by rescaling the metal abundances of this set by a factor 0.4 ( @xmath83 ) or 2.5 ( @xmath84 ) . the slab mass is @xmath85 f_{\rm cov } f_{\rm fill } [ n/(10 ^ 4\,{\rm cm^{-3}})]$ ] , where @xmath86 is the mass of the h atom . since we expect that @xmath87 ( y05 , s10 ) , ionization models were calculated only when @xmath88 . models were required to agree with two basic observational constraints from the spectrum shown in z081.3 years ) . instead the two spectra described by z07 ( taken in 20042005 , i.e. in the previous @xmath03.9 years ) appear quite different ( fwhm @xmath89@xmath90 , @xmath91 ) . z07 warn that uncertainties amount to a factor of @xmath92 , and the discrepancy might be explained by the ( very different ) observational setups . ] 1 . the flux of the [ oiii]@xmath8 line , @xmath93 . z08 do not quote an error on this measurement : the equivalent width comparison performed by s11 suggests an uncertainty of 2030 per cent ; but the discrepancy with the z07 spectrum suggests a larger amount . furthermore , we account for the approximations of the ionization models by increasing such uncertainty . thus , we consider viable models with @xmath94 within a factor of 3 from the z08 estimate . the absence of any other emission lines in all the spectra : we exclude all the models where the luminosity in any optical line , h@xmath20 , [ oi]@xmath95 , [ oii]@xmath14 , [ nii]@xmath15 , and [ sii]@xmath96 . ] exceeds @xmath97 . about 4 per cent of the models we calculated models , of whom 2144 are viable ( 655 if we consider only models with @xmath98 ) . if we require @xmath99 the number of viable models is 1278 ( 262 if we further impose @xmath98 ) , and requiring @xmath100 further decreases this number to 652 ( 100 of whom have @xmath98 ) . ] satisfy the above criteria . in fig . 1 , we show the distribution of the various parameters ( including the ionization parameter , defined as the number ratio of ionizing photons to h atoms , @xmath101 , where @xmath102 is the number of ionizing photons - i.e. with energy @xmath103 - emitted by the xrs in the unit time ) for these viable models , and how these distributions are influenced by the choice of priors about the models ( e.g. the effects of using different thresholds on the ejecta mass , or imposing a minimum for the nova - xrs distance ) . such distributions show us where a large amount of parameter space is available . it can be seen that : 1 . it is almost impossible to have viable models if @xmath104 ; 2 . thin shells ( @xmath105 ) are favoured , especially when high-@xmath106 models are excluded ; 3 . models where the xrs is inside the nova shell ( i.e. with @xmath107 ) are moderately unlikely , especially when high-@xmath108 models are excluded ; 4 . shell densities are in the @xmath109 range ; but models with high values of @xmath51 show a narrower distribution , favouring the low-@xmath48 part ; 5 . @xmath110 models are favoured ; but it is possible to get reasonable results if @xmath111 or even @xmath83 . the ionization parameter is mostly close to @xmath112 , a fairly typical value for hii regions ; this is somewhat remarkable , as the ionizing spectrum is not produced by hot stars , and most of the radiation energy is in the form of quite energetic ( 0.11 kev ) photons ( rather than photons just above the lyman limit ) . lcccc parameter & model a & model b & model c & model d + @xmath67 [ cm ] & 17.25 & 17.00 & 16.75 & 16.50 + @xmath68 [ cm ] & 15.5 & 15.0 & 15.0 & 15.5 + @xmath52 & 0.03 & 0.03 & 2/3 & 0.10 + @xmath69 [ @xmath70 & 4.7 & 5.5 & 5.0 & 5.5 + @xmath113 & -2.81 & -2.68 & -2.58 & -2.58 + @xmath50 & 0.50 & 0.29 & 0.50 & 0.29 + @xmath114 [ @xmath76 & 21 & 20 & 21.5 & 21.5 + @xmath106 [ 10@xmath115 m@xmath27 ] & 10.6 & 6.7 & 14.9 & 7.0 + @xmath58 [ 10@xmath116 cm ] & 6.2 & 3.5 & 5.6 & 2.0 + @xmath117 [ yr ] & 13.1 & 7.4 & 11.9 & 4.2 + [ oiii ] emission should evolve on a timescale @xmath117 . thus , our scenario can be tested by monitoring the rz 2109 spectrum . we arbitrarily select four viable models ( see table 2 ) for a study of the time evolution of line emission . this was done by assuming that the nova shell keeps expanding with @xmath118 ( cfr . z08 , s11 ) , and properly scaling the relevant parameters ( @xmath119 , @xmath47 and @xmath48 if xrs is outside the expanding shell ; @xmath51 , @xmath47 and @xmath48 otherwise ) . as the evolution of the thickness @xmath47 of the nova shell is unclear , we consider two different cases : ( a ) @xmath120constant ( appropriate for small velocity dispersions within the ejecta ) , and ( b ) @xmath121 ( appropriate for a significant ejecta velocity dispersion ) . we calculated the properties of each of these four models at five times : three past epochs ( april 2005 , december 2007 , and march 2009 , i.e. the times of the last z07 observation , of the z08 observation , and of the last s11 observation , respectively ) , and two future ones ( december 2012 and december 2017 , i.e. 5 and 10 years after the z08 observation ) . the behaviour of the results are summarized in fig . 2 , where we show the time evolution of the ionization parameter , of the ionization state of oxygen , and of the intensities of five lines ( or groups of lines ) . such lines are representative of the main ionization stages and of the most abundant heavy elements in the ejecta , but are _ not _ always the most intense : an exhaustive list of possibly detectable optical and uv lines can be found in the appendix . in 7 out of 8 models the ionization parameter increases with time , so that heavy elements ( such as oxygen ) move to higher ionization states . therefore , high - ionization lines ( such as the ovi lines around @xmath1221035 ) tend to grow stronger , whereas low - ionization lines ( e.g. [ nii]@xmath76548,6583 ) tend to weaken with time , and intermediate - ionization lines ( e.g. the civ doublet at @xmath71549,1551 ) grow to a peak , and then start to decline . the case ( a ) of model c is the main exception : since it assumes that ( i ) the xrs is within the expanding shell , and ( ii ) the shell thickness is constant , the density evolution in the shell ( @xmath123 ) compensates the geometrical dilution of the ionizing flux , resulting in a constant ionization parameter . as a consequence , the ionization state of oxygen ( and of all the other heavy elements ) remains practically constant , and line fluxes tend to decrease very slowly . the four models were required to agree with the december 2007 observations , so their spectra for this epoch are dominated by the [ oiii ] doublet , with @xmath94 within 20 per cent from the z08 value . in addition , [ neiii]@xmath124 and [ oiii]@xmath17 might be strong enough to be detectable : the latter was marginally detected by z08 ; as to [ neiii ] , we note that the z08 spectrum is very noisy at @xmath1254000 , and this line might be confused with other features in that region . all the models ( except model d - see next subsection ) are in quite satisfactory agreement with the 2009 observation , too : this is simply because the 1.3 years interval from december 2007 to march 2009 is a relatively small fraction of @xmath126 . instead we find larger ( and more interesting ) differences if we look at the april 2005 spectrum : models a and c predict that the [ oiii ] emission was as strong ( a ) or stronger ( c ) than in 2007 , and require that the discrepancy among the fluxes reported by z07 and z08 derives from observational factors only . instead , models b and d provide a different explanation : their quite short @xmath117 imply large differences in @xmath58 : in 2005 @xmath58 was only 2/3 ( model b ) or 1/3 ( model d ) of its value in 2007 ; such a difference not only reduces the amount of energy intercepted by the ejecta , but at the same time lowers the ionization parameter , favouring low - ionization states and moving the emission from [ oiii ] to [ oii ] ( or even [ oi ] ) lines . the predictions for 2012 and 2017 offer an opportunity to test these models , with model d predicting that all anomalous emission lines should disappear before 2012 , model b predicting large declines , and models a and c predicting a slower , steady decline . as we already mentioned , such decline is mostly understood in terms of an increasing ionization parameter , so that the o@xmath127 abundance drops . it should be noted that all the above predictions suffer from a number of approximations : 1 . the structure of a nova shell is not uniform . there are large density fluctuations both at early times ( @xmath128 years after the eruption ; see e.g. the images of nova cygni 1992 in chochol et al . 1997 ) and at late times ( @xmath129 years after the eruption ; see e.g. the images of t pyx in shara et al . 2 . the residual nuclear burning on the wd responsible for the nova eruption might produce a significant amount of uv / x - ray photons for a few years ( see henze et al . 2010 ) : furthermore , when @xmath130 , @xmath131 , and even moderate wd luminosities ( say @xmath132 ) are enough for the wd emission to dominate over the one from the xrs . therefore , the predictions for model b in 2005 , for model d in 2007 , and especially for model d in 2005 might suffer from the fact that our models do not consider the wd emission . the cloudy calculations assume steady state , but models with a long ( @xmath1331 yr ) recombination timescale would require a time - dependent approach . for h recombination , @xmath134 ( where @xmath135 is the h recombination coefficient ) , so that models with @xmath136 might suffer from this problem ; however , the recombination timescales of highly ionized metals are shorter ( by factors of 210 ) than @xmath137 , and predictions about their lines should be reliable for @xmath138 . our geometry is extremely simple , and might suffer from assuming that the whole nova shell is at the same distance from the xrs : this is reasonable if the xrs is within the shell ( model c ) , or if @xmath139 , but there are a number of cases ( models a and b at the 2012 and 2017 epochs ; model d at all epochs except 2005 ) where @xmath140 ( so that the ionization parameter in the xrs - approaching part of the shell might be @xmath141 times higher than in the xrs - receding part ) , and the xrs is engulfed by the shell ; we try to adapt to the new situation by switching to a model with the xrs at the centre of the shell ( @xmath142 , @xmath143 ) , but these two set of predictions should be considered extremely uncertain . ] . the timing of the evolution depends upon the exact value of @xmath144 : the adopted 1.5 value is reasonable ( s11 estimates @xmath145 for the broad component of the line profile - see the next subsection ) , but not unique : a lower(higher ) choice of @xmath146 would delay(fasten ) the predicted evolution ( for example , if @xmath147 , our predictions for december 2012 and december 2017 would apply to june 2015 and december 2022 , respectively ) . this is particularly relevant for model d : the [ oiii ] flux if fig . 2 and table a4 evolves too rapidly ( the 2007 flux agrees with the observations , but both the 2005 and the 2009 fluxes are too low ) ; however , the model should not be discarded ( yet ) because these discrepancies can still be reconciled using a value of @xmath144 ( @xmath148 ) that is still in agreement with the s11 measurements . some of these issues might be resolved by adopting more complicated modelling ( e.g. assume a spectrum / temporal evolution for the wd emission , use some density distribution , adopt a more realistic geometry , etc . ) ; however , a more detailed modelling , while providing more reliable predictions , would also require us to introduce a number of additional parameters . this is not only a problem in terms of resources ( such as computing power ) , but would further complicate the interpretation of the results . s11 model the spectral profile of the [ oiii]@xmath149 complex as the superposition of the identical profiles of the two lines of the doublet , each one made of a central , relatively narrow ( fwhm@xmath150 ) gaussian peak , plus a much broader ( fwhm@xmath151 ) ` flat ' component ( possibly the result of a bipolar outflow ) ; notably , the latter component is not really flat : the red side is about 1.4 times brighter than the blue side . as noted by s11 , this is consistent with a nova shell between the xrs and the observer ) isotropic component , plus a fast bipolar one ( @xmath152 ) . nova observations ( e.g. iijima & naito 2011 ) and theoretical models ( see e.g. the ejecta velocity plots in s10 ) support these hypotheses by finding / predicting complex profiles for nova emission lines . ] . if this is the case , the flux difference among the two sides suggests that in 2007 - 2008 @xmath58 was of the order of 10 per cent of the distance between the nova and the xrs : this would favor models with @xmath153 0.01 . furthermore , since @xmath58 increases with time , the asymmetry between the red and blue sides should increase as well ( at least as long as the xrs is outside the nova shell ) . if the nova lies between us and the xrs , the ejecta might even intercept the line of sight to the xrs , increasing the value of @xmath65 . we checked whether such an increase can explain the fading of the xrs seen in the 2004 _ xmm _ observation ( see m2007 ) , and in the subsequent _ swift _ and _ chandra _ observations ( see maccarone et al . the effective x - ray extinction due to isotropic nova ejecta is @xmath154 where @xmath155 is a geometrical factor accounting for the number of times that the photons cross the shell ( 1 if the xrs is within the shell ; 2 if the xrs is outside the shell ) , and for the inclination of the photon direction with respect to the normal to the shell ( note that @xmath155 can be quite larger than 2 if the line of sight to the xrs is almost tangential to the ejecta shell ) . maccarone et al ( 2010 ) suggest that the weakness of the xrs in rz 2109 during the _ chandra _ 2010 observation might be explained by a large increase in the effective @xmath65 in front of the xrs . we used the web version of w3pimms ( http://heasarc.nasa.gov/tools/w3pimms.html ) to estimate that @xmath156 is necessary to explain the _ chandra _ 2010 observations . this is marginally consistent with our scenario : if we set @xmath157 , @xmath84 , @xmath158 , we can obtain such a high @xmath65 for @xmath159 . in turn , this implies that the nova eruption occurred in 20002004 ( for @xmath160 ) , which might be consistent with the timing of the z07 observations ( 20042005 ) . however , the available parameter space is very small : a slight change in the choice of @xmath155 , @xmath50 and @xmath106 can reduce the upper limit on @xmath161 , moving the nova occurrence to an epoch _ after _ the first z07 observation . therefore , we deem quite unlikely that the weakening of the xrs is ( entirely ) due to the additional absorption from the nova shell ; however , if this is the case , the xrs should re - brighten within a few years . we estimate the probability of having a nova shell with properties suitable to explain the rz 2109 observations within a distance @xmath162 from the xrs as @xmath163 where @xmath164 and @xmath41 were introduced in sec . 2 , @xmath165 is the number of stars in a gc , @xmath166 is the average stellar mass in a gc , @xmath167 is the number density of stars at the centre of a gc ( @xmath168@xmath169 ; see e.g. pryor & meylan 1993 ) , @xmath170 is the fraction of novae whose ejecta can produce the observed line emission ( a highly uncertain quantity ; since relatively massive ejecta are needed , we use 0.01 as a reference ) , and @xmath171 v_{\rm ej,3}^{-1}$ ] acts as the visibility timescale for the [ oiii ] line emission . such a probability is of the same order of magnitude as that predicted in other scenarios such as the disruption of a wd by an imbh ( clausen & eracleous 2011 quote a disruption rate of @xmath172 from sigurdsson & rees 1997 ) , and it must be noted that the large uncertainty about several parameters in the above formula might lead to a substantial increase of @xmath173 . we note that , if the xrs is outside the nova shell , a further factor @xmath174 accounting for the possible beaming of the xrs radiation should be introduced into eq . ( 2 ) : therefore , an observational confirmation of this scenario might disfavour the models of the xrs ( e.g. king 2011 ) that assume a beamed emission ( appropriate for this geometry ) should be multiplied by a factor of @xmath175 : if @xmath176 , no viable model of this kind would survive . ] . we used the photoionization code cloudy to check whether the peculiar emission spectra from xrs - hosting gcs ( especially rz 2109 ) can be explained as the result of the photoionization of a nova shell by the radiation emitted by the xrs . we found that this is definitely possible , but requires that the nova ejecta are relatively massive ( @xmath177 ) , and that the xrs - nova distance is quite small ( @xmath5 ) . the combination of these two requirements is quite unlikely ; on the other hand , it can explain observations in a straightforward manner ( although the line profile needs further investigation ) . while not entirely satisfactory , this compares well with the two other proposed scenarios for rz 2109 : ( i ) the disruption of a wd by an imbh ( an event as unlikely as the one we discuss ) explains the line profile and the absence of h lines ( i10 , clausen & eracleous 2011 ) , but appears to be unable to produce the high [ oiii ] luminosity ( porter 2010 , s11 ) ; ( ii ) the ionization of material ejected from the xrs itself , while not suffering from the `` unlikely coincidence '' problem , does not explain either the absence of h emission lines , nor some properties of the line profile ( asymmetry in the flat component , origin of the gaussian component ) . finally , we predict the future evolution of the line emission in the xrs+nova scenario , by looking at both optical and uv lines . we find that in most cases the nova shell tends to move towards a higher ionization state , so that the line flux should move from moderate - ionization lines such as [ oiii]@xmath8 to high - ionization lines ( e.g. the ovi lines in the 10321038wavelength range ) . such predictions might provide a crucial test for discriminating among the various hypotheses described above . we thank the anonymous referee for a careful reading of the manuscript , and for several very valuable suggestions ( e.g. about uv lines ) , and thank a. wolter , l. zampieri , a. bressan , d. fiacconi , s. hachinger , r. porter , and t. maccarone for useful discussions and comments . this research has made use of nasa s astrophysics data system bibliographic services , and of tools provided by the high energy astrophysics science archive research center ( heasarc ; which is provided by nasa s goddard space flight center ) . allende prieto c. , lambert d.l . , asplund m. , 2001 , apj , 556 , l63 allende prieto c. , lambert d.l . , asplund m. , 2002 , apj , 573 , l137 brassington n.j . , et al . , 2010 , apj , 725 , 1805 chochol d. , grygar j. , pribulla t. , komzik r. , hric l. , elkin v. , 1997 , a&a , 318 , 908 clausen d. , eracleous m. , 2011 , apj , 726 , 34 della valle m. , rosino l. , bianchini a. , livio m. , 1994 , a&a , 287 , 403 gehrz r.d . , truran j.w . , williams r.e . , starrfield s. , 1998 , pasp , 110 , 3 ferland g.j . , shields g.a . , 1978 , apj , 226 , 172 ferland g.j . , korista k.t . , verner d.a . , ferguson j.w . , kingdon j.b . , verner e.m . , 1998 , pasp , 110 , 761 grevesse n. , sauval a.j . , 1998 , space science review , 85 , 161 heinke c.o . , 2011 , to appear in the proceedings of the conference `` binary star evolution : mass loss , accretion and mergers '' ( eds . v. kalogera and m. van der sluys , aip conf . ser . ) , mykonos , greece , june 22 - 25 , 2010 , ( arxiv.org/1101.5356 ) holweger h. , 2001 , joint soho / ace workshop `` solar and galactic composition '' ( bern , switzerland , march 69 2001 ) , edited by r.f . wimmer - schweingruber , aipc proceedings , vol . 598 , p. 23 henze m. , et al . , 2009 , a&a , 500 , 769 henze m. , et al . , 2010 , a&a , 523 , 89 iijima t. , naito h. , 2011 , a&a , 526 , a73 irwin j.a . , brink t.g . , bregman j.n . , roberts t.p . , 2010 , apj , 712 , l1 [ i10 ] kalogera v. , henninger m. , ivanova n. , king a.r . , 2004 , apj , 603 , l41 king a. , 2011 , apj , 732 , l28 maccarone t.j . , kundu a. , zepf s.e . , rhode k.l . , 2007 , nature , 445 , 183 [ m07 ] maccarone t.j . , kundu a. , zepf s.e . , rhode k.l . , 2010 , mnras , 409 , l84 maccarone t.j , kundu a. , zepf s.e . , rhode k.l . , 2011 , mnras , 410 , 1655 maccarone t.j . , warner b. , 2011 , mnras , 410 , l32 mannucci f. , della valle m. , panagia n. , cappellaro e. , cresci g. , maiolino r. , petrosian a. , turatto m. , 2005 , a&a , 433 , 807 moore k. , bildsten l. , 2011 , apj , 728 , 81 peacock m.b . , maccarone t.j . , knigge c. , kundu a. , waters c.z . , zepf s.e . , zurek d.r . , mnras , 402 , 803 pietsch w. , et al . , 2007 , a&a , 465 , 375 porter r.l . , 2010 , mnras , 407 , l59 prialnik d. , 1986 , apj , 310 , 222 pryor c. , meylan g. , 1993 , in asp con ference series 50 , structure and dynamics of globular clusters , eds djorgovski s.g . & meylan g. ( san francisco , ca : asp ) , 357 shafter a.w , quimby r.m . , 2007 , apj , 671 , l121 shara m.m . , zurek d.r . , williams r.e . , prialnik d. , gilmozzi r. , moffat a.f.j . , 1997 , aj , 114 , 258 shara m.m . , zurek d.r . , baltz e.a . , lauer t.r . , silk j. , 2004 , apj , 605 , l117 shara m.m . , yaron o. , prialnik d. , kovetz a. , zurek d. , 2010 , apj , 725 , 831 [ s10 ] shih i.c . , kundu a. , maccarone t.j . , zepf s.e . , joseph t.d . , 2010 , apj , 721 , 323 sigurdsson s. , phinney e.s . , 1995 , apjs , 99 , 609 sigurdsson s. , rees m.j . , 1997 , mnras , 284 , 318 starrfield s. , truran j.w . , sparks w.m . , kutter g.s . , 1972 , apj , 176 , 169 steele m.m . , zepf s.e . , kundu a. , maccarone t.j . , rhode k.l . , salzer j.j . , 2011 , apj , in press ( arxiv:1107.4051 ) [ s11 ] yaron o. , prialnik d. , shara m.m . , kovetz a. , 2005 , apj , 623 , 398 [ y05 ] zepf s.e . , maccarone t.j . , bergond g. , kundu a. , rhode k.l . , salzer j.j . , 2007 , apj , 669 , l69 [ z07 ] zepf s.e . , et al . , 2008 , apj , 683 , l139 [ z08 ] in this appendix , we list the detailed predictions about line fluxes at earth , and about the properties of the nova ejecta ( e.g. the ionization state of c , n , o , and ne ) from the four models described in the text , evolved to five past or future epochs : april 2005 , december 2007 , march 2009 ( chosen to coincide with the times of past optical observations ) , december 2012 , and december 2017 . we report all the emission lines in the uv - optical range ( 913@xmath17810000 ) reaching a flux @xmath179 in at least one of the five considered epoch , plus lyman@xmath180 and h@xmath180 . for well - separated multiplets we list only the strongest line ( for example , in the case of the [ oiii]@xmath74959,5007 doublet we only list [ oiii]@xmath1225007 ) , whereas close multiplets are listed as a single line ( for example the [ oii]@xmath73726,3729 doublet is listed as [ oii]@xmath1223727 ) . all the fluxes are in units of @xmath181 , and were obtained assuming a distance @xmath182 . no reddening correction was applied . all the data columns - except the one referring to the z08 observation , which is taken as a reference point - report two set of data : the first was obtained assuming that the shell thickness @xmath183 remains constant at the value used for december 2007 ( implying that the density @xmath48 scales as @xmath184;assumption a ) , whereas the second was obtained assuming that @xmath47 increases linearly with time ( i.e. that @xmath185 remains constant at the value estimated for december 2007 ; in such case the density @xmath48 scales as @xmath186 ; assumption b ) . the two cases are generally separated by a `` / '' , but in the case of ionization states it was necessary to use two separate lines . lccccc line & april 2005 & december 2007 & march 2009 & december 2012 & december 2017 + + niii@xmath187 & 0.020/0.014 & 0.043 & 0.063/0.074 & 0.129/0.094 & 0.077/0.026 + ovi@xmath188@xmath189 & 0/0 & 0.005 & 0.020/0.035 & 0.232/0.446 & 0.654/0.900 + lyman @xmath180 @xmath190 & 0.168/0.182 & 0.103 & 0.081/0.072 & 0.045/0.033 & 0.029/0.018 + ov]@xmath191 & 0/0 & 0.004 & 0.011/0.016 & 0.076/0.131 & 0.170/0.178 + nv@xmath192 & 0.005/0.002 & 0.055 & 0.131/0.189 & 0.667/0.878 & 0.977/0.704 + oiv]@xmath193 & 0.004/0.002 & 0.038 & 0.090/0.130 & 0.460/0.548 & 0.516/0.279 + niv]@xmath194 & 0.012/0.006 & 0.081 & 0.172/0.235 & 0.714/0.770 & 0.647/0.267 + civ@xmath195 & 0.068/0.004 & 0.031 & 0.057/0.071 & 0.162/0.151 & 0.119/0.043 + oiii]@xmath196 & 0.030/0.020 & 0.092 & 0.135/0.147 & 0.168/0.088 & 0.073/0.025 + niii@xmath197 & 0.067/0.046 & 0.196 & 0.286/0.314 & 0.383/0.203 & 0.146/0.037 + ciii]@xmath198 & 0.026/0.022 & 0.059 & 0.078/0.082 & 0.092/0.049 & 0.033/0.007 + @xmath199neiv]@xmath200 & 0.007/0.003 & 0.041 & 0.085/0.114 & 0.326/0.348 & 0.310/0.156 + @xmath199oii]@xmath201 & 0.050/0.100 & 0.014 & 0.011/0.008 & 0.004/0.002 & 0.002/0.001 + @xmath199nev]@xmath202 & 0.003/0.002 & 0.012 & 0.023/0.031 & 0.103/0.153 & 0.178/0.171 + @xmath199oii]@xmath203 & 0.055/0.098 & 0.012 & 0.010/0.008 & 0.005/0.002 & 0.003/0.001 + @xmath199neiii]@xmath204 & 0.166/0.148 & 0.232 & 0.242/0.229 & 0.139/0.058 & 0.041/0.010 + @xmath199oiii]@xmath205 & 0.030/0.023 & 0.058 & 0.071/0.071 & 0.056/0.025 & 0.019/0.006 + @xmath199oiii]@xmath206 & 2.77/2.24 & 3.6 & 3.43/3.14 & 1.48/0.555 & 0.385/0.097 + @xmath199nii]@xmath207 & 0.023/0.053 & 0.005 & 0.004/0.003 & 0.001/0.001 & 0.001/0 + h@xmath180 @xmath208 & 0.014/0.015 & 0.008 & 0.006/0.004 & 0.003/0.002 & 0.002/0.001 + @xmath199nii]@xmath209 & 0.433/0.893 & 0.055 & 0.037/0.029 & 0.012/0.006 & 0.006/0.001 + @xmath199oii]@xmath210 & 0.068/0.140 & 0.019 & 0.015/0.012 & 0.006/0.003 & 0.003/0.001 + + @xmath69 [ @xmath70 & 4.87/4.96 & 4.70 & 4.63/4.59 & 4.45/4.33 & 4.26/4.04 + @xmath211}$ ] & 10.4 & 13.1 & 14.4 & 18.1 & 23.1 + @xmath212}$ ] & 4.9 & 6.2 & 6.8 & 8.6 & 10.9 + @xmath52 & 0.020 & 0.030 & 0.0354 & 0.533 & 0.0831 + @xmath113 & -2.98/-3.07 & -2.81 & -2.74/-2.70 & -2.56/-2.44 & -2.37/-2.15 + c(i@xmath213vii ) & -,11,69,6,13,1,- & -,3,55,11,27,4,0 & -,2,45,13,33,7,0 & -,0,17,12,45,23,2 & -,0,5,6,40,40,9 + & 0,31,56,5,8,0,- & & -,1,40,13,36,9,1 & -,0,8,9,44,33,5 & -,0,1,2,26,50,21 + n(i@xmath213viii ) & 0,25,63,7,4,1,-,- & -,2,70,15,9,4,0,- & -,1,61,20,11,7,1,- & -,0,24,30,15,25,5,0 & -,0,7,21,14,40,16,1 + & 3,48,42,4,2,0,-,- & & -,1,56,22,12,9,1,- & -,0,11,28,15,35,10,1 & -,0,2,9,9,43,32,5 + o(i@xmath213ix ) & 0,31,58,9,2,0,0,-,- & -,2,69,23,4,2,0,-,- & -,1,58,31,6,3,1,-,- & -,0,19,53,13,10,5,0,- & -,0,6,41,20,15,15,2,0 + & 3,52,39,6,1,0,-,-,- & & -,1,53,35,7,4,1,0,- & -,0,8,50,18,13,10,1,- & -,-,2,22,20,17,30,8,0 + ne(i@xmath213xi ) & 0,25,69,5,1,0,-,-,-,-,- & -,1,79,15,3,0,-,-,-,-,- & -,1,70,23,5,1,-,-,-,-,- & -,0,29,49,19,3,0,-,-,-,- & -,0,10,44,36,9,1,0,0,-,- + & 2,39,55,3,1,-,-,-,-,-,- & & -,1,65,27,7,1,0,-,-,-,- & -,0,13,50,30,6,1,0,-,-,- & -,-,3,26,47,18,4,1,1,0,- + lccccc line & april 2005 & december 2007 & march 2009 & december 2012 & december 2017 + + niii@xmath187 & 0.010/0.002 & 0.132 & 0.216/0.220 & 0.115/0.042 & 0.038/0.005 + ovi@xmath188@xmath189 & 0/0 & 0.032 & 0.167/0.285 & 1.06/1.39 & 1.47/0.961 + lyman @xmath180 @xmath190 & 0.523/0.872 & 0.342 & 0.245/0.208 & 0.136/0.094 & 0.095/0.057 + ov]@xmath191 & 0/0 & 0.034 & 0.118/0.173 & 0.371/0.287 & 0.234/0.073 + nv@xmath192 & 0.009/0.002 & 0.252 & 0.744/1.01 & 1.76/1.30 & 1.09/0.301 + oiv]@xmath193 & 0.007/0.001 & 0.181 & 0.495/0.637 & 0.607/0.268 & 0.200/0.030 + niv]@xmath194 & 0.020/0.002 & 0.341 & 0.850/1.07 & 0.982/0.420 & 0.303/0.040 + civ@xmath195 & 0.016/0.002 & 0.127 & 0.228/0.250 & 0.175/0.068 & 0.045/0.005 + oiii]@xmath196 & 0.067/0.003 & 0.306 & 0.347/0.285 & 0.077/0.024 & 0.022/0.002 + niii@xmath197 & 0.140/0.006 & 0.655 & 0.809/0.710 & 0.221/0.065 & 0.056/0.004 + ciii]@xmath198 & 0.074/0.053 & 0.144 & 0.151/0.128 & 0.035/0.009 & 0.006/0 + nii]@xmath214 & 0.205/0.316 & 0.014 & 0.009/0.006 & 0.003/0.001 & 0/0 + cii]@xmath215 & 0.067/0.114 & 0.007 & 0.004/0.003 & 0.001/0 & 0/0 + @xmath199neiv]@xmath200 & 0.005/0 & 0.087 & 0.180/0.206 & 0.132/0.046 & 0.028/0.002 + @xmath199oii]@xmath201 & 0.546/0.798 & 0.024 & 0.011/0.006 & 0.002/0 & 0/0 + @xmath199nev]@xmath202 & 0.004/0 & 0.083 & 0.204/0.256 & 0.303/0.168 & 0.110/0.011 + @xmath199oii]@xmath203 & 0.070/0.076 & 0.004 & 0.002/0.001 & 0.001/0 & 0/0 + @xmath199neiii]@xmath204 & 0.316/0.305 & 0.358 & 0.231/0.156 & 0.017/0.003 & 0.002/0 + @xmath199oiii]@xmath205 & 0.071/0.003 & 0.206 & 0.176/0.129 & 0.024/0.006 & 0.005/0 + @xmath199oiii]@xmath206 & 1.37/0.050 & 4.84 & 3.36/2.33 & 0.368/0.095 & 0.080/0.005 + @xmath199nii]@xmath207 & 0.333/0.500 & 0.013 & 0.006/0.004 & 0.001/0 & 0/0 + @xmath199oi]@xmath216 & 0.336/0.581 & 0 & 0/0 & 0/0 & 0/0 + h@xmath180 @xmath208 & 0.027/0.030 & 0.025 & 0.017/0.014 & 0.008/0.005 & 0.004/0.002 + @xmath199nii]@xmath209 & 0.748/0.812 & 0.027 & 0.013/0.008 & 0.004/0.001 & 0.001/0 + @xmath199oii]@xmath210 & 0.726/1.06 & 0.033 & 0.015/0.009 & 0.002/0 & 0/0 + + @xmath69 [ @xmath70 & 5.83/6.00 & 5.50 & 5.38/5.32 & 5.10/4.90 & 4.82/4.48 + @xmath211}$ ] & 4.7 & 7.4 & 8.7 & 12.4 & 17.4 + @xmath212}$ ] & 2.2 & 3.5 & 4.1 & 5.9 & 8.2 + @xmath52 & 0.0139 & 0.030 & 0.0399 & 0.0759 & 0.1429 + @xmath113 & -3.01/-3.18 & -2.68 & -2.56/-2.50 & -2.28/-2.08 & -2.00/-1.66 + c(i@xmath213vii ) & 1,60,33,2,3,0,- & -,2,41,14,36,6,0 & 0,1,24,14,47,14,1 & -,0,4,7,44,39,6 & -,0,1,2,26,52,19 + & 2,74,24,0,0,-,- & & -,0,18,13,49,19,1 & -,0,1,3,32,50,14 & -,-,0,0,9,46,46 + n(i@xmath213viii ) & 20,63,14,2,1,0,-,- & -,1,65,19,9,5,0,- & -,1,43,29,14,13,1,- & -,0,7,23,18,40,11,1 & -,0,2,8,11,46,29,3 + & 33,66,1,0,-,-,-,- & & -,0,32,31,15,18,2,0 & -,0,3,11,14,48,23,2 & -,-,0,1,3,30,50,15 + o(i@xmath213ix ) & 32,52,13,2,0,-,-,-,- & -,1,62,28,6,2,1,-,- & -,0,36,44,11,6,2,0,- & -,0,4,34,25,18,17,1,- & -,-,1,12,18,22,40,8,0 + & 43,56,0,-,-,-,-,-,- & & -,0,25,49,14,8,4,0,- & -,-,1,16,22,23,33,5,0 & -,-,0,2,6,11,52,26,2 + ne(i@xmath213xi ) & 5,39,52,3,0,-,-,-,-,-,- & -,1,60,28,10,2,0,-,-,-,- & -,0,33,38,23,6,1,0,-,-,- & -,-,3,20,42,27,6,2,1,0,- & -,-,0,5,24,36,19,10,6,0,- + & 8,47,45,0,-,-,-,-,-,-,- & & -,0,22,39,29,9,1,0,-,-,- & -,-,1,8,32,36,14,6,3,0,- & -,-,-,0,4,15,20,24,35,1,0 + lccccc line & april 2005 & december 2007 & march 2009 & december 2012 & december 2017 + + ciii@xmath217 & 0.349/0.308 & 0.284 & 0.247/0.223 & 0.175/0.125 & 0.116/0.069 + niii@xmath187 & 1.140/1.130 & 0.878 & 0.751/0.656 & 0.529/0.348 & 0.359/0.180 + ovi@xmath188@xmath189 & 2.360/1.180 & 2.090 & 1.920/2.270 & 1.380/2.090 & 0.880/1.620 + lyman @xmath180 @xmath190 & 0.332/0.448 & 0.212 & 0.178/0.161 & 0.121/0.088 & 0.078/0.044 + ov]@xmath191 & 0.890/0.505 & 0.717 & 0.631/0.722 & 0.413/0.550 & 0.240/0.331 + nv@xmath192 & 4.930/3.100 & 4.240 & 3.810/4.160 & 2.670/3.070 & 1.690/1.790 + oiv]@xmath193 & 3.850/2.590 & 3.160 & 2.730/2.740 & 1.760/1.570 & 1.030/0.736 + niv]@xmath194 & 6.590/4.400 & 5.400 & 4.640/4.640 & 2.960/2.510 & 1.690/1.060 + civ@xmath195 & 1.940/1.300 & 1.570 & 1.340/1.340 & 0.842/0.694 & 0.467/0.282 + oiii]@xmath196 & 1.150/1.580 & 0.733 & 0.605/0.473 & 0.439/0.230 & 0.321/0.115 + niii@xmath197 & 3.210/4.040 & 2.200 & 1.820/1.450 & 1.280/0.659 & 0.889/0.295 + ciii]@xmath198 & 1.230/1.470 & 0.867 & 0.722/0.581 & 0.497/0.261 & 0.329/0.114 + @xmath199oiii]@xmath218 & 0.113/0.184 & 0.059 & 0.047/0.034 & 0.032/0.015 & 0.022/0.006 + @xmath199neiv]@xmath200 & 1.520/1.000 & 1.460 & 1.340/1.340 & 0.982/0.855 & 0.637/0.447 + @xmath199nev]@xmath202 & 0.865/0.546 & 0.713 & 0.631/0.705 & 0.423/0.536 & 0.248/0.322 + @xmath199neiii]@xmath204 & 0.964/1.620 & 0.528 & 0.425/0.315 & 0.301/0.130 & 0.218/0.055 + @xmath199oiii]@xmath205 & 0.485/0.789 & 0.254 & 0.200/0.149 & 0.136/0.063 & 0.095/0.028 + @xmath199oiii]@xmath206 & 9.350/16.90 & 4.630 & 3.720/2.680 & 2.730/1.150 & 2.050/0.498 + h@xmath180 @xmath208 & 0.023/0.033 & 0.014 & 0.012/0.010 & 0.008/0.006 & 0.005/0.003 + @xmath199nii]@xmath209 & 0.032/0.054 & 0.036 & 0.038/0.032 & 0.048/0.026 & 0.056/0.017 + + @xmath69 [ @xmath70 & 5.19/5.29 & 5.00 & 4.92/4.88 & 4.73/4.60 & 4.52/4.28 + @xmath211}$ ] & 9.2 & 11.9 & 13.2 & 16.9 & 21.9 + @xmath212}$ ] & 4.3 & 5.6 & 6.2 & 8.0 & 10.3 + @xmath52 & 2/3 & 2/3 & 2/3 & 2/3 & 2/3 + @xmath113 & -2.58/-2.68 & -2.58 & -2.57/-2.53 & -2.58/-2.45 & -2.57/-2.33 + c(i@xmath213vii ) & -,1,34,17,37,11,1 & -,1,26,17,41,14,1 & -,1,24,17,42,15,2 & -,2,25,15,42,15,2 & -,2,26,14,41,15,2 + & -,1,50,14,28,6,0 & & -,1,20,16,44,18,2 & -,1,13,12,45,25,4 & -,1,10,8,41,33,8 + n(i@xmath213viii ) & -,0,30,34,12,20,4,0 & -,0,21,35,13,25,5,0 & -,0,20,34,14,26,6,0 & -,1,20,32,14,27,6,1 & -,1,23,29,14,26,7,1 + & -,0,49,27,10,12,2,0 & & -,0,15,33,14,29,8,1 & -,0,10,26,14,36,13,1 & -,0,8,17,12,39,21,3 + o(i@xmath213ix ) & -,0,21,51,12,9,6,1,- & -,0,12,53,15,12,8,1,- & -,0,11,51,15,12,8,1,- & -,0,12,48,16,13,9,1,- & -,1,14,45,16,13,10,2,0 + & -,0,39,42,9,6,3,0,- & & -,0,8,49,17,13,10,2,0 & -,0,6,39,19,15,17,4,0 & -,0,5,28,18,16,24,8,1 + ne(i@xmath213xi ) & -,0,29,48,19,4,1,0,-,-,- & -,0,20,52,23,5,1,0,-,-,- & -,0,18,52,24,5,1,0,0,-,- & -,0,20,49,24,5,1,0,0,-,- & 0,0,23,47,23,5,1,0,0,-,- + & -,0,49,37,12,2,0,0,-,-,- & & -,0,14,51,28,6,1,0,0,-,- & 0,0,10,44,35,9,2,1,0,-,- & -,0,8,36,40,12,3,1,1,0,- + lccccc line & april 2005 & december 2007 & march 2009 & december 2012@xmath219 & december 2017@xmath219 + + ciii@xmath217 & 0/0 & 0.061 & 0.063/0.039 & 0.003/0 & 0.002/0 + niii@xmath187 & 0/0 & 0.213 & 0.208/0.138 & 0.009/0 & 0.005/0 + ovi@xmath188@xmath189 & 0/0 & 0.371 & 1.41/1.83 & 0.714/0.093 & 0.309/0 + lyman @xmath180 @xmath190 & 0.391/0.411 & 0.403 & 0.191/0.145 & 0.040/0.015 & 0.023/0 + ov]@xmath191 & 0/0 & 0.149 & 0.492/0.599 & 0.050/0.003 & 0.021/0 + nv@xmath192 & 0/0 & 0.746 & 2.01/2.27 & 0.155/0.013 & 0.050/0 + oiv]@xmath193 & 0/0 & 0.493 & 1.11/1.09 & 0.028/0.001 & 0.014/0 + niv]@xmath194 & 0/0 & 0.784 & 1.72/1.65 & 0.018/0.001 & 0.009/0 + civ@xmath195 & 0/0 & 0.216 & 0.456/0.391 & 0.006/0.001 & 0.003/0 + @xmath199neiv]@xmath220 & 0/0 & 0.086 & 0.161/0.142 & 0.006/0.001 & 0.003/0 + oiii]@xmath196 & 0/0 & 0.296 & 0.198/0.088 & 0.003/0 & 0.002/0 + niii@xmath197 & 0.001/0 & 0.697 & 0.496/0.251 & 0.004/0 & 0.003/0 + ciii]@xmath198 & 0.008/0.004 & 0.242 & 0.173/0.086 & 0.001/0 & 0.001/0 + nii]@xmath214 & 0.118/0.092 & 0.005 & 0.002/0.001 & 0/0 & 0/0 + cii]@xmath215 & 0.062/0.063 & 0.005 & 0.002/0.001 & 0/0 & 0/0 + @xmath199neiv]@xmath200 & 0/0 & 0.137 & 0.341/0.348 & 0.023/0.003 & 0.013/0 + @xmath199oii]@xmath201 & 0.283/0.200 & 0.007 & 0.002/0.001 & 0/0 & 0/0 + mgii@xmath221 & 0.065/0.031 & 0.004 & 0.001/0 & 0/0 & 0/0 + @xmath199nev]@xmath202 & 0/0 & 0.127 & 0.394/0.473 & 0.059/0.005 & 0.026/0 + @xmath199nei]@xmath222 & 0.056/0.084 & 0 & 0/0 & 0/0 & 0/0 + @xmath199neiii]@xmath204 & 0.079/0.062 & 0.355 & 0.151/0.051 & 0.001/0 & 0.001/0 + @xmath199oiii]@xmath205 & 0/0 & 0.173 & 0.088/0.031 & 0.001/0 & 0/0 + @xmath199oiii]@xmath206 & 0.007/0.002 & 4.30 & 1.73/0.480 & 0.005/0 & 0.003/0 + @xmath199nii]@xmath207 & 0.185/0.168 & 0.005 & 0.001/0.001 & 0/0 & 0/0 + @xmath199oi]@xmath216 & 0.348/.454 & 0 & 0/0 & 0/0 & 0/0 + h@xmath180 @xmath208 & 0.022/0.030 & 0.031 & 0.014/0.010 & 0.002/0.001 & 0.001/0 + @xmath199nii]@xmath209 & 0.244/0.159 & 0.013 & 0.003/0.001 & 0/0 & 0/0 + @xmath199oii]@xmath210 & 0.375/0.264 & 0.010 & 0.003/0.001 & 0/0 & 0/0 + + @xmath69 [ @xmath70 & 6.20/6.55 & 5.50 & 5.30/5.20 & 4.46/4.15 & 4.11/3.63 + @xmath211}$ ] & 1.5 & 4.2 & 5.5 & 9.2 & 14.2 + @xmath212}$ ] & 0.7 & 2.0 & 2.6 & 4.3@xmath219 & 6.7@xmath219 + @xmath52 & 0.02 & 0.10 & 0.160 & 2/3@xmath219 & 2/3@xmath219 + @xmath113 & -3.28/-3.63 & -2.58 & -2.38/-2.28 & -1.76/-1.44 & -1.76/-1.28 + c(i@xmath213vii ) & 41,57,2,-,-,-,- & -,2,65,8,20,5,0 & -,0,20,12,40,22,4 & -,-,0,0,4,30,66 & -,-,0,0,5,31,63 + & 59,40,1,-,-,-,- & & -,0,7,10,42,33,8 & -,-,-,-,1,15,85 & -,-,-,-,-,3,97 + n(i@xmath213viii ) & 74,26,0,-,-,-,-,- & -,1,68,15,7,8,2,0 & -,0,21,27,12,28,10,1 & -,-,0,0,0,9,42,48 & -,-,0,0,0,9,41,48 + & 80,20,-,-,-,-,-,- & & -,-,7,25,12,36,17,2 & -,-,-,-,-,2,23,75 & -,-,-,-,-,0,5,95 + o(i@xmath213ix ) & 70,30,0,-,-,-,-,-,- & -,1,64,23,6,4,3,0,- & -,0,18,40,16,11,12,3,0 & -,-,0,1,1,2,21,49,26 & -,-,0,1,1,2,20,49,26 + & 76,24,-,-,-,-,-,-,- & & -,-,4,39,19,13,18,6,0 & -,-,-,-,0,0,6,38,56 & -,-,-,-,-,-,0,10,90 + ne(i@xmath213xi ) & 61,30,10,-,-,-,-,-,-,-,- & -,1,67,21,10,2,0,0,-,-,- & -,0,20,40,31,8,1,1,0,-,- & -,-,0,3,15,21,14,11,27,9,1 & -,-,1,3,15,20,13,11,28,9,1 + & 71,23,6,-,-,-,-,-,-,-,- & & -,-,6,38,40,12,3,1,1,0,- & -,-,0,0,2,5,5,6,39,35,6 & -,-,-,-,-,-,-,0,8,46,46 +	we study the possibility that the very broad ( @xmath0 1500 km s@xmath1 ) and luminous ( @xmath2 ) [ oiii ] line emission observed in the globular cluster rz 2109 might be explained with the photoionization of nova ejecta by the bright ( @xmath3 ) x - ray source hosted in the same globular cluster . we find that such scenario is plausible and explains most of the features of the rz 2109 spectrum ( line luminosity , absence of h emission lines , peculiar asymmetry of the line profile ) ; on the other hand , it requires the nova ejecta to be relatively massive ( @xmath4 ) , and the nova to be located at a distance @xmath5 from the x - ray source . we also predict the time evolution of the rz 2109 line emission , so that future observations can be used to test this scenario . novae , cataclysmic variables - galaxies : individual ( ngc 4472 ) - galaxies : star clusters : individual ( rz 2109 ) - x - rays : binaries - ism : jets and outflows
we start from the bdg hamiltonian ( eq . ( [ eq : h ] ) of the main text ) and compute the thermodynamic potential @xmath129 at inverse temperature @xmath130 using the quantum partition function @xmath131 . this allows us to extrapolate to the groundstate energy in the limit @xmath132 . following ref . @xcite ( see p. 254 ) , we introduce the real parameter @xmath78 multiplied onto the exchange part of the bdg hamiltonian , @xmath133 written with @xmath134 nambu matrices @xmath135 . from this definition , the change of thermodynamic potential due to the spin lattice can be found as an integral of @xmath136 over @xmath78 : @xmath137.\label{eq : trvg}\end{aligned}\ ] ] here we use the matrix notation @xmath138 and @xmath139 , with the nambu matrix structure being implied . the trace involves a sum over spin lattice positions , @xmath140 , nambu indices and matsubara frequencies . notice that the green function @xmath141 is defined with averaging , @xmath142 , with respect to @xmath143 . in terms of the bare green function , @xmath144 , the dressed green function matrix is obtained by summing the dyson series @xmath145 we note that the 4-spinor notation , @xmath146 exhibits the usual particle - hole redundancy since @xmath147 . the energy overcounting is avoided by the factor of 1/2 multiplying the hamiltonian ( [ eq : hlam ] ) . this factor of 1/2 also compensates a factor of 2 appearing at each vertex in the dyson series ( [ eq : dyson ] ) due to the fact that the perturbation series for the green function involves both anomalous @xmath148 as well as normal @xmath149 nambu contractions . since , however , the two are related by multiplication with @xmath150 from the right , their contributions are equivalent and give rise to a compensating factor of 2 at each vertex @xmath151 . as a result , one arrives at the dyson series containing only nambu normal green functions . inserting into , the @xmath78-integration can be carried out order by order replacing @xmath152 by @xmath153 , and the infinite series can be recognized as a logarithm : @xmath154.\label{eq : de}\end{aligned}\ ] ] since we are interested in the magnetic ordering arising from the induced interaction between different spins , we focus on the part of the energy shift which depends on the spin lattice spacing , @xmath18 . the only @xmath18-independent terms in the dyson equation are those which only involve repeated scattering from a single magnetic atom . to facilitate this separation we split the bare green function matrix into a sum of local ( @xmath155 ) , and non - local ( @xmath156 ) terms : @xmath157 the @xmath18 dependent part of is therefore found by subtracting the local part , i.e. @xmath158- { \rm tr}\ln\left[1-{\mathbb j}{\mathbb g}_{loc}^{0}\right]\right ) = -\frac{1}{2\beta}{\rm tr}\ln\left[1-{\mathbb t}\tilde{\mathbb g}^{0}\right],\end{aligned}\ ] ] where the local @xmath49-matrix has been introduced as @xmath159 to investigate the magnetic structure , we use a variational approach in which we minimize the electronic ground state energy with respect to the magnetic ordering profile . we consider a planar spiral order ansatz whose spiral axis of rotation is defined along @xmath125 @xmath160 which interpolates between ferromagnetic ( f ) @xmath161 and anti - ferromagnetic ( af ) @xmath162 order @xcite . within the spiral ansatz , the @xmath16 dependence of @xmath163 can be conveniently written as @xmath164 , where @xmath165 . since @xmath166 , the green function commutes with @xmath167 , and we may rewrite eq . ( [ eq : f ] ) as @xmath168 the local green function and the @xmath49-matrix are given by @xmath169 as noted in the main text the determinant of @xmath49 exhibits a pole on the real energy axis @xmath170 where @xmath51 is defined in eq . . for @xmath171 , with debye frequency @xmath172 representing the high - frequency cut - off in the bcs green function , the non - local green function takes the form @xmath173,\end{aligned}\ ] ] which applies to propagation along the spin chain , provided @xmath174 . the matrix @xmath175 in eq . ( [ eq : m ] ) exhibits translational invariance , allowing it to be diagonalized using a set of bloch states labeled by quasi - momentum @xmath176 . in this basis , the operator trace over spin lattice positions in eq . ( [ eq : f ] ) becomes an integral over the brillouin zone . focusing on the low - temperature limit @xmath177 , we replace the matsubara summation by a frequency integral to obtain the variational groundstate energy for a chain of @xmath178 spins @xmath179.\ ] ] in eq . ( [ eq : e ] ) we introduced the fourier transform of the non - local part of the green function @xmath180 the remaining trace over the @xmath181 spin / particle - hole space in eq . ( [ eq : e ] ) can be performed explicitly . the groundstate magnetic order is characterized by the wavevector @xmath68 which minimizes eq . ( [ eq : e ] ) . the minimization procedure is carried out numerically , with the main results for the magnetic order of the spin chain displayed in the exact phase diagram of fig . [ fig : phase - diagram ] of the main text . we note that in the dilute limit @xmath182 we have @xmath183 and thus the off - diagonal components of @xmath184 are small . this is because @xmath185 rapidly falls off as @xmath186 and thereby provides a controlled expansion for @xmath187 . we may then approximate the energy by expanding ( [ eq : f ] ) to second order in @xmath188 : @xmath189.\label{eq : e - pert}\ ] ] carrying out the remaining trace and frequency integral leads to eq . ( [ eq : j - ex ] ) of the main text in the limit @xmath190 . for @xmath191 the terms in square brackets in eq . ( [ eq : j - ex ] ) are modified by powers of @xmath192 , but the entire expression @xmath193 becomes exponentially suppressed in this limit and we do not consider it further . as discussed in the main text , the perturbative expansion leading to the eq . ( [ eq : e - pert ] ) fails for @xmath99 due to strong ysr - cooper pair hybridization . formally , the @xmath194 singularity in eq . ( [ eq : j - ex ] ) is regularized by keeping all terms under the logarithm in eq . ( [ eq : f ] ) . in that case one must understand the structure of the term under the logarithm in eq . ( [ eq : e ] ) , which encodes the subgap ysr spectrum as well as its topological properties . the dispersion relation for the chain of ysr states is determined by having zero determinant under the logarithm in eq . ( [ eq : e ] ) . the sub - gap excitations induced by the spin lattice thus correspond to roots of the equation @xmath195,\end{aligned}\ ] ] where we have pulled out a factor of @xmath196 for convenience , since it has no sub - gap zeroes . one may also derive eq . ( [ eq : spec ] ) from the dyson series for the exact electron green function ( see also ref . @xcite ) , given by @xmath197 where , within the spiral ansatz of eq . ( [ eq : spans ] ) , we have @xmath198 . multiplying eq . ( [ eq : spec1 ] ) on the left by @xmath199 and on the right by @xmath200 we obtain @xmath201 in momentum space eq . ( [ eq : spec2 ] ) becomes , in the notation of the main text , @xmath202 while the inverse of eq . ( [ eq : spec3 ] ) may be written as @xmath203 splitting the bare green function into non - local , and local parts @xmath204 with @xmath205 defined in eq . , the local term and the self - energy may be combined into the @xmath49-matrix whereby @xmath206 after analytical continuation , @xmath207 , the poles of @xmath208 are found as zeroes of the determinant of @xmath209 . the subgap solutions of eq . ( [ eq : app - g ] ) occur solely from the zero determinant of the term in parentheses , leading to the same condition as eq . ( [ eq : spec ] ) , which may be rewritten as @xmath210,\end{aligned}\ ] ] with @xmath211 and @xmath212 here the functions @xmath213 are defined as @xmath214.\end{aligned}\ ] ] for a given @xmath215 , eq . has four roots placed symmetrically around zero , two of which lie in the continuum above the gap . the sub - gap excitation energy @xmath216 generally takes a rather complicated form , which can be simplified by expanding to second order in @xmath217 and @xmath218 , to arrive at : @xmath219 where @xmath220 and @xmath221 . the identical spectrum eq . ( [ eq : spec1 ] ) was found in ref . we may now use the more general result ( [ eq : ginvdet ] ) to investigate the topological properties of the ysr chain . one may determine the topological classification by studying the exact inverse green function at zero frequency , which defines a ` topological hamiltonian ' @xcite for the sub - gap electron degrees of freedom that propagate along the spin chain . as we show in sec . [ sec : top1 ] , the precise form of the topological hamiltonian is somewhat arbitrary and thus allows one to choose the most convenient representation . to that end , we define the topological hamiltonian as the zero - frequency limit of the @xmath134 matrix used in eq . @xmath222 with coefficients defined as the zero - frequency limit of eq . . we have written @xmath223 in a suggestive form to make the analogy with the nanowire hamiltonian of refs . @xcite . in this case , one would interpret @xmath224 as an effective zeeman field , @xmath225 as singlet / triplet pairing , @xmath226 as kinetic energy and @xmath227 as spin - orbit coupling . as discussed in refs . @xcite , the topological class of @xmath228 is bdi due to hidden time - reversal , and chiral symmetries , and is therefore characterized by a @xmath103 topological invariant . this can be shown by noting that , in addition to the usual particle - hole symmetry @xmath229 where @xmath230 and @xmath231 denotes complex conjugation , there is an effective ( anti - unitary ) time - reversal operator @xmath232 such that @xmath233 . the operator @xmath234 satisfies @xmath235 and is distinct from the canonical time - reversal operator @xmath236 , @xmath237 , which does not commute with @xmath228 in the presence of the zeeman field @xmath224 . the existence of @xmath234 allows one to define a chiral operator @xmath238 , @xmath239 , that anti - commutes with @xmath228 : @xmath240 . this implies that @xmath223 can be block off - diagonalized in the eigenbasis of @xmath241 , a representation that may be achieved with the unitary transformation @xmath242 : the @xmath103 topological invariant measures the degree of the map from the brillouin zone to the u(1 ) phase of the complex number @xmath244 , @xmath245 , and is therefore represented by a winding number @xmath246 . from eq . ( [ eq : a ] ) we find @xmath247 with @xmath248 the value of @xmath246 is quantized and can only change when @xmath249 becomes ill - defined . this occurs when @xmath250 , and coincides with the closing of the spectral gap @xmath251 , see eq . ( [ eq : au ] ) . since the function @xmath252 vanishes at the time - reversal symmetric points @xmath253 , the location of the gap - closing transitions simplify in this case to give @xmath254 where @xmath255 with @xmath256 , respectively . in the limit @xmath182 we find @xmath257 which is equivalent to the gap closing transition at @xmath258 in the ysr band for @xmath182 ( @xmath259 ) . furthermore , in the limit @xmath64 this leads to @xmath260 see fig . [ fig : phase - diagram ] . other gap closing transitions for @xmath261 are not precluded and do occur . however , their locations in the phase diagram are difficult to determine analytically for arbitrary @xmath262 . an example of such behavior was discussed in the main text in the limit @xmath182 and was shown to lead to the gapless ferromagnetic line ( [ eq : f - line ] ) . in this section we show that the definition of the hamiltonian ( [ eq : htop ] ) leads to the same topological classification as the hamiltonian @xmath263 defined by the prescription of ref . @xcite to be the negative inverse green function at zero - frequency . according to eq . this gives @xmath264 with the ( zero ) frequency variable suppressed . we may write @xmath265 ( cf . ( [ eq : h_t ] ) ) , with @xmath266 and @xmath267 . thus , if @xmath268 anti - commute with the chiral operator @xmath269 , then @xmath270 has the same chiral symmetry as @xmath228 . since one readily sees @xmath271 , we need only prove that @xmath272 . since @xmath273 clearly anti - commutes with @xmath274 , one may simply take the inverse of the equation @xmath275 to prove that @xmath276 also anti - commutes with @xmath274 . here we have utilized the fact that @xmath276 is an invertible matrix for every @xmath215 due to the bare electron spectrum being gapped by superconducting correlations ( the zero eigenvalue may occur only in @xmath228 ) . in the eigenbasis of @xmath241 , the matrices @xmath277 also become purely off - diagonal , and the off - diagonal element of @xmath278 is therefore of the form @xmath279 , where @xmath224 and @xmath280 are the off - diagonal components of @xmath276 and @xmath281 , respectively . the winding number @xmath246 associated with phase of the complex number @xmath282 is unchanged because the determinants of @xmath283 never vanish , and hence @xmath246 may only change when @xmath250 . thus , the topological classifications of @xmath263 and @xmath228 are equivalent . 99 ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty https://global.oup.com/academic/product/many-body-quantum-theory-in-condensed-matter-physics-9780198566335?cc=dk&lang=en&#[__ ] ( , , ) link:\doibase 10.1103/physrevb.91.064502 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.155420 [ * * , ( ) ] http://stacks.iop.org/0953-8984/25/i=15/a=155601 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.077001 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.177002 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.150408 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.245438 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.115109 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.060507 [ * * , ( ) ] http://stacks.iop.org/1367-2630/17/i=2/a=023051 [ * * , ( ) ]	we study a chain of magnetic moments exchange coupled to a conventional three dimensional superconductor . in the normal state the chain orders into a collinear configuration , while in the superconducting phase we find that ferromagnetism is unstable to the formation of a magnetic spiral state . beyond weak exchange coupling the spiral wavevector greatly exceeds the inverse superconducting coherence length as a result of the strong spin - spin interaction mediated through the subgap band of yu - shiba - rusinov states . moreover , the simple spin - spin exchange description breaks down as the subgap band crosses the fermi energy , wherein the spiral phase becomes stabilized by the spontaneous opening of a @xmath0wave superconducting gap within the band . this leads to the possibility of electron - driven topological superconductivity with majorana boundary modes using magnetic atoms on superconducting surfaces . the prospect of performing topological quantum computation @xcite has stimulated intense investigations into condensed - matter systems harboring majorana bound states @xcite . one potential platform involves magnetic atoms arranged in a regular lattice on an @xmath1wave superconducting substrate @xcite . this system has received a renewed interest due to recent scanning tunneling microscopy ( stm ) data possibly supporting the existence of majorana bound states in a self - assembled one - dimensional ( 1d ) array of atomically - spaced fe atoms on the surface of superconducting pb @xcite . we focus on the case of an stm - assembled magnetic atom chain whose spacing is several substrate lattice sites @xcite , where the overlap of atomic wavefunctions is negligible . in that case each atom acts as an isolated magnetic moment giving rise to localized , sub - gap yu - shiba - rusinov ( ysr ) states @xcite within the superconductor . overlap between ysr states leads to the formation of an effectively spinless sub - gap band that can undergo a topological superconducting transition controlled by the magnetic ordering of the magnetic atoms ( spin chain ) @xcite , see fig . [ fig : schematic ] . on the other hand , the ordering of the spin chain is dictated by the indirect exchange mediated by electrons in the superconductor , allowing the possibility of an electron - driven self - organized " topological superconducting phase . wave pairing due to hybridization with singlet cooper pairs in the host . this hybridization can drive the spin chain into a spiral state that , in turn , induces spontaneous topological superconductivity within the ysr chain . ] although many important aspects of spin chains on superconductors have been theoretically studied by several authors @xcite ( see also @xcite ) , a generic conceptual account of the collective ordering mechanism of the magnetic and electronic degrees of freedom in a bulk ( @xmath2 ) superconductor remains elusive ( the 1d case is special due to @xmath3 nesting and was studied in @xcite , see also @xcite ) . in this paper , we establish the existence of spiral magnetic order and self - organized topological superconductivity in the case of a magnetic adatom chain on a 3d substrate using an exactly solvable minimal model . we provide the conceptual framework to describe the ordering mechanism in a general setting , which we find to be entirely distinct from the @xmath3 mechanism that can occur for a 1d substrate @xcite . in particular , we show that ferromagnetism of the spin chain is unstable to the formation of a spiral state due to the presence of singlet superconductivity in the host . the corresponding spiral wavevector exhibits a nonmonotonic dependence on the spin - lattice exchange coupling to the superconductor . the wavevector grows rapidly as the exchange increases from zero due to a strong indirect spin interaction mediated by ysr states , and decreases after the ysr band crosses the fermi energy due to the double - exchange mechanism . in the latter regime the ysr chain exhibits self - sustained @xmath0wave superconductivity with the possibility of majorana boundary modes , depending on the lattice spacing . our main results are summarized in fig . [ fig : phase - diagram ] , with details related to various technical calculations delegated to the supplemental material @xcite . a system of magnetic atoms exchange coupled to a 3d superconductor may be described by the following bogoliubov - de gennes hamiltonian @xmath4 where @xmath5 ( @xmath6 , @xmath7 and @xmath8 are the electron momentum , position and mass , respectively , while @xmath9 is the fermi momentum ) , @xmath10 is the 4-component nambu spinor written in terms of electron annihilation ( creation ) operators @xmath11 with spin projection @xmath12 . here @xmath13 and @xmath14 are , respectively , pauli matrices acting in the spin and particle - hole spaces while @xmath15 is the sum of spins @xmath16 at sites @xmath17 with lattice spacing @xmath18 and @xmath19 is the exchange interaction constant . we provide the exact phase diagram of eq . ( [ eq : h ] ) within the classical spin approximation @xcite assuming a planar spiral ansatz , @xmath20 , where the spiral wavevector @xmath21 is treated as a variational parameter determined by minimizing the total electronic energy of the system , @xmath22 . the main results of this procedure are shown in fig . [ fig : phase - diagram ] , with technical details provided in the supplemental material @xcite . below , we focus on extracting only the most important and physically relevant ordering mechanisms , which we verify against the exact results . _ magnetic atom order _ @xmath23 the spin - chain magnetic order is easiest to determine in the limit of small exchange coupling , where one may integrate out electrons to second order in @xmath24 to arrive at the effective spin hamiltonian @xmath25 the electron - induced spin - spin exchange coupling in eq . ( [ eq : hspin ] ) is @xmath26 @xcite @xmath27,\ ] ] where @xmath28 is the fermi velocity and @xmath29 is the coherence length of the superconductor . we note that solving for the superconducting order parameter self - consistently is not important in the dilute atom regime ( @xmath30 ) since the ysr energy and electron propagator between spins are hardly modified @xcite . we do not consider the dense atom limit @xmath31 , where a collective depletion of @xmath32 near the chain should be taken into account . the first term in square brackets in eq . is the familiar rudermann - kittel - kasuya - yosida ( rkky ) interaction @xcite , while the second , strictly antiferromagnetic , term arises from the anomalous component of the electronic spin susceptibility and disfavors the pair - breaking effect of a polarized exchange field . the spin - chain groundstate energy is given by the minimum fourier component of the exchange interaction , @xmath33 . in the normal state ( @xmath34 ) , one finds from eq . ( [ eq : jr ] ) a ferromagnet ( @xmath35 ) in the range @xmath36 with integer @xmath37 and an antiferromagnet ( @xmath38 ) otherwise . we emphasize that the @xmath39 spiral phase that may arise when the electron gas is also one - dimensional @xcite does not occur here ; the spin chain is always in a collinear state . however , superconducting correlations in a 3d electron gas generally make ferromagnetism unstable towards spiral formation due to the long - range superconducting correction in eq . ( [ eq : jr ] ) ( in a 1d electron gas the @xmath39 order is hardly modified by superconductivity @xcite ) . indeed , for @xmath40 the minimum fourier component , @xmath41 , is shifted to the wavevector @xmath42 . this magnetic instability is similar to the anderson - suhl transition in 2d and 3d spin lattices @xcite and results from the compromise between the shorter - range ferromagnetic rkky interaction and the longer - range antiferromagnetic interaction . the spiral formation is stable and strongly reinforced by the inclusion of higher order terms beyond the born approximation in eq . ( [ eq : jr ] ) . the crucial effect responsible for this behavior is the formation and overlap of the subgap ysr bound states induced by the magnetic atoms . formally , these subgap states are encoded in the single magnetic atom @xmath43matrices @xmath44 where @xmath45 is the normal state density of states at the fermi energy . the presence of the bound state is signified by the subgap pole of @xmath46 on the real energy axis @xmath47 , @xmath48 to capture the modification of eq . ( [ eq : jr ] ) due to the presence of the ysr bound states , one may replace the bare magnetic atom potential with its @xmath43matrix , eq . ( [ eq : t - mat ] ) . computing the exchange coupling to second order in @xmath49 leads to @xcite @xmath50\right],\label{eq : f-2}\end{aligned}\ ] ] where @xmath19 is expressed through @xmath51 using eq . ( [ eq : epsilondef ] ) . while in the limit of small @xmath19 , i.e. @xmath52 , we recover the perturbative result , we find for small @xmath51 an enhanced antiferromagnetic component @xmath53 due to the exchange coupling mediated via the ysr states @xcite . this effect strongly increases the spiral wavevector of the spin chain . indeed , in the range @xmath54 for integer @xmath37 , one finds @xmath55 , with the groundstate wavevector @xmath56 rapidly growing as @xmath51 approaches zero . this behavior is consistent with the exact numerical result , shown in fig . [ fig : phase - diagram]b , as long as @xmath57 is not too small . the apparent divergence of eq . as @xmath58 signifies its inapplicability for small enough @xmath51 , since the effective exchange energy can not physically exceed the splitting between ysr energies caused by their finite overlap . this suggests that below a characteristic value of @xmath51 , a crossover ( for two spins ) to a new regime occurs . to see this we follow @xcite and consider two ysr states hybridizing with a cooper pair from the superconducting condensate this process can be described by the matrix element @xmath59 representing the overlap between ysr states , see fig . [ fig : schematic ] . the energy difference between the ground state and the excited state with ysr states occupied / unoccupied is @xmath60 . as long as @xmath61 , the change of the groundstate energy can be computed perturbatively , @xmath62 , leading to the second term in the brackets of eq . ( [ eq : j - ex ] ) . thus , for the spin chain , eq . is valid down to the scale @xmath63 ( assuming @xmath64 ) , at which point @xmath65 . when the fermi energy crosses the ysr band @xmath66 , the perturbation theory in @xmath67 fails since cooper pairs in the bulk strongly hybridize with the ysr states . this signals a fundamental change of the groundstate , and corresponds to a @xmath0wave superconducting phase transition in the ysr chain which harbors majorana boundary states ( provided spiral order remains stable ) . we now wish to show that in the regime of strong hybridization ( where the effective spin model eqs . , is inapplicable ) ferromagnetism remains unstable to spiral formation ( i.e. the topological superconductivity is stable ) . the complication is that a new competing tendency arises once the fermi energy lies in the ysr band : the gain of kinetic energy associated with electrons hopping along the ysr chain is maximized for a ferromagnetic arrangement of the spin chain since this maximizes the ysr bandwidth , analogous to the double - exchange mechanism . however , the presence of a spiral leads to a superconducting gap in the ysr band , leading to a concomitant gain of the corresponding condensation energy , depicted in fig . [ fig : bdg ] . at small @xmath68 , it is easy to see that the energy gain from the opening of the superconducting gap @xmath69 is always larger than the loss of kinetic ( or rkky ) energy @xmath70 , where the coefficient of the logarithm is proportional to the square of the ysr gap at the ysr fermi points @xmath71 ( see fig . [ fig : bdg ] ) expanded for small @xmath68 , @xmath72 ( @xmath73 for @xmath74 since cooper pairs can not hybridize with a ferromagnetic chain ) . qualitatively , we thus find that ferromagnetism is generically unstable to the formation of spiral phase with wavevector @xmath75 , where the parameter @xmath76 entering the exponent depends on @xmath77 , and is provided below eq . . the case @xmath35 can only occur if the gap at the ysr fermi points , and thus @xmath78 , vanishes for @xmath79 . we study this possibility and the quantitative variation of @xmath68 on @xmath51 by evaluating the ysr contribution to the total energy numerically ( the above gap rkky contribution , first term in the brackets of eq . , only weakly depends on @xmath80 here ) @xmath81 here the hopping @xmath82 and @xmath0wave pairing @xmath83 energies are the same as those derived in ref . @xcite , up to the energy shift @xmath51 which enters eq . ( [ eq : f - ysr ] ) only as an effective chemical potential for the ysr band . the precise forms of @xmath84 are provided in the supplemental material @xcite , but are not essential in the following ; their salient features are sketched in fig . [ fig : bdg ] . in particular , one sees that @xmath85 indeed may have extra nodes away from @xmath86 due to the long - range nature of the pairing ( i.e. higher odd harmonics beyond @xmath87 are relevant ) . generally we find that when @xmath88 have simultaneous zeroes double - exchange wins and the system exhibits a gapless electronic spectrum and ferromagnetism in the spin chain . remarkably , the exact location of these gapless critical lines in the phase diagram of fig . [ fig : phase - diagram]a can be extracted : @xmath89.\ ] ] in the spiral regions of fig . [ fig : phase - diagram ] , the magnetic order in the close vicinity of eq . ( [ eq : f - line ] ) is exponentially suppressed @xmath90 where @xmath91 . away from this line @xmath68 grows to the scale @xmath65 at @xmath92 , beyond which it crosses over to the @xmath43matrix dependence @xmath93 , as shown in fig . [ fig : phase - diagram]b for the case @xmath94 ( @xmath95 ) . this leads to the non - monotonic behavior of @xmath68 on the ysr energy @xmath51 , with a peak around @xmath96 of order @xmath97 . the magnetic order determined from minimizing the electronic energy of eq . ( [ eq : h ] ) , shown in fig . [ fig : phase - diagram]b , indeed exhibits this crossover behavior and is well - described by the @xmath43matrix approximation when @xmath98 and the ysr approximation when @xmath99 . of the ysr band , eq . ( [ eq : f - ysr ] ) for @xmath100 a spectral gap opens and the system gains the condensation energy , corresponding to the gray area between the full and dashed lines for momenta near @xmath71 . this occurs only if the pairing at these points is non - zero : @xmath101 . when the zeroes of @xmath102 and @xmath85 ( depicted by dots ) coincide , the energy ( [ eq : f - ysr ] ) is minimized at @xmath35 , corresponding to a gapless electron system and a ferromagnetic spin chain . this phase occurs along lines in the phase diagram given by eq . ( [ eq : f - line ] ) . ] we now briefly discuss the topological properties of the model in eq . ( [ eq : h ] ) , delegating a more detailed account to the supplemental material . as discussed in refs . @xcite , the topological class of the ysr chain in the planar spiral configuration is bdi due to hidden time - reversal and chiral symmetries . different topological phases are therefore distinguished by a @xmath103 topological invariant ( winding number ) , which can only change when the gap in the ysr spectrum closes . this occurs when the ysr band crosses the fermi energy or when the system is tuned across the gapless ferromagnetic line eq . ( [ eq : f - line ] ) , leading to the topologically non - trivial and distinct ts@xmath104 phases with winding numbers @xmath105 ( for a fixed rotation axis of the spiral ) , shown in fig . [ fig : phase - diagram]a . as a result , the boundary between ts@xmath106 and ts@xmath107 systems will host a pair of zero - energy majorana bound states protected by chiral symmetry , whereas the boundary of either system with vacuum will only host a single majorana bound state . the presence of two majorana bound states also occurs for a single system in either of the ts@xmath104 phases if there is a @xmath108 domain wall @xcite . _ experimental outlook _ @xmath23 the ysr topological gap , @xmath109 , obtains a broad maximum as a function of magnetic atom spacing ( i.e. from one period @xmath37 of the topologically non - trivial part of the phase diagram to the next , @xmath110 ) . this is because if the magnetic atoms are spaced too closely ( @xmath111 ) @xmath112 becomes small , while if the atoms are too dilute ( @xmath113 ) the overall energy scale @xmath114 is diminished . the maximum thus occurs for an optimal spacing @xmath115 determined by @xmath116 , giving @xmath117 . at this spacing , for @xmath118 , we find @xmath119 , assuming @xmath120 k ( by comparison for fe on pb experiments show @xmath121 @xcite ) . the influence of anisotropy and temperature on the topological phase are expected to be unimportant as long as the associated energy scale remains below @xmath109 . the magnetic order can also be destroyed by thermal fluctuations , but such effects will be quenched either by finite chain size or anisotropy @xcite . for the case of single - ion anisotropy , described by the hamiltonian @xmath122 , we find that when @xmath123 ( easy - plane ) the spiral is merely fixed to lie in the @xmath124-plane ( i.e. the spiral axis lies along @xmath125 ) , while for @xmath126 the spiral axis lies in the @xmath127plane and is orthogonal to the @xmath125 axis . for sufficiently strong anisotropy @xmath128ev we expect the groundstate to become a topologically trivial , collinear magnetic phase . this restricts the self - organized topological phase to easy - plane systems or those with sufficiently weak easy - axis anisotropy . 60ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty \doibase http://dx.doi.org/10.1016/s0003-4916(02)00018-0 [ * * , ( ) ] http://stacks.iop.org/0034-4885/75/i=7/a=076501 [ * * , ( ) ] link:\doibase 10.1146/annurev - conmatphys-030212 - 184337 [ * * , ( ) ] http://stacks.iop.org/0268-1242/27/i=12/a=124003 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.096407 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.161408 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.077001 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.177002 [ * * , ( ) ] link:\doibase 10.1126/science.1222360 [ * * , ( ) ] http://dx.doi.org/10.1038/nphys2479 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.241401 [ * * , ( ) ] link:\doibase 10.1021/nl303758w [ * * , ( ) ] , link:\doibase 10.1103/physrevlett.110.126406 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.020407 [ * * , ( ) ] link:\doibase 10.1126/science.1259327 [ * * , ( ) ] http://journals.aps.org/prl/abstract/10.1103/physrevlett.115.197204 [ * * , ( ) ] http://arxiv.org/abs/1505.06078 [ ] link:\doibase 10.1103/physrevb.90.235433 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.106801 [ * * , ( ) ] link:\doibase 10.7498/aps.21.75 [ * * , ( ) ] link:\doibase 10.1143/ptp.40.435 [ * * , ( ) ] http://adsabs.harvard.edu/abs/1969jetpl...9...85r [ * * , ( ) ] , link:\doibase 10.1103/revmodphys.78.373 [ * * , ( ) ] link:\doibase 10.1126/science.275.5307.1767 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.83.176 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.155420 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.180505 [ * * , ( ) ] http://dx.doi.org/10.1038/nphys2299 [ * * , ( ) ] http://dx.doi.org/10.1038/nphys1915 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.111.186805 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.111.147202 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.111.206802 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.060401 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.060507 [ * * , ( ) ] http://stacks.iop.org/1367-2630/17/i=2/a=023051 [ * * , ( ) ] link:\doibase 10.1103/physrevb.91.064505 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.085124 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.247205 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.195442 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.144505 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.020503 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1007/s002570050313 [ * * , ( ) ] link:\doibase 10.1103/physrevb.66.064526 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.087202 [ * * , ( ) ] link:\doibase 10.1103/physrevb.56.11213 [ * * , ( ) ] link:\doibase 10.1103/physrev.96.99 [ * * , ( ) ] link:\doibase 10.1143/ptp.16.45 [ * * , ( ) ] link:\doibase 10.1103/physrev.106.893 [ * * , ( ) ] link:\doibase 10.1103/physrev.116.898 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.82.3045 [ * * , ( ) ] link:\doibase 10.1103/physrevb.91.064502 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.150408 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.245438 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.115109 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.100506 [ * * , ( ) ] * self - organized topological superconductivity in a yu - shiba - rusinov chain * * supplemental material * in this supplemental material we compute the exact groundstate energy of the bdg hamiltonian as a function of the spiral wavevector in sec . [ sec : g - energy ] and derive the formula determining the exact subgap ysr spectrum in sec . [ sec : ysr - spec ] , which allows us to classify and determine the exact topological phase boundaries in sec . [ sec : top ] .
the born - oppenheimer approximation is a method which deals with a coupled system of heavy and light objects@xcite . from atomic physics to nuclear physics , it has played an important role . moreover , the importance of this method gets greater as the necessity of the adiabatic control of quantum states arises in the quantum information science@xcite . nevertheless , the criteria of the validity of the method is not well defined yet , even still controversial in some aspects@xcite . therefore , more theoretical work is needed through applications to specific models . in this paper , we focused on the effect of chaotic dynamics in the born - oppenheimer approximation method . since the born - oppenheimer approximation basically deals with coupled systems , there is always the possibility of non - integrability generating chaotic dynamics@xcite . for this study , we choose the fermi accelerator as a model of hamiltonian chaos@xcite . originally , the fermi accelerator has been proposed to explain the origin of fast cosmic rays and developed further by ulam@xcite . however , this model has had more importance as a standard hamiltonian system to exhibit chaotic dynamics . because of its simple composition and yet rich dynamics , it has been intensely studied both classically@xcite and quantum mechanically@xcite . the feature of harmonically driven barriers is frequently encountered in various physical situations@xcite , hence it provides very useful physical insights in spite of its simple structure . furthermore , physical features of the model have been realized experimentally in an atomic system@xcite . besides the considerations of various regimes , several varieties of the fermi accelerator have also been proposed , including relativistic models , of the type which embodies thermodynamic considerations@xcite and mechanical fermi accelerator with damping@xcite . our model can however be contrasted with the previous ones in terms of energy conservation . while the general fermi accelerator has a periodic external driving force , the two moving objects in our system simply exchange energy with each other , but always maintaining the total energy conserved . ( in the relativistic regime , the system can have a constant energy effectively in spite of the driving , but it is conditionally given . see @xcite ) accordingly , this model may be called an ` autonomous fermi accelerator'(afa ) , and the property of energy conservation enables us to calculate the exact eigenstates . with this set of eigenstates , we analyze the dynamics of the afa and derive the condition for the validity of the born - oppenheimer approximation in the chaotic regime . this paper is organized as following . in sec . ii , we describe the autonomous fermi accelerator schematically and present the analysis of its classical dynamics by phase space portraits . the phase space of the system shows under what conditions the chaotic dynamics sets in . then we apply the born - oppenheimer approximation to the system in sec . we then compare the approximated calculation with the exact numerical solution which is obtained in sec . iii . through this comparison , the connection between the conventional criteria of the born - oppenheimer approximation and the chaotic dynamics can be derived . ) and the total energy(@xmath0 ) for each plot are assigned as @xmath1 , @xmath2 for ( a ) and ( c ) , and @xmath3 , @xmath2 for ( b ) and ( d).,width=340 ] the autonomous fermi accelerator consists of a free particle confined between two infinite potential barriers . both barriers are not penetrable and one of them has a finite mass and moves in a harmonic potential , while the other one is fixed . the schematical description is shown in fig . [ fig1 ] . the hamiltonian of the autonomous fermi accelerator is given by @xmath4 where @xmath5 , @xmath6 and @xmath7 are a mass , momentum and postion variable for the oscillating heavy wall , while @xmath8 , @xmath9 and @xmath10 are those for the free particle . we set the fixed wall at @xmath11 and the minimum point of the harmonic potential at @xmath12 . also , we assume that every collision between the particle and the wall is elastic . first , we analyze the classical dynamics of this system by poincare surface of sections . the surface of section can be constructed stroboscopically by sampling the dynamical status of either the particle or the wall at each collision . the obtained results with different parameters are presented in fig . the right column of the figure shows the phase space plots of the light particle and the left one shows those of the center of masses . in this figure , we can notice that the feature of energy conservation is vividly shown ; namely that all dynamical phases are restricted on a spherical surface given by @xmath13 as fig . [ fig2 ] shows , the phase space of the system is quite clearly divided into two parts , a stochastic and a regular part . the border between them is conditioned by the existence of the double collision . here , the double collision means the situation that the light particle is hit by the heavy wall twice until it reaches the fixed wall . the presence of this type of collision diverges the light particle s trajectory drastically , and results in the stochastic dynamics . apparently , it can be a necessary condition for the double collision that the heavy wall is faster than the light particle . [ fig2 ] shows such tendency well . the stochastic region is located around the equator of the sphere in the phase space , while the regular island is around the pole . we also mention a structural detail in phase space : around the border between the chaotic and the regular region there exists stochastic web structure"@xcite . however , this structure is ignorable in the regime of wave mechanical parameters in this study . the structural properties of the phase space , including the portion of the regular and chaotic region , is controlled mainly by the mass ratio(@xmath14 ) . by the comparison of ( a ) and ( b ) in fig . [ fig2 ] , we can confirm the tendency that the regular island is shrinked as @xmath14 is increased . also the phase space is affected by the total energy(@xmath0 ) , but the portion of regular island is not changed as much as by the mass ratio , according to our numerical study . if the system is quantized , the total hamiltonian becomes @xmath15 and the schrdinger equation and the boundary condition are given by @xmath16 @xmath17 application of the born - oppenheimer approximation means to fix all parameters which are given instantaneously by the heavy object . in our system , this can be done by fixing the distance between the walls , to be @xmath7 , at every moment . then , a set of basis states can be constructed as follows : @xmath18 where @xmath19 using the completeness of the basis set above , the schrdinger equation ( [ scheq ] ) can be reduced in the following way : @xmath20 where @xmath21 . if eq . ( [ oned ] ) is inserted into ( [ boapp ] ) , one is able to expand the equation with the mass - ratio , @xmath22 as a parameter . by taking up to the first order of this expansion , the following equation is obtained , @xmath23 consequently , the governing equation is reduced to a simple 1-dimensional equation of motion with a stationary bounding potential and the third term of the equation can be thought of as a pressure generated by the frequent collisions of the light particle . according to the theory of hamiltonian chaos , there is no possibility of chaos if we consider the dynamics at the pure classical level . however , if we take the level transition into account , the effective potential can be thought of to be time - dependent , and it seems more reasonable as the counterpart of the classical chaotic system . along this line of reasoning , if an eigenmode is corresponding to chaotic dynamics , we can make an assumption that the lowest order of approximation would fail to describe the eigenmode . the validity of this assumption , will be discussed in the next section through a discussion of the exact solution which is numerically obtained . . : ( a ) regular mode : @xmath24 ( c ) chaotic mode : @xmath25 . ( b ) , ( d ) : corresponding husimi distributions to ( a ) and ( c ) , respectively , width=377 ] as mentioned in the introduction , the exact solutions of the system are also available . to calculate them , the following transform is performed . @xmath26 then the schrdinger equation becomes @xmath27 as the result of the transformation , the 1-dimensional dynamics of the system becomes a 2-dimensional dynamics in a triangular well , the three sides of which are two infinite potential barriers along the x - axis and @xmath28 , and one harmonic potential barrier to the x - direction . here , the angle @xmath29 between the two infinite barriers is given by : ( a ) regular mode : @xmath30 ( c ) chaotic mode : @xmath31 . ( b ) , ( d ) : corresponding husimi distributions to ( a ) and ( c ) , respectively.,width=377 ] @xmath32 where @xmath14 is the mass ratio(@xmath33 ) . it should be noted that @xmath29 would reach @xmath34 as @xmath14 approaches infinity . in this work , we are concerned only with the normal situation of born - oppenheimer approximation , namely that @xmath5 is much larger than @xmath8 . then the boundary condition is given as @xmath35 using a numerical approach , we obtain all possible eigenvalues and corresponding eigenmodes with two different mass - ratio parameters , @xmath36 and @xmath37 . as the numerical algorithm , ` finite element method'@xcite is used . [ fig3 ] and fig . [ fig4 ] show configurational plots and husimi distributions of eigenmodes at @xmath36 and @xmath38 , respectively . using the normal derivative of configurational mode distributions along @xmath39 , we can calculate the husimi distributions , and these are corresponding to the poincar surface of section of a center of mass . hence , the corresponding dynamics for each mode can be confirmed by comparison of the husimi distributions with the classical phase spaces in fig . [ fig2 ] . in fig . [ fig3 ] and [ fig4 ] , ( a ) and ( b ) of each figure shows a mode which is localized the regular regime in the phase space , and ( c ) and ( d ) shows ones in the chaotic regime . from the obtained computational results , the validity of the born - oppenheimer approximation is investigated . for this purpose , the fourier analysis is performed to find the eigenmode composition of the light particle . @xmath40 the result of the analysis shows the different features of distributions , depending on the underlying dynamics . for @xmath36 , the distribution is well localized regardless to the dynamics ( see fig . [ fig5](a ) . ) . for @xmath38 , however , the distribution shows the interesting features depending on the dynamics ( fig . [ fig5](b ) ) . the mode which is placed in the chaotic region shows a broad distribution , with tails reaching the regular region , but the one in the regular region still has a well localized distribution . considering the scheme of the born - oppenheimer approximation , modes with the high localization are expected to be well - approximated . to confirm this , we extract the wave function of the moving wall with the highest level ( @xmath41 ) in each distribution in the following way , @xmath42 fig . [ fig5 ] presents a comparison between the wave function extracted from the exact wave function in fig . [ fig3](c ) and the one approximated by the eq . ( [ approximation ] ) . at @xmath36 , the extracted wave function agrees very well with the approximation , regardless to where in phase space region a mode is situated ( fig . [ fig5](c ) ) . in contrast , the mode in the chaotic region for @xmath43 does not show a good agreement with the approximation , while the one in the regular region still abides by the aforementioned approximation ( fig . [ fig5](d ) ) . of course , the disagreement is due to the level transitions during the dynamics . ( @xmath36 ) under the instantaneous basis of the light particle . ( b ) the same one for fig . [ fig4 ] . in each figure , the upper distribution is for modes in the chaotic region and the lower one is for ones in the regular region . ( c ) , ( d ) the extracted wave functions of the moving wall(solid line ) and the approximated wave functions(dashed line ) , width=302 ] now , it is necessary to consider the general criteria for the variation of born - oppenheimer approximation@xcite , which is given as follows , @xmath44 this inequality simply guarantees that the heavy object is moving slowly enough to prevent level transitions . if we apply this inequality to our system , the condition of this inequality is not changed much by the dynamical status . in other words , the ratio between the left term and the right term in eq.([ineq ] ) is given as around @xmath45 for @xmath36 and around @xmath46 for @xmath38 , no matter whether a mode is chaotic or regular . considering the near - integrability in the case of regular dynamics , it is obvious that the simple approximation works well in the regular region , because we can define good quantum numbers for each degree of freedom . in addition , we can find the dynamical interpretation of the inequality ( [ ineq ] ) . in our system , the satisfaction of ( [ ineq ] ) guarantees that the chaotic region is small enough in comparison to planck s constant @xmath47 . therefore , it is impossible that more than two modes are situated in the chaotic region . accordingly , no transition is possible . in this work , we study the validity of the born - oppenheimer approximation in the presence of chaotic dynamics . especially , the criteria under which the approximation is working is interpreted from the dynamical point of view . we apply the approximation method to the autonomous fermi accelerator which has a clear division of regular dynamics and chaotic dynamics in its phase space . in the regular region , the approximation is obviously well suited to calculate the eigenmodes , whereas it is just conditionally suited in the chaotic region . however , when the dynamics of the system satisfies the general condition for the approximation , that is , the evolution of the system is slow enough to prevent the level transitions , the area of the chaotic region becomes small enough in comparison to planck s constant . thus , the chaotic dynamics can not be resolved by wave functions . thereby , the born - oppenheimer approximation can approximate the wave functions very well in spite of the chaotic nature of the dynamics . it is important to remark that in the region where the born - oppenheimer approximation is working , we can apply another approximation to our system . then we can follow the procedure of ref . @xcite to construct the density matrix and the reduced density matrix of the interesting subsystem ( light particle or oscillating wall ) . we leave this for a future work . we also remark that it would be quite instructive to study higher order corrections to the born - oppenheimer approximation @xcite in connection with chaotic dynamics . we also leave this for a future study . + + this work was supported in part by the max - planck - institute for the physics of complex systems ( mpipks ) in dresden , the emmy - noether - programme of the dfg ( germany ) and by the brazilian agencies , cnpq and fapesp . 99 m. born and v. fock , z. phys . * 51 * 165 ( 1928 ) . jie song_et al . lett . , * 80 * , 60001 ; a. m. childs _ et al . a * 65 * 012322 ( 2002 ) . k. marzlin and b. sanders , phys . lett . * 93 * , 160408 ( 2004 ) . d. m. tong , k. singh , l. c. kwek and c. h. oh , phys . rev . lett . * 98 * , 150402 ( 2005 ) . a. j. lichtenberg and m. a. lieberman , _ regular and stochastic motion " _ , ( springer - verlag , newyork , 1983 ) . e. fermi , phys . rev . * 75 * , 1169 ( 1949 ) . s. ulam , _ in proc . 4th berkeley simposium on mathematics and probability 3 " _ 315 , ( berkeley , los angeles , 1961 ) . f. lenz , f. k. diakonos , and p. schmelcher , phys . * 100 * , 014103 ( 2008 ) . r. blmel and b. esser , phys . 72 * , 3658 ( 1994 ) . p. seba phys . * 41 * , 2306 ( 1991 ) . f. saif , i. bialynicki - birula , m. fortunato , and w. p. schleich , phy . a * 58 * , 4779 ( 1998 ) . f. r. n. kock , f. lenz , christoph petri , f. k. diakonos , and p. schmelcher , phys . e. * 78 * , 056204 ( 2008 ) . f. marquardt , j. g. e. harris , and s. m. girvin , phys . 96 * , 103901 ( 2006 ) . a. steane , p. szriftgiser , p. desbiolles , and j dalibard , phys . rev * 74 * , 4972 ( 1995 ) . l. d. pustylnikov , russian math . surveys * 50:1 145 ( 1995 ) . a. c. j. luo and y. guo , mathmatical problems in engineering . 2009 * , article i d 298906 ( 2009 ) . g. m. zaslavsky , r. z. sagdeev , d. a. usikov and a. a. chernikov , _ weak chaos and quasi - regular patterns " _ , ( cambridgy university press , 1991 ) . ramdas ram - mohan , _ finite element and boundary element applications in quantum mechanic " _ , ( oxford university press , usa , 2002 ) . m. s. hussein and v. kharchenko , ann . phys . * 250 * , 352 ( 1996 ) . a. mostafazadeh , phys . rev.a , * 55 * , 1653 ( 1997 ) . r. mackenzie , e marcotte and h. paquette , phys . a * 73 * , 042104 ( 2006 ) . alfred shapere and frank wklczek , _ geometric phases in physics " _ , ( world scientific , singaport , 1989 ) .	we study the validity of the born - oppenheimer approximation in chaotic dynamics . using numerical solutions of autonomous fermi accelerators , we show that the general adiabatic conditions can be interpreted as the narrowness of the chaotic region in phase space .
for the design of simple transfers in astrodynamics , hohmann transfers are widely used . they consist basically in a maneuvre in a system represented by a two body problem ( 2bp ) and their solutions , consequently given by keplerian ellipses . starting from a circular orbit around a main body , a transfer to a different ( inner or outer ) circular orbit can be achieved with just two different impulses of properly defined sense and magnitude ( see e.g. @xcite ) . in cases where 2bp is not a suitable approximation , a similar approach can be considered with a different simplified version of the hamiltonian representing the system . the aim of the method lays on the idea of designing a transfer between orbits that are _ exact solutions _ of an _ integrable approximation _ of the studied system , using a set of impulses . since , in general , physical problems have more than one degree of freedom , a suitable construction of a normal form approximating the hamiltonian of the system is needed . furthermore , many d.o.f . make representations in configuration space inadequate , since they do not give precise hints of the time evolution of the orbits at glance . thus , the baseline is the construction of a normalized integrable approximation of the model to study , that should provide : i ) analytical solutions for the motions , and ii ) suitable surfaces of sections , tools to be used for the analysis of the effects of the impulses that conform the trasfer . + in the last decades , several semi - analytical results have been obtained in order to ensure the stability of the motions of some trojan asteroids , orbiting around the lagrangian equilibrium points @xmath2 of the sun jupiter system . all those works share a same common structure , summarized as follows : each approach is based on an explicit algorithm , that can be translated on a computer so as to calculate the expansion of a suitable normal form , providing a good local approximation of the complete hamiltonian of the cprtbp . such a normal form is used to approximate the orbits of some objects and to prove their stability , provided they are close enough to the equilateral lagrangian points . as far as we know , the wider coverage of the tadpole orbits around @xmath2 is given in @xcite , that is based on the kolmogorov normal form . here , we try to improve those results , by revisiting the approach developed in @xcite . from that article we borrow two main ideas : first , we perform the initial expansions of the cprtbp hamiltonian in a suitable set of poincar delaunay like coordinates ( while polar coordinates were used in @xcite ) . that particular type of canonical variables allows us to clearly distinguish a pair of slowly varying coordinates from those quickly changing their values therefore , we average the hamiltonian with respect to the angle related to the fast dynamics . this second main idea is implemented here , by using the modern lie series formalism so as to construct an _ integrable _ normal form ; this allows us to approximate also the tadpole orbits going rather far from @xmath3 . in fact , as a major novelty with respect to @xcite ( that is based on purely analytical techniques ) , we have translated our procedure in some codes , producing suitably truncated expansions of the normal form and , therefore , explicit numerical results . having the tools described before , we create an algorithm which is able to test the effectiveness of different impulses , after choosing an arbitrary starting point . this approach can be applied in many different systems . in particular , we focus our work in the area surrounding the lagrangian equilibrium points @xmath2 of the earth - moon system . this kind of stability region is interesting since it is located close to the earth , and could provide a natural trapping zone for small bodies as astronomical observatories , obsolet spacecrafts or space debris in general . at the end of the paper , we characterize the effects of different instantaneous impulses and we choose the best candidates for effective transfers in the framework provided by the normal form approximating the cprtbp hamiltonian . let us introduce the standard framework of the cprtbp , as done , e.g. , in @xcite . in such a model , the motion of the two biggest bodies ( hereafter , the primaries ) is not influenced by the third one , considered massless , so the orbits of the primaries are keplerian ellipses and the third body moves under the gravitational attraction exerted by the other two . additionally , we assume that both those keplerian orbits are circular and the massless body is coplanar with the primaries . let us define the heliocentric vectors @xmath4 with @xmath5 , being @xmath6 , @xmath7 and @xmath8 the position vectors of the biggest primary , the smallest one and the third body , respectively , in an _ inertial _ frame . as usual , we also define the units of measures in order to set the rotation period @xmath9 and the gravitational constant @xmath10 , and we denote with @xmath11 and @xmath12 the masses of the smallest primary and the largest one , respectively . these settings imply that the semi - major axis of the ellipse described by the orbit of @xmath13 is equal to @xmath14 , while @xmath15 . for our purposes , it is convenient to adopt the hamiltonian formalism in the non - inertial _ synodic _ frame that co - rotates with the primaries . we define the axes so that the largest primary is always located at the origin of the synodic frame , while the fixed position of the smallest primary is such that @xmath16 . let us first introduce the action angle canonical coordinates @xmath17 , which can be seen as modified delaunay variables for the _ planar keplerian problem_. they are given by @xmath18 where @xmath19 , @xmath20 , @xmath21 and @xmath22 are the semi - major axis , the eccentricity , the mean anomaly and the longitude of the perihelion of the massless body , being the angle @xmath22 measured in the _ inertial _ frame . moreover , @xmath23 and @xmath24 denote the mean anomaly and the perihelion longitude of the smallest primary , respectively . thus , @xmath25 corresponds to the _ synodic mean longitude_. the hamiltonian ruling the motion of the third body in the non - inertial synodical frame can now be written as follows : @xmath26 where the so called disturbing function @xmath27 is such that @xmath28 in order to remove the singularity of the delaunay like variables when @xmath29 ( i.e. , for circular orbits ) , it is convenient to introduce canonical coordinates @xmath30 , similar to the poincar coordinates for the _ planar keplerian problem_. thus , let us define @xmath31 where the values of @xmath32 are significantly small in a region surrounding the lagrangian points , for instance , in the case of tadpole or horseshoe orbits . in fact , in those cases , it holds true since @xmath33 and @xmath34 . let us recall that those variables have been adopted in @xcite to study the cprtbp . before constructing a normal form , the starting hamiltonian ( in particular , the disturbing function ) must be expanded in poincar like coordinates @xmath30 . such a non trivial operation is described in @xcite , which also includes all the detailed mathematica codes explicitly producing those expansions . this approach ensures that the starting hamiltonian @xmath35 can be written in the following form as a function of the poincar like coordinates : @xmath36 where @xmath37 and @xmath38 @xmath39 , being @xmath40 a fixed positive integer . @xmath41 is the set of functions such that * a function @xmath42 if the generic terms appearing in its taylor fourier expansion , which are of type @xmath43 or @xmath44 , satisfy the following relations about their coefficients : latexmath:[\[c_{m_1,m_2,m_3,k}=d_{m_1,m_2,m_3,k}=0 \qquad{\rm when}\qquad 2m_1+m_2+m_3\neq l \quad{\rm or}\quad in principle , the expansion can be seen as a reorganization of the taylor fourier series giving the disturbing function ; this is made in a suitable way that allows us to successfully perform the construction of the normal form . furthermore , the criterion for the choice of @xmath40 is such that the size ( in any common functional norm ) of @xmath46 , @xmath47 , @xmath48 is approximately the same for any value of the index @xmath49 . in other words , the disturbing function is splitted in terms @xmath50 , @xmath51 , @xmath48 by using the fourier decay of the coefficients for increasing values of the harmonic @xmath52 ; let us recall that in the present work @xmath11 is regarded as a fixed small parameter of the system . let us emphasize that the expansion contains also a non - trivial information about the keplerian part ( that corresponds to the whole hamiltonian if @xmath53 ) . in fact , since @xmath54 and @xmath55 , then one can deduce that @xmath56 when the index @xmath49 is odd . this is in agreement with the expansion of the starting hamiltonian @xmath35 , when the disturbing function is neglected and the actions @xmath57 and @xmath58 are substituted according to formula . if we explicitely write the first main terms of the keplerian part @xmath59 ^ 2= -\frac{3}{2}+\gamma-\frac{3}{2}(\rho+\gamma)^2\ , \label{eq : quadr - part - kepl - appr}\ ] ] we conclude that the angular velocities have different order of magnitude @xmath60 because @xmath61 and the values of the actions @xmath32 and @xmath58 are small in a region surrounding the lagrangian points . thus , @xmath25 can be seen as a slow angle and @xmath62 as a fast angle . then , this motivates to average the hamiltonian over the fast angle ( see , e.g. , 52 of @xcite ) , in order to focus mainly on the secular evolution of the system . we remove all the terms depending on the fast angle @xmath62 , by performing a sequence of canonical transformations . in the following , this strategy will be translated in an explicit algorithm . such a procedure will allow us to produce a final hamiltonian satisfying two important properties : at the same time it provides a good approximation of the starting system , it only depens on the actions and one of the angles , i.e. , it is integrable . as discussed above , the normalization algorithm defines a sequence of hamiltonians . this is done by an iterative procedure ; let us describe the basic step which introduces @xmath63 starting from @xmath64 when both the values of the indexes @xmath65 and @xmath66 are positive . we assume that the expansions of @xmath64 is such that @xmath67 where @xmath68 @xmath69 , @xmath70 @xmath71 , @xmath72 @xmath73 , @xmath74 @xmath75 . let us recall that the hamiltonian is suitable to start the procedure with @xmath76 , after having set @xmath77 . in formula , one can distinguish the normal form terms from the perturbing part ; the latter depends on @xmath78 in a _ generic _ way , while in the first two terms of the fast variables can be replaced by the action @xmath79 . the @xmath80th step of the algorithm formally defines the new hamiltonian as @xmath81 where @xmath82 is the lie series operator , with @xmath83 ( being @xmath84 the classical poisson bracket ) , @xmath22 a generic function defined on the phase space and @xmath85 any generating function ( for an introduction to canonical transformations by lie series in the context of the hamiltonian perturbation theory , see , e.g. , @xcite ) . the new generating function @xmath86 is determined so as to remove from the main perturbing term is expected to decrease when the indexes @xmath87 or @xmath66 are increased , because the values of @xmath11 , @xmath32 and @xmath88 are assumed to be small ] @xmath89 its subpart that is not in normal form . this is done by solving the following homological equation with respect to @xmath90 : @xmath91 where we require that @xmath92 is the new term in normal form , i.e. @xmath93 . when @xmath94 and @xmath95 , then there exists a generating function @xmath96 and a normal form term @xmath97 solving the homological equation . [ lem : sol_homol_eq ] we limit ourselves to just sketch the procedure that can be followed so as to explicitly determine a solution of and , therefore , prove the statement above . first , we replace the fast coordinates @xmath78 with the pair of complex conjugate canonical variables @xmath98 such that @xmath99 and @xmath100 . moreover , the homological equation has to be expanded in taylor series with respect to @xmath98 , using the slow coordinates @xmath101 as fixed parameters ( because they are not affected by the poisson bracket @xmath102 , since @xmath103 do not depend on them ) . therefore , we solve term - by - term the equation in the unknown coefficients @xmath104 and @xmath105 such that @xmath106 \ , \qquad z_{r_2}^{(r_1)}(\rho , z,\lambda,\imunit{\overline z})= \sum_{m}\big[\zeta_{m}(\rho,\lambda)z^{m}(\imunit{\overline z})^{m}\big]\ .\ ] ] at last , we express the expansions above by replacing @xmath98 for @xmath78 , so as to obtain the final solutions in the form @xmath90 and @xmath93 . the following property of the poisson brackets is very useful for our purposes , and since the proof is immediate , it is omitted . let @xmath107 and @xmath22 be two generic functions such that @xmath108 and @xmath109 , then @xmath110 [ lem : alg - prop - pois - brackets ] in order to provide an algorithm easy to translate in a programming language , we are going to give explicit formulas for the new hamiltonian @xmath63 and for its expansion that can be written as follows : @xmath111 for the sake of simplicity in the calculation of @xmath112 , we redefine the same quantity several times using the same symbol . this notation of the algorithm is more similar to its translation in a programming code , and , thus , more useful : let us introduce the recursive operation @xmath113 , where the previously defined quantity @xmath114 is redefined as @xmath115 . therefore , we initially define @xmath116 then , we consider the contribution of the terms generated by the lie series applied to each function belonging to the normal form part as follows : @xmath117 where the upper limits @xmath118 and @xmath119 on the indexes @xmath120 and @xmath49 , respectively , are such that @xmath121 for what concerns the contributions given by the perturbing terms making part of the expansion of @xmath64 in , we have @xmath122 where the limiting values for the indexes , that are @xmath118 and @xmath123 , are defined in . the redefinition rules and are set so that the new perturbing part generated by the lie series in is coherently split in different terms according to their order of magnitude in @xmath11 and their total polynomial degree in the actions . in fact , by applying repeatedly proposition [ lem : alg - prop - pois - brackets ] to the redefinitions in , it is possible to inductively verify that @xmath124 @xmath125 . therefore , the terms making part of the hamiltonian @xmath63 in the expansion share the same properties with those appearing in ; this ensures that the normalization algorithm can be iterated so as to construct @xmath126 from an ideal point of view , we would be interested in producing the final hamiltonian @xmath127 , where @xmath128 is defined as @xmath129 @xmath130 . in fact , such a hamiltonian would be integrable , because it would depend from the fast variables just through the action @xmath131 , due to the special form of the expansion . in general , the problem , depends just on the fast action @xmath131 . ] of the restrictions of domains prevents the convergence of the limits ( with respect to both indexes @xmath65 and @xmath132 , when the standard @xmath133norm is used ) ; thus , it is not possible to define the integrable hamiltonian on any open set ( see , e.g. , @xcite ) . actually , from a mathematical point of view , expansions of type are _ asympotic series _ with respect to both @xmath65 and @xmath132 ; this means that we have to truncate the indexes , but in a way that optimize our result . for this purpose , we can proceed as follows . first , we introduce the functions @xmath134 and @xmath135 that make explicit the splitting between the integrable and the perturbing parts in the expansion , so that @xmath136 therefore , we look for the pair of upper indexes @xmath137 minimizing the @xmath133norm of @xmath138 ( on the set of values of the variables that we are interested to study ) . this approach can be implemented with a suitable scheme of analytic estimates , in order to reduce exponentially the remainder @xmath139 with respect to the small parameters of the problem ( see @xcite , @xcite and @xcite ) . such an optimal choice about the final values of the indexes @xmath65 and @xmath66 allows us to reformulate the algorithm , in such a way that it requires just @xmath140 normalization steps , constructing the finite sequence of hamiltonians @xmath141 , where @xmath142 @xmath143 . while the algorithm has been rearranged so as to be performed in a finite number of normalization steps , it is evident that the redefinition rules reported in formulas and would require to calculate infinitely many poisson brackets . in order to avoid such a problem , we have to establish two truncation rules on the terms appearing in the expansions , so as to fix ( a ) their maximal exponent @xmath144 related to the order of magnitude @xmath145 , ( b ) their total maximal degree @xmath146 on the index @xmath49 ( that is equal to twice the degree in @xmath32 plus the one in @xmath147 and that in @xmath148 ) . if our formal algorithm is subject to a further restriction , such that it is limited to the calculation of functions belonging to classes of type @xmath41 with @xmath149 and @xmath150 , then it can be proved that it requires a _ finite _ total number of poisson brackets . therefore , this newly restricted version of our algorithm is suitable to be translated in a programming code . in principle , the values of @xmath146 and @xmath144 should be chosen in order to optimize the final results ; in practice , they are usually fixed ( as well as the final indexes @xmath151 and @xmath152 ) so as to fit with the available computational resources . it is well known that lie series induce canonical transformations in a hamiltonian framework ; this fundamental feature will allow us to design a numerical integration method , by using both the normal form discussed above and the corresponding canonical coordinates . in order to explicitly realize such a project , we have to introduce some more complicate notations . let us denote with @xmath153 the set of canonical coordinates related to the @xmath80th step . by appying the so called exchange theorem ( see , e.g. , @xcite ) , we have that @xmath154 where the variables related to the previous step , namely @xmath155 , are given as @xmath156 the r.h.s . of the equation above means that four lie series must be applied separatedly to each variable , in order to properly define all the coordinates for the canonical transformation @xmath157 . the whole normalization procedure can be described by the canonical transformation @xmath158 such a composition of all the intermediate changes of variables can be used for providing the following semi - analytical scheme to integrate the equations of motion : @xmath159 where @xmath160 is the flow induced on the canonical coordinates by the generic hamiltonian @xmath161 for an interval of time equal to @xmath162 . let us emphasize that the above integration scheme provides an approximate solution ; from an ideal point of view ( i.e. , if all the expansions were performed without errors and truncations ) , formula would be exact if @xmath163 would correspond to the complete hamiltonian @xmath164 . on the other hand , @xmath163 is _ integrable _ and its flow is easy to compute , it is convenient to introduce the temporary action angle variables @xmath165 such that @xmath166 and @xmath167 , where @xmath168 is a constant of motion for the normal form @xmath169 . by considering @xmath168 as a fixed parameter and using the standard quadrature method for conservative systems with @xmath14 d.o.f . , one can compute @xmath170 and @xmath171 at any time @xmath172 . the same can be done for the evolution of @xmath173 , by evaluating the integral corresponding to the differential equation @xmath174 . for practical purposes , the application of the classical quadrature method can be replaced by any numerical integrator that is precise enough . finally , the values of @xmath175 and @xmath176 can be directly calculated from those of the corresponding action angle variables , that are @xmath177 and @xmath173 . ] , reasons why using @xmath163 becomes valuable . the approximate solution provided by the scheme is as more accurate as smaller the perturbing part @xmath139 is with respect to @xmath163 ( see their definitions in ) . as a final remark , we stress that it requires a _ finite _ total number of operations , if we limit the the expansions of the canonical transformations as discussed in the previous subsection . thus , also the whole integration scheme can be translated in a programming code . in order to test the accuracy of our integrable normal form approximating the hamiltonian , we have to choose a suitable surface of section for the comparison with the complete problem . the most logical choice about the sectioning of the flow is given by the surface defined by @xmath178 , because @xmath62 is a fast angle ( recall the discussion about formula ) . in fig . [ fig : surf_sec ] we show the results of the comparison between the complete cprtbp hamiltonian and the normal form approximating it , in the case of the earth - moon system , which is defined by its value of the mass parameter @xmath179 . in the left panel , we show the surface of section numerically computed by considering the equations of motion related to the cprtbp hamiltonian . we take a set of 10 equispaced initial conditions , with @xmath180 , @xmath181 , @xmath182 . the value for @xmath58 has been set in such a way that all the initial conditions keep the same value for the jacobi constant as in @xmath183 . these orbits has been integrated up to recover 1000 points over the surface defined by @xmath178 . those points are drawn in black in the space of variables ( @xmath25,@xmath32 ) . is located at ( @xmath184 , @xmath185).,title="fig : " ] is located at ( @xmath184 , @xmath185).,title="fig : " ] is located at ( @xmath184 , @xmath185).,title="fig : " ] for the same initial conditions as before , we have integrated the orbits also according to the semi - analytical scheme , up to collect also 1000 points over the surface . this computations have been automatically made by a code written in * c * , in such a way to use the expansions of the normal form @xmath163 and the canonical transformation @xmath186 , which were preliminarly produced by using mathematica . the truncations have been made according to the following values of the parameters ruling the extension of those expansions : @xmath187 , @xmath188 , @xmath189 , @xmath190 and @xmath191 . these values imply reasonable computing times . the results are expressed in the middle panel of fig . [ fig : surf_sec ] . finally , in the right panel , we show the comparison between the two surfaces of section . for all the orbits close to the equilibrium point @xmath183 , the correlation is extremely good , thought the approximation fails on reproducing exactly the tadpole orbits far from @xmath183 . since the new normal form has @xmath14 d.o.f . , it provides the integrable approximations of the motions of small bodies in the vicinity of the stable equilibrium points and suitable surfaces of section that are easy to compute . these tools can be used in many different studies . as explained in the introduction , we focus in the design of maneuvres of small bodies ( as spacecrafts or space debris ) that are originally out of the stability region filled by tadpole orbits and we would like to situate into it . with such an aim , we design a method that allows to study the effectiveness of a set of impulses , done after considering a fixed initial position . as outline of the algorithm , we have to * choose a starting point @xmath192 , which corresponds to the position of the body to transfer at @xmath193 and translate it to cartesian coordinates @xmath194 . * give an impulse @xmath195 where @xmath196 and @xmath197 are the modulus and direction of the impulse , respectively , in cartesian coordinates . this gives a new orbit , whose evolution in normalized variables can be computed with the integrable approximation of the hamiltonian and checked in the surfaces of section . * decide the acceptance or rejection of the impulse , according to whether it fulfills or not the conditions of the transfer . as a matter of fact , this mechanism can be applied for a physically adequate range of moduli and directions and not one orbit by one , in order to speed up the computations . + we apply the method described above to the case of the earth - moon system . as starting point @xmath198 , we choose a position slightly outside the real stability region is in the chaotic region related to the stable / unstable manifolds emanating from @xmath199 . it has been fixed in order to be one of the points closest to @xmath0 among those belonging to the chaotic set and to the axis of the abscissas of the surface of section defined by @xmath178 . ] , estimated by numerical integrations of the full problem . @xmath192 is given by @xmath200 , @xmath201 , @xmath202 and @xmath203 . for the transfer orbits , we apply @xmath204 different impulses , for which @xmath205 and @xmath206 . following 2.8 in @xcite , and , we translate every orbit to normalized variables . in fig . [ fig : impulses ] , we present the results of the translation of the new coordinates calculated after each impulse ( for the first new point of the orbit on the surface of section ) in the corresponding variables . the color scale represents the size of the impulse @xmath196 in the left panel and its direction @xmath197 in the right panel . as expected , a stretching effect occurs while translating to the new system , and the impulses are more effective ( i.e. , same values imply a bigger difference with respect to the original point @xmath192 ) in some directions . , @xmath32 ) . the color scale represents the magnitude of the impulse @xmath196 . right panel , same as before with color scale representing the direction of the impulse @xmath197.,title="fig : " ] , @xmath32 ) . the color scale represents the magnitude of the impulse @xmath196 . right panel , same as before with color scale representing the direction of the impulse @xmath197.,title="fig : " ] in order to distinguish which impulses generate orbits that go deeper inside the stability region , we isolate the impulses pointing towards the inner part . we refine the values of the angles considered , taking @xmath207 equidistant values @xmath197 , where @xmath208 . we integrate their evolution with respect to the normal form @xmath209 and we find the curves described by these orbits on the surface of section . in the averaged normal form , each of these curves depicts the area associated to one of the actions defining the 2-d torus that is invariant with respect to the motion . since as closer the torus is to the equilibrium point @xmath183 , as smaller this area is , we take as criterion for a good transfer a reduction of this value with respect to the initial one ( _ minimum action criterion _ , since such an area corresponds to an action ) . in fig . [ fig : areas ] , we show the results of the computation of the area in two different color - scales . in the left panel , we can see for all the considered combinations of modulus @xmath196 and angles @xmath197 , the value of the area described by the transfered orbit . the lower border of the plot represents the orbits with @xmath210 , so their areas ( @xmath211 ) are equal and can be used as reference value . in the right panel , only orbits corresponding to areas @xmath212 are considered , the greater ones are fixed equal to @xmath213 . this allows a better discrimination between all the orbits providing suitable impulses . and @xmath197 , the color scale represents the results of the calculations using the averaged normal form . right panel , same as before where the computed areas are reported only if they are smaller than the value ( @xmath211 ) corresponding to the original initial condition @xmath198.,title="fig : " ] and @xmath197 , the color scale represents the results of the calculations using the averaged normal form . right panel , same as before where the computed areas are reported only if they are smaller than the value ( @xmath211 ) corresponding to the original initial condition @xmath198.,title="fig : " ] both plots shows that the selection of the angle for the impulse is not trivial , since small differences generate very different results . consider , for instance , the impulses with @xmath214 and with @xmath215 , for @xmath216 ) . the former give the best approach to the optimal transfer we look for , while the latter generate orbits which are actually further from @xmath0 than that related to the initial condition @xmath198 . in general , since we take small values for the impulses , in the correct directions , bigger @xmath196 implies smaller final areas . the best choices for the transfer correspond to the impulses for which @xmath217\,$ ] , and suitable values for their sizes , following the darker stripes in fig . [ fig : areas ] . in this work we explicitly construct an integrable normal form approximating the cprtbp hamiltonian . this is done by reformulating the approach described in @xcite so as to use some more modern techniques , developed in the framework of the hamiltonian perturbation theory and mainly based on the lie series formalism . this allow us to design an algorithm that can be fully translated in programming codes . in particular , we produce a truncation of the normal form @xmath209 , whose the expansion in highlights that it is an average of the cprtbp hamiltonian with respect to the angle associated to its fast dynamics . the first results provided by this revisited approach are encouraging : in some suitable surface of sections our algorithm provides very good approximations of the tadpole orbits close enough to @xmath3 . however , we are aware of the fact that the accuracy of our expansions must be strongly improved in order to face challenging concrete problems in a region far from those equilibrium points . in our opinion , the main constraint on the quality of our results is due to the truncations on the fourier series in the slow angle @xmath25 . we think that this limitation can be removed , by representing the dependence on @xmath25 in a suitable way , so as to avoid fourier expansions . we plan to investigate such a new approach in the near future . furthermore , we show a first astrodynamical application starting from our calculation of the integrable normal form @xmath209 , which approximates the cprtbp hamiltonian . we design an algorithm that allowed to compute optimal transfers between orbits in the neighborhood of the equilateral lagrangian equilibrium points . we generate impulses in cartesian variables according to formula , on a grid of values for @xmath196 , the magnitude of the impulse , and @xmath197 , the direction . for those , we are able to discriminate the _ suitable transfers _ , that imply a final orbit closer to the equilibrium point , according to a new criterion . using our normal form as an approximation of the complete cprtbp , we estimate the area enclosed by the final orbit , and minimizing this quantity , we select the best candidates for the transfer . a careful inspection of the plots shows that the best candidates are highly depending on the size of the impulse , and very sensitive to the changes on the angle @xmath197 . the authors would like to thank c. efthymiopoulos , because of his constant support which allowed us to finalize our mathematica codes . we are indebted also with c. sim who suggested to one of us to reconsider the fundamental work @xcite by garfinkel . during this work , r.i.p . was supported by the astronet - ii marie curie training network ( pitn - ga-2011 - 289240 ) , while u.l . was partially supported also by the research program `` teorie geometriche e analitiche dei sistemi hamiltoniani in dimensioni finite e infinite '' , prin 2010jj4kpa_009 , financed by miur .	in this work , we study the motions in the region around the equilateral lagrangian equilibrium points @xmath0 and @xmath1 , in the framework of the circular planar restricted three - body problem ( hereafter , cprtbp ) . we design a semi - analytic approach based on some ideas by garfinkel in @xcite : the hamiltonian is expanded in poincar delaunay coordinates and a suitable average is performed . this allows us to construct ( quasi ) invariant tori that are moderately far from the lagrangian points l4-l5 and approximate wide tadpole orbits . this construction provides the tools for studying optimal transfers in the neighborhood of the equilateral points , when instantaneous impulses are considered . we show some applications of the new averaged hamiltonian for the earth - moon system , applied to the setting - up of some transfers which allow to enter in the stability region filled by tadpole orbits .
nucleon knockout reactions in nuclear collisions at and above the fermi energy in nuclei have become an important tool to determine the occupancy of single - particle states and the correlation effects in the nuclear many - body system ( see , e.g. refs . @xcite ) . in peripheral , sudden collisions of fast - moving projectiles with a target nucleus , a single nucleon is removed from the projectile , producing projectile - like residues in the exit channel , which are measured . referred to the center - of - mass system of the projectile , the transferred momentum is @xmath0 . for the knockout reactions in the sudden approximation , this must equal the momentum of the struck nucleon before the collision . the standard reaction models assume that the ground state of the projectile of spin and parity @xmath1 can be approximated by a superposition of configurations of the form @xmath2^{j^{\pi}}$ ] where @xmath3 denote the core states and the @xmath4 are the quantum numbers for the single particle wave functions in a spherical mean field potential . the measured partial cross sections to individual final levels @xmath5 of the core allow to extract , by comparison with theoretical calculations , the spectroscopic factors for the individual core - single particle configurations . in complete analogy to the use of angular distributions in transfer reactions , the orbital angular momentum @xmath6 is revealed by the shape of the momentum distributions @xmath7 . it is obvious then that the accuracy of the extracted spectroscopic factors , a measure of the occupancy of the single - particle orbitals in nuclei , and the conclusions that may follow from these spectroscopic factors about correlations inside nuclei depend in direct measure on the accuracy of the cross section calculations . similarly , the cross sections for one - proton removal reactions @xmath8 are directly related to the non - resonant part of the astrophysical s - factors for the inverse radiative proton capture y(p,@xmath9)x ( see , e.g. , refs . @xcite ) . again , the results and their reliability depend directly on the reliability and accuracy of the reaction calculations ( as discussed in @xcite ) . the one - nucleon removal cross section is calculated in most reaction models as an incoherent sum of the contributions of all core - single particle configurations making the ground state of the fast moving projectile : @xmath10 where @xmath11 and @xmath12 are the spectroscopic factors of each configuration and the single particle removal cross section , respectively @xcite . a similar relation is valid for momentum distributions . systematic studies of projectiles and reactions allow the determination of the ordering , spacing and the occupancy of orbitals , essential in assessing how nuclei evolve in the presence of large neutron or proton excess . much was done in this respect in the last decade in various laboratories . this information can be compared to many - body nuclear structure calculations which are now able to reproduce the measured masses , charge radii and low - lying excited states of a large number of nuclei . it was found that , e.g. , for very exotic nuclei , the small additional stability that comes with the filling of a particular orbital can have profound effects upon their existence as bound systems , their lifetime and structure , and lead to the discovery of magic numbers that do not manifest along the valley of stability . extensions of the nucleon knockout formalism including the treatment of final - state interactions have been discussed in ref . @xcite where it is shown that coulomb final - state interactions are of relevance . in the meantime , inclusion of higher - order effects @xcite and a theory for two - nucleon knockout @xcite have been developed . knockout reactions represent a particular case for which higher projectile energies allow a simpler theoretical treatment of the reaction mechanism , due to the simplicity of the reaction mechanism and the assumption of a single - step process . a microscopic approach to direct reactions uses an effective nucleon - nucleon ( nn ) interaction ( e.g. those of ref . @xcite ) to start with . this interaction is often used to construct an optical potential with its imaginary part assumed to relate to the real part and its strength adjusted to reproduce experimental data . the real and imaginary parts of the potential can also be independent as in refs . @xcite , where the procedure starts from a nn effective interaction with independent real and imaginary parts . for collisions at high energies ( @xmath13 ) , it is possible to show that instead of nucleon - nucleon interactions one can use nucleon - nucleon cross sections as the microscopic input @xcite . in this case , an effective treatment of pauli - blocking of nucleon - nucleon scattering is needed , as it manifests through medium modification of nucleon - nucleon cross sections . it is well known that medium modification of the nucleon - nucleon cross sections is necessary for an adequate numerical modeling of heavy - ion collision dynamics in central collisions . in these collisions , the ultimate purpose is to extract information about the nuclear equation of state ( eos ) by studying global collective variables describing the collision process . in direct reactions , such as one - nucleon removal reactions , medium effects of nn scattering are smaller because mostly low nuclear densities are probed . a first study of this effect in knockout reactions was carried out in ref . nonetheless , no comparison with experimental data was provided . in this work we explore further consequences of medium corrections and final state interactions in knockout reactions . we study medium effects in the nn cross section in knockout reactions using the methods reported in ref . @xcite , namely with a geometrical treatment of pauli - blocking and with the dirac - brueckner theory in terms of baryon densities . we also explore the effect of final state interaction , in particular the effects of coulomb distortion in the entrance and final reaction channels . this is of relevance as an increasing number of experiments use heavy targets with a large nuclear charge . we compare our results of knockout cross section and momentum distribution calculations to a large number of published experimental data . the purpose is to improve the accuracy of the extracted spectroscopic factors that will lead to better understanding nuclear structure and to check and improve the reliability of the use of one - nucleon removal reactions as indirect methods in nuclear astrophysics . the geometrical treatment of pauli corrections is performed using the isotropic nn scattering approximation because the numerical calculations can be largely simplified if we assume that the free nucleon - nucleon cross section is isotropic . in this case , a formula which fits the numerical integration of the geometrical model reads @xcite @xmath14{c}\displaystyle{1-\frac{37.02 \tilde{\rho}^{2/3}}{e } } , \ \ \ { \rm if } \ \ e>46.27 \tilde{\rho}^{2/3}\\ \ , \\ \displaystyle{\frac{e}{231.38\tilde{\rho}^{2/3}}},\ \ \ \ \ { \rm if } \ \ e\le 46.27 \tilde{\rho}^{2/3}\end{array } \right . \label{vm1}\end{aligned}\ ] ] where @xmath15 is the laboratory energy in mev , @xmath16 , @xmath17 , @xmath18 is the local density of nucleus @xmath19 , and @xmath20 @xmath21 . the parameters and models for the @xmath22 and @xmath23 densities which have been used to describe the nuclei in this work are presented in table [ grdens ] . the brueckner method goes beyond the simple geometrical treatment of pauli blocking . some of the brueckner results that we used in this analysis have been reported in refs . @xcite , where a simple parametrization was given . it reads ( the misprinted factor 0.0256 in ref . @xcite has been corrected to 0.00256 ) @xmath24 \nonumber \\ & \times&\frac{1 + 0.0034e^{1.51 } \rho^2}{1 + 21.55\rho^{1.34 } } , \nonumber\\ \sigma_{pp } & = & \left [ 23.5 + 0.00256\left ( 18.2-e^{0.5}\right)^{4.0}\right]\nonumber \\ & \times&\frac{1 + 0.1667e^{1.05 } \rho^3}{1 + 9.704\rho^{1.2 } } . \label{brueckner}\end{aligned}\ ] ] the limits of validity of this parametrization are clearly associated with the limits of validity of the brueckner calculations , which are valid only below the pion - production threshold . a modification of this parametrization was introduced in ref . @xcite and consists in combining the free nucleon - nucleon cross sections parametrized in ref . @xcite with the results of brueckner theory reported in refs . @xcite . clllll nuclei & & & & & + & & & & & + ( target ) & + @xmath25be & ho@xmath26 & 2.50(9 ) & 2.367 & 1.77(6 ) & 0.631 + @xmath27c & ho@xmath28 & - & 2.332 & 1.584 & - + + @xmath29be & ho@xmath26 & 2.50(9 ) & 2.372 & 1.77(6 ) & 0.631 + @xmath30n & ho@xmath26 & 2.540(20 ) & 2.410 & 1.729(6 ) & 1.291 + @xmath31b & ldm@xmath32 & - & - & - & - + @xmath33 mg & hfb@xmath34 & - & 2.92 & - & - + @xmath35o & ldm@xmath32 & - & - & - & - + @xmath36 mg & hfb@xmath34 & - & 3.187 & - & - + current theoretical models for the calculation of momentum distributions and cross sections in high - energy nucleon - removal reactions follow a semiclassical probabilistic approach , described , e.g. , in refs . @xcite . the method relies on the use of survival amplitudes " ( or s - matrices ) in the eikonal approximation , @xmath37=\exp\left[-\frac{i}{\hbar v}\int_{-\infty}^\infty u_{it}({\bf r})dz\right ] , \label{sib}\ ] ] where @xmath38 , and @xmath39 is the particle(@xmath19)-target(@xmath40 ) optical potential . in ref . @xcite , a relation has been developed between the optical potential and the nucleon - nucleon scattering amplitude . such a relation is often referred in the literature as the t-@xmath41 approximation " . the t-@xmath41 approximation is the basis of most calculations of elastic and inelastic scattering involving radioactive nuclei , as experimentally deduced optical potentials are not often available . in this approximation , the eikonal phase becomes @xmath42 where @xmath43 is the fourier transform of the nuclear densities of the projectile and target , and @xmath44 is the high - energy nucleon - nucleon scattering amplitude at forward angles , which can be parametrized as @xmath45 there are many ways of introducing final state interactions in direct nuclear reactions , some of which are discussed in refs . @xcite . besides coulomb repulsion , included by modifying the straight - line trajectories accordingly , we have also modified the integral in eq . by using the optical potential including the coulomb potential that modifies the s - matrices according to @xmath46 . the coulomb phase in @xmath47 is calculated by assuming a uniform charge distribution with radius @xmath48 , and is given by @xmath49 \right\ } \ , \label{chico_0}\ ] ] where @xmath50 is the step function . the value @xmath48 is chosen to be small enough so that the nuclear s - matrices are basically zero below @xmath51 because of the strong absorption at small impact parameters . in this section the results for momentum distributions and nucleon - removal cross sections are compared to several experimental data . the focus is the importance of medium corrections of nucleon - nucleon cross sections and of coulomb distortions . both effects are expected to decrease as the bombarding energy increases . it is important to include such effects in order to minimize the uncertainty in the extraction of spectroscopic factors , especially at low bombarding energies . to substantiate this assertion , we analyze low energy data on knockout reactions and compare to high energy data . in order to identify the separate contribution of these two factors , we do not vary the geometry of the nucleon binding potentials used to calculate the single particle radial wave functions . that was identified in the literature as another major factor in the calculations , leading to large variations in the extracted spectroscopic factors . _ bottom and left scale : _ radial wave functions in arbitrary units ( solid curves ) for the valence nucleon - core system and for a few representative reactions considered in this work . we have taken only one configuration in cases of systems with multiple configurations . ] c,@xmath31b)x at 35 mev / u as a function of target t. the solid ( dashed ) curves are calculations with ( without ) coulomb distortion . _ lower panel _ : same as above , but for @xmath27c target and for different beam energies . ] the relevance of medium corrections are motivated by the effect summarized in figure [ fig01 ] . the cases are shown in figure [ fig01 ] are for @xmath27c(@xmath52c,@xmath31b ) at 35 mev / u , @xmath25be(@xmath53o,@xmath30n ) at 56 mev / u , @xmath27c(@xmath35al,@xmath33 mg ) at 50 mev / u and @xmath25be(@xmath54be,@xmath29be ) at 60 mev / u . later we will discuss more results for each system separately . the dashed and dotted curves show the ratio between the average in - medium and the free nucleon - nucleon cross section as a function of impact parameter ( see top scale ) . we define the average nucleon - nucleon cross section at the distance of closest approach between the projectile and the target using eq . and the definition @xmath55 where @xmath56 is the impact parameter vector , perpendicular to the beam axis . in figure [ fig01 ] , the dashed and dotted curves are for core - target and valence nucleon - target average nucleon - nucleon cross sections , respectively . notice that the target center of mass is located on the right of the top axis scale . also shown in the figure are the radial wave functions ( solid curves ) for the valence nucleon - core system and for a few representative reactions considered in this work . for simplicity , in this figure we have used only one of the main configurations for the projectile ground state ( more detailed and complete calculations are reported later on in this section ) . the binding energies effectively " decrease from the top to low panels . we mention effectively " because , although the binding energy of the valence proton in @xmath35al is smaller than for the valence neutron in @xmath54be , the coulomb barrier creates an effectively larger binding in @xmath35al . it is clearly noticeable from figure [ fig01 ] that the wavefunctions of weakly bound systems extend far within the target where the nucleon - nucleon cross sections are strongly modified by the medium . we have to emphasize that the shaded areas in fig . [ fig01 ] are relevant to stress the importance of medium effects at surface region since the reaction is peripheral due to strong absorption at @xmath57 . momentum distributions and nucleon removal cross sections in knockout reactions are thus expected to change appreciably with the inclusion of medium corrections of nucleon - nucleon cross section . such corrections are also expected to play a more significant role for loosely - bound systems . in the following , we discuss coulomb corrections . here we consider the simplest and most straightforward correction one can do , namely the inclusion of a coulomb phase , which accounts for the distortion of the elastic scattering of the core fragment . it has been usually taken for granted that longitudinal momentum distributions are little affected by elastic scattering of the core fragment , the reason being that the longitudinal forces acting on the core fragment reverse sign as the projectile passes by the target , leading to a reduced distortion effect @xcite . further , as has been shown in ref . @xcite , the transverse momentum distributions are strongly affected by both nuclear and coulomb elastic scattering . for heavier targets the distortions are predominantly due to coulomb repulsion @xcite . it is worthwhile mentioning that the implications of the findings on coulomb distortion effects presented in ref . @xcite have been neglected in the literature . in order to avoid dealing with the effects of the coulomb scattering , experiments are usually performed with light targets , such as @xmath58be and relatively high energies , @xmath59 mev / nucleon . in this work we show that these argumentations are not always valid and need to be studied with care . as discussed in the previous section , in the presence of the coulomb field the eikonal s - matrices factorize as the product of the nuclear and the coulomb contributions : @xmath60 . although this does not make any difference for the total stripping cross sections ( see eq . ( 20 ) of ref . @xcite ) , it has an impact on the diffraction dissociation cross section ( through the second term of eq . ( 21 ) of ref . this means that not only transverse , but also longitudinal momentum distributions will be affected by the coulomb field . this is shown in figure [ fig02 ] for the longitudinal momentum distributions of several systems which we will consider in details later in this section . it is evident from the upper panels of this figure that longitudinal momentum distributions in knockout reactions t(@xmath52c,@xmath31b)x ( and their total cross sections ) are strongly influenced by the coulomb field of the target t at bombarding energies of 35 mev / nucleon . the solid ( dashed ) curves are calculations with ( without ) the inclusion of coulomb scattering . it is also evident that even for the case of light targets , such as @xmath58be and @xmath61li , the distributions change appreciably . the lower panels show calculations for the same reaction , but for @xmath27c targets and as a function of the bombarding energy . it is clear that distortions are important even for usually considered safe " energies , such as 100 mev / nucleon . c(@xmath52c,@xmath31b ) ( solid curve ) and @xmath27c(@xmath35al,@xmath33 mg ) ( dashed curve ) at 35 mev / u . we artificially vary the separation energy @xmath62 of the proton in @xmath52c and in @xmath35al . the dotted curve shows the calculation for @xmath27c(@xmath35al,@xmath33 mg ) at 50 mev / u . ] we found that the effect of coulomb scattering is relatively larger for systems with smaller sizes . this is illustrated in figure [ fig03 ] where we present our calculations for the total nucleon removal cross sections for the reactions @xmath27c(@xmath52c,@xmath31b ) ( solid curve ) and @xmath27c(@xmath35al,@xmath33 mg ) ( dashed curve ) at 35 mev / u . we artificially vary the binding energy of the proton in @xmath52c and in @xmath35al . as the separation energy increases so does the percent difference of the cross section , @xmath63 , where @xmath64 is the cross section without the coulomb scattering phases . with increasing separation energy the relative valence nucleon - core distance decreases and the nucleon removal cross section decreases , but the relative importance of the coulomb scattering increases . for very small energies the effects of coulomb dissociation ( not considered here ) should also become relevant and increase the magnitude of the cross sections . the relative importance of the coulomb scattering for the removal cross sections decreases with the bombarding energy . this is shown in figure [ fig03 ] with the calculation for @xmath27c(@xmath35al,@xmath33 mg ) at 50 mev / u ( dotted curve ) . these preliminary discussions support our conclusion that both medium effects and coulomb distortion play a relevant role in knockout reactions . next we consider a series of published data for which neither medium or coulomb corrections were included . we thus quantify the changes in the extracted values of spectroscopic factors in case these effects were to be included in the experimental analysis . recently , the @xmath27c(@xmath35al,@xmath33mg)x knockout reaction has been studied at 50 mev / nucleon to investigate the ground state properties of @xmath35al @xcite . it was shown that the ground - state structure of @xmath35al is a configuration mixing of a _ d_-orbital valence proton coupled to four core states of @xmath33 mg - 0@xmath65 , 2@xmath66 , 4@xmath66 , 4@xmath67 . the ground state spin and parity of @xmath35al as @xmath68 has been confirmed . this experiment had the advantage that exclusive measurements were done and momentum distributions were determined for the four major configurations in the ground state of the projectile ( @xmath35al ) . c(@xmath35al,@xmath33mg)x at 50 mev / nucleon . the solid line has both coulomb and medium corrections . the dashed - curve has no medium corrections . the dashed - dotted line includes calculations without coulomb corrections . the dotted curve neither includes medium effects nor coulomb corrections . ] in this work , we have analyzed the @xmath27c(@xmath35al,@xmath33mg)x system to check the relevance of coulomb and medium effects . the @xmath69 wave functions for the valence proton were generated in a spherical woods - saxon ( ws ) potential with the parameters given in table [ bound ] . in the optical limit of the glauber theory and the t-@xmath41 approximation ( explained in detail in refs . @xcite ) , the eikonal phase is obtained from the input of the nuclear ground state densities and the energy dependent nucleon - nucleon cross sections . the ground state density parameters and models used in this work are shown in table [ grdens ] , and our results are presented in fig . [ fig04 ] and table [ t_preal23 ] . [ cols="^,^,^,^,^,^,^ " , ] it is evident from fig . [ fig01 ] that @xmath52c has the smallest effective " size and that @xmath54be has the biggest size among the low energy systems in this study . the nuclear size is important for low energy cases because the diffraction dissociation becomes dominant when the nuclear size is smaller , but the stripping becomes dominant when the nuclear size is bigger . the reason for this is that a large projectile feels the nuclear interaction already at large impact parameters . a small projectile can come closer to the target where the coulomb interaction is larger . the evidence of this can be seen in table [ cross ] . it is thus clear why medium and coulomb corrections are more important in the @xmath58be(@xmath54be,@xmath29be ) and the @xmath27c(@xmath52c,@xmath31b ) cases , respectively . the momentum distribution of the one - neutron removal residues from the @xmath27c(@xmath70o,@xmath35o)x reaction was measured for the first time at 920 mev / nucleon and reported in ref . the data could be explained with a spectroscopic factor @xmath62=1.74(19 ) of an almost pure 2@xmath71 single - particle state for the valence neutron . this work , together with recent theoretical calculations , suggests that @xmath70o is a newly discovered doubly magic nucleus . the one - neutron removal cross section was found to be 63(7 ) mb . the calculations in ref . @xcite were based on a few - body glauber formalism @xcite for two configurations : ( a ) @xmath35o@xmath72 in @xmath73 orbital and ( b ) @xmath35o@xmath74 in @xmath69 orbital . the wave functions for the configurations are obtained with a woods - saxon potential by adjusting the depth of the potential to reproduce the one - neutron separation energy @xmath75=3.61(27 ) mev @xcite . using a pure 2@xmath71 configuration with @xmath62=1 leads to a cross section of 34 mb . the calculation is in agreement with the data when it is multiplied by @xmath62=1.74(19 ) . this large spectroscopic factor indicates that the single - particle strength of the valence neutron is strongly weighted in the 2@xmath71 state . in the present work we have reproduced the data of ref . @xcite also by assuming a 2@xmath71 orbital only . the potential parameters for the bound state wave function are given in table [ bound ] and the ground state density for the @xmath35o@xmath76 core is obtained using liquid droplet model ( ldm ) densities @xcite , as indicated in table [ grdens ] . to understand the differences between medium effect models , four different calculations including coulomb corrections have been made for this system . the calculated one - neutron removal cross sections are 58.58 mb , 54.08 mb , 78.74 mb and 53.25 mb using free @xcite , pauli corrected ( eq . 2 ) , brueckner ( eq . 3 ) and phenomenological parameterizations @xcite of the nucleon - nucleon cross sections , respectively . except for the result obtained with the brueckner theory , they are all in agreement with the previous work and with the data . the relative difference between the results obtained using brueckner corrections and with free nucleon - nucleon cross sections is about 34% . however , we do not consider a real discrepancy , as the brueckner parametrization have been extrapolated well beyond their validity . brueckner calculations are limited by the pion - production threshold , and should only be valid for projectile energies below 300 mev / nucleon . thus we verify that the experimental data for the reaction @xmath27c(@xmath70o,@xmath35o)x at 920 mev / u is well reproduced with the use of free nucleon - nucleon cross sections . the changes introduced by pauli - blocking with the geometric model are small , and the phenomenological account of medium effects at this high energy also basically agree with the results using free cross sections . c(@xmath70o,@xmath35o)x reaction at 920 mev / nucleon . the curves are calculated with the free nn cross sections ( solid ) , with a geometrical account of pauli blocking ( dashed ) , a phenomenological fit from ref . @xcite ( dotted ) , and a correction from brueckner theory ( dashed - dotted ) . the data has been taken from ref . @xcite . ] the ground state structure of @xmath77 mg , a nucleus belonging to the @xmath78 island of inversion , has been studied in ref . @xcite by means of nucleon - removal reactions on a carbon target at 898 mev / nucleon . the longitudinal momentum distribution of the @xmath36 mg core was measured and the one - neutron removal cross section was found to be 74(4 ) mb . most of the contribution to the ground state structure of @xmath77 mg was shown to arise from the @xmath79 orbital . the longitudinal momentum distribution obtained in ref . @xcite can not be reproduced with a pure single particle state . it has been discussed in details in ref . @xcite the reason why a configuration mixing of different single - particle states is needed . two different configuration mixings for the ground state of @xmath77 mg were assumed . the first one is @xmath80 and the second is @xmath81 where @xmath82 are the spectroscopic amplitudes for each single - particle orbital . the values of the corresponding spectroscopic factor @xmath83 were found by @xmath84 minimization and their values for the second configuration are @xmath85 , @xmath86 , @xmath87 and @xmath88 @xcite . c(@xmath77mg,@xmath36mg)x system at 898 mev / nucleon . the data has been taken from r. kanungo _ the curves are calculated with the free nn cross sections ( solid ) , with a geometrical account of pauli blocking ( dashed ) , a phenomenological fit from ref . @xcite ( dotted ) , and a correction from brueckner theory ( dashed - dotted ) . ] in our calculations we have chosen the second configuration set used in ref . @xcite , eq . [ sconf ] , since the @xmath77 mg ground state is usually accepted to be @xmath89 . we apply the same procedure as described before to obtain bound state wave functions and eikonal phases . the parameters for the bound state potentials and ground state densities are shown in table [ bound ] and table [ grdens ] , respectively . we have used nearly the same spectroscopic factors within the error bar range of ref . @xcite to make a consistent comparison of the medium effects . our results yield a small but relevant variation of the one - neutron removal cross sections using the free , pauli corrected , and phenomenological nn cross sections , namely 83.70 mb , 77.90 mb , and 77.63 mb , respectively . as observed in the case of the @xmath27c(@xmath70o,@xmath35o)x at 920 mev / u , the use of brueckner corrected nn cross sections yields 112.92 mb , about 35% relative to the calculations using the free nn cross sections . for the same reason as with the previously considered reaction , this discrepancy is meaningless as one extrapolates the brueckner results beyond their regime of validity . both reactions considered above are very illustrative as they show a great consistency between the calculations performed by different authors , with somewhat different methods . they also show the expected relevance of medium corrections at intermediate and low energy collisions . as we mentioned above , medium effects have been routinely neglected in the experimental analysis of knockout reactions . but their relevance has been known for a long time in the analysis of elastic and inelastic scattering , as well as of total reaction cross sections @xcite . the effects are larger at lower bombarding energies , where pauli blocking strongly reduces the nucleon - nucleon cross sections in the medium . a systematic study of these effects has been presented in ref . @xcite . c taken from ref . the curves are calculated with the free nn cross sections from ref . @xcite ( solid ) , with a geometrical account of pauli blocking ( dashed ) , a phenomenological fit from ref . @xcite ( dotted ) , and a correction from brueckner theory ( dashed - dotted ) . the triangle - dotted curve is calculated with the same free nn cross sections from ref . @xcite , but with an another hfb calculation @xcite for the @xmath27c ground state density . ] to corroborate these statements , in figure [ fig10 ] we show the data on p + @xmath27c reaction cross sections taken from the ref . @xcite in the energy region of our interest , 20 - 100 mev / nucleon . the cross sections were calculated from the relation @xmath90,\ ] ] where @xmath91 has been calculated using eqs . ( [ sib],[eikphase],[fnn ] ) and the carbon matter density from a hartree - fock - bogoliubov calculation @xcite . several distinct calculations are shown . the solid curve uses eq . with the free nucleon - nucleon cross sections and the carbon matter density from a hfb calculation @xcite , whereas the triangle - dotted curve ( the triangles are not data , but used for better visibility ) uses a different hfb density @xcite , consistent with the calculations presented in ref . @xcite . as expected , that the agreement between the two calculations is very good . the other curves in figure [ fig10 ] show the same calculation procedure , but including medium corrections for the nucleon - nucleon cross section . the results are evidently very different than the previous ones . the dotted ( dashed - dotted ) [ dashed ] curves use phenomenological ( brueckner ) [ pauli geometrical ] recipes for medium effects on the cross sections . based on the large error bars and spread of the experimental data , it is hard to judge what model adopted for medium corrections yields the best agreement with the data . it is clear that the inclusion of medium effects change the results to yield a closer reproduction of the data . these findings are in agreement with our present understanding of medium modifications of the reaction cross sections and of several other reaction channels involving heavy ion scattering at intermediate energies ( @xmath92 mev / nucleon ) . often neglected effects , such as medium modifications of the nucleon - nucleon cross sections and coulomb distortion , modify appreciably the nucleon knockout cross sections . as we have shown , these effects do not lead to an appreciable modification of the shapes of momentum distributions . this is explained by the fact that the momentum distributions are largely the fourier transforms of the contributing parts of the single - particle wavefunctions , overwhelmingly their asymptotic regions , which are the whittaker functions for protons or the hankel functions for neutrons , sensitive only to the orbital momentum and the nucleon binding energies . we have shown these features explicitly by comparing our results with a large number of available experimental data . as expected on physics grounds , these corrections are larger for experiments at lower energies , around 50 mev / nucleon , and for heavy targets . as more experiments make use of heavier targets , it is worthwhile to illustrate the relevance of coulomb corrections . medium effects in knockout reactions have also been frequently ignored in the past . we show that they also have to be included in order to obtain a better accuracy of the extracted spectroscopic factors . although these conclusions might not come as a big surprise , they have not been properly included in many previous experimental analyses . this work was partially supported by the us - doe grants de - sc004972 and de - fg02 - 08er41533 and de - fg02 - 10er41706 , and by the research corporation . m. karako has also been partially supported by the tubitak grant 109t373 . a. gade , p. adrich , d. bazin , m. d. bowen , b. a. brown , c. m. campbell , j. m. cook , t. glasmacher , p. g. hansen , k. hosier , s. mcdaniel , d. mcglinchery , a. obertelli , k. siwek , l. a. riley , j. a. tostevin , d. weisshaar , phys . c 77 ( 2008 ) 044306 .	background : : one - nucleon removal reactions at or above the fermi energy are important tools to explore the single - particle structure of exotic nuclei . experimental data must be compared with calculations to extract structure information , evaluate correlation effects in nuclei or determine reaction rates for nuclear astrophysics . however , there is insufficient knowledge to calculate accurately the cross sections for these reactions . purpose : : we evaluate the contributions of the final state interaction ( fsi ) and of the medium modifications of the nucleon - nucleon interactions and obtain the shapes and magnitudes of momentum distributions . such effects have been often neglected in the literature . method : : calculations for reactions at energies 35 - 1000 mev / nucleon are reported and compared to published data . for consistency , the state - of - the - art eikonal method for stripping and diffraction dissociation is used . results : : we find that the two effects are important and their relative contributions vary with the energy and with the atomic and mass number of the projectile involved . conclusions : : these two often neglected effects modify considerably the one - nucleon removal cross sections . as expected , the effect are largest at lower energies , around 50 mev / nucleon and on heavy targets .
in the constellation of the brilliant elementary particle experiments of the last decades , the b - factories , belle @xcite and babar @xcite , were ones of the most remarkable . their results on cp - violation in b - mesons prove the validity of the km mechanism , which inventors , m. kobayashi and t. maskawa , won the nobel prize in physics in 2008 . belle was successful not only in cp - violation studies . the huge amount of the data collected by the belle detector ( more than one billion @xmath1-pairs ) allows to make a number of discoveries in hadron spectroscopy . more than 10 new quarkonium - like states were discovered including puzzling x(3872 ) state and new charged z - resonances which do not fit conventional quarkonium model . the belle successor , belle ii experiment @xcite , is aimed to collect statistics sample of 50 ab@xmath2 by the end of 2022 . major accelerator upgrade includes changes of the crossing angle ( 22 mrad to 83 mrad ) , introduction of the nano - beam scheme , _ i.e. _ decreasing of horizontal emittance ( @xmath3 nm to @xmath4 nm for positrons and @xmath5 nm to @xmath6 nm for electrons ) and beta - functions at ip ( @xmath7 from @xmath8 mm to @xmath9 mm ) , rising beam currents ( from @xmath10 a to @xmath11 a for positrons and from @xmath12 a to @xmath13 a for electrons ) . the downside of these changes is significant rise of backgrounds . higher luminosity delivered by the accelerator requires adequate upgrade of the detector components to sustain effective operation at new environment . new belle ii detector have 2-layer depfet pixel ( pxd ) and 4-layer dssd silicon strip ( svd ) detectors for vertexing . tracking is done with the upgraded drift chamber ( cdc ) with increased size and smaller cells . particle identification is done with time of propagation counter ( top ) combining advantages of the excellent time resolution time - of - fight system and erenkov ring imaging babar - like dirc in barrel region and aerogel rich ( arich ) detector with focusing radiator in endcaps . electromagnetic calorimeter with new electronics for barrel and pure csi crystals instead of csi(tl ) in endcaps is able to operate at much higher backgrounds . muon and @xmath0 system ( klm ) is replaced with new one based on scintillator technique in endcaps and in two innermost barrel layers . the @xmath0 and muon detector ( klm ) of the belle detector consists of an alternating sandwich of @xmath14-cm thick iron plates and active detector elements located outside of the superconducting solenoid . the octagonal barrel covers the polar angle range from @xmath15 to @xmath16 , while the endcaps extend this coverage from @xmath17 to @xmath18 . there are 14 detector layers and 14 iron plates in each endcap . each layer operates independently . the belle klm is based on resistive plate chambers ( rpc ) technology . high voltage is distributed over the two glass electrodes with bulk resistivity of @xmath19 but transparent to the fast electric pulses . the throughgoing charged particle ionizes gas along its track and form streamer discharge in the gas - filled gap between electrodes . the signal induced on the couple of orthogonal pick - up strip electrodes is large enough to be discriminated without amplification . in the belle environment the performance of the most of the barrel layers was about @xmath20 decreasing down to @xmath21 for the innermost one . for the endcap efficiency ranges between @xmath22 and @xmath23 . the lowest numbers are for the outermost layers where the background is higher and came mostly from the accelerator tunnel . since at high background loads of belle ii experiment efficiency of the rpcs drops to zero , new belle ii endcap klm system ( eklm ) utilizes another technique based on scintillator strips read - out by wavelength shifting ( wls ) fibers . the light is registered by novel devices silicon photomultipliers ( sipm ) . the base element of the new system is scintillator strip of polystyrene doped with ptp ( @xmath24 ) and popop ( @xmath25 ) produced by extrusion technique at `` uniplast '' enterprise ( vladimir , russia ) . it has a cross - section of @xmath26 mm@xmath27 and length varying from @xmath28 mm to @xmath29 mm . in total there are 41 different strip lengths . diffusive reflective cover of the surface is produced by chemical etching . a @xmath30 mm wide and @xmath31 mm deep groove is milled in each strip to hold the wls fiber . an optical contact between the scintillator and kuraray y11 wls fiber put into the groove is provided by optical gel sl-1 by surel ( st .- peterburg , russia ) . one wls fiber end is connected to a photosensor while the other is mirrored using a silver - shine paint . the groove is covered by aluminized reflective tape . the schematic view of the strip is presented in fig . [ strip ] . the photosensor is hamamatsu sipm ( s10362 - 13 - 025ckx ) . it is pressed against the polished fiber end by the rubber ring in the connector . sipms are connected to the preamplifiers by twisted - pair cables . for the sake of efficient handling strips are glued together into segments of 15 strips . a @xmath32 mm plastic plates are glued to the strips at both segment s sides with double - sided sticky tape . the whole eklm system consists of 112 identical modules . each module contains 150 strips ( 10 segments ) in two identical orthogonal planes , preamplifiers for these strips and is enclosed into aluminum box . the segments supported by aluminum i - beams that run along the segment boundaries . the i - beams supporting the @xmath33- and @xmath34-plane segments are glued into a supporting grid . the schematic view of the module is shown in fig . [ sector ] . ( 1684,0 ) ( 225,-43)(-2,0)43 ( 225,-43)(0,-2)43 ( 225,-43)(-1,-1)20 ( 225,-43)(-4,1)65 ( 225,-43)preamplifier boards ( 255,-105)(-1,-1)45 ( 255,-105)(-1,0)47 ( 255,-105)(-1,-2)45 ( 255,-105)support grid to control strip quality during the production process each strip is subjected to cosmic ray test , while the light yield caused by cosmic muons passing trough the strip is measured . the triggers are located at the middle of the strip and at both its ends . the results of the tests are shown in fig . [ lightyield ] . the quality of the strips is found to be very good . the number of photo electrons from the end , nearest to the sipm , is shown by solid curve . only strips with light yield more than @xmath35 p.e . are accepted for the further production . this number is almost @xmath32 times larger than original goal @xcite . the mass production of the detector components have been started in itep in september 2012 . till the end of the september 2013 more than @xmath36 is complete : 14075 strips are produced and tested , 13808 successfully passed the cosmic test ; 875 segments are assembled . three ( out of four ) batches of the detector components are sent to kek , the last one will be sent in november . the work is supported by the russian ministry of education and science contracts 1366.2012.2 , and 14.a12.31.0006 ; and president grant mk-2877.2012.2 .	the belle detector operated at kekb b - factory in 1999 - 2010 was one of the most remarkable experiments in the field of elementary particle physics of the last decades . the belle successor , belle ii collaboration , is aimed to operate the belle ii detector at superkekb factory at 40 times higher luminosity . increased luminosity imposes new requirements on the detector elements : they have to survive at higher radiation levels , to operate at higher loads and at higher backgrounds . the belle @xmath0 and muon system based on the resistive plate chambers ( rpc ) technology worked well during all data taking period , however at belle ii environments its performance decreases to negligible level due to increasing load and high neutron background . to sustain detector operation it will be replaced by the new system based on the scintillation strips read - out by silicon photomultipliers . the latter technology allows not only to reach time resolution at level of 1 ns but also perform the amplitude measurements . nowadays the production of the new eklm system s elements are under way . the assembly at kek is started this fall .
the rather extensive literature concerning the `` inverse scattering problem '' and the `` inverse sturm - liouville problem '' on graphs have mostly followed separate pathways except for a very few results @xcite . in the following paragraphs , we briefly recall the previous results on these subjects and at the end we will situate the result of this paper with respect to the others . indeed , as it will be seen later , the inverse sturm - liouville problem considered in this paper raises from the necessity of finding a minimal setup for solving the inverse scattering problem . the paper @xcite considers a star - shape graph consisting of @xmath0 infinite branches and solves the inverse scattering problem assuming the measurement of @xmath1 reflection coefficients . next , in the paper @xcite , harmer provides an extension of the previous result with general self - adjoint boundary conditions at the central node . this however necessitates the knowledge of @xmath0 reflection coefficients . + the paper @xcite studies the relation between the scattering data and the topology of the graph . they show that the knowledge of the scattering matrix is not enough to determine uniquely the topological structure of a generic graph . in @xcite , avdonin and kurasov consider again a star - shape graph with @xmath0 finite branches . they , show that the knowledge of one diagonal element of the response operator allows one to reconstruct the potential on the edge corresponding to this element . as mentioned above , in parallel to the research on inverse scattering problems , another class of papers consider the inverse problem for sturm - liouville operators on compact graphs . these results can be seen as extensions of the classical result provided by borg @xcite , on the recovering of the sturm - liouville operator from two spectra on a finite interval . the main progress in this field has been made by yurko @xcite , @xcite and @xcite . the paper @xcite deals with the inverse spectral problem on a tree . the idea is to generalize the borg s result in the following sense : for a tree with @xmath2 boundary vertices , it is sufficient to know @xmath2 spectra , corresponding to @xmath2 different settings for boundary conditions at the extremities , to retrieve the potentials on the tree . in the recent paper @xcite , the same kind of result is proposed for a star - shape graph including a loop joined to the central node . finally , @xcite provides a generalization of @xcite to higher order differential operators on a star - shape graph . + in @xcite , the author proves that under some restrictive assumptions on the spectrum of a sturm - liouville operator on a star - shape graph with some fixed boundary conditions , the knowledge of this spectra can determine uniquely the sturm - liouille operator . in this paper , we consider the inverse scattering problem motivated by the application in fault - detection / diagnostic of star - shape lc transmission networks . we are interested in minimal experimental setting providing enough information on the network and the potentials on its branches . the graph consists of @xmath0 finite branches joined at a central node and we add a infinite branch to this central node for the experimentation . we will see that the knowledge of the reflection coefficient is equivalent to the knowledge of the spectra for the sturm - liouville operator defined on the compact part of the graph for various boundary conditions at the central node . by considering the associated inverse sturm - liouville problem , we will show that under some assumptions on the geometry of the metric graph , the knowledge of only two reflection coefficients , corresponding to two settings for the boundary conditions at the terminal nodes , is enough to determine the potentials ( at least locally ) . in the next section , we explain the application under study and the associated experimental setup . we will show that the inverse scattering problem for the so - called telegrapher s equation ( lossless case ) on the network is equivalent to an inverse scattering problem for a schrdinger operator over the metric graph of the network . in section [ sec : direct ] , we will consider the direct scattering problem and we will characterize the reflection coefficient in terms of the fundamental solutions for sturm - liouville operators on branches . in section [ sec : geometry ] , we will show that the knowledge of only one reflection coefficient is enough to identify the lengthes of the branches of the metric graph . this result will be useful to locate hard faults ( open or short circuits ) . finally , in section [ sec : potential ] we consider the main problem of recovering the potentials from the knowledge of one or two reflection coefficients . this result will be useful to locate soft faults ( local variations of the electrical characteristics).we will prove the equivalence of the inverse scattering problem with an associated inverse sturm - liouville problem on the compact part of the graph . this inverse problem is then treated applying the methods extending the classical result by borg @xcite . the electric signal transmission through a wired network is , generally , modeled with the `` telegrapher s equation '' and characterized by the parameters @xmath3 ( functions of the space position @xmath4 along the transmission lines ) representing the inductance @xmath5 , capacitance @xmath6 , resistance @xmath7 and loss conductance @xmath8 per unit of length . these parameters allow a rather complete and understandable description of the transmission lines and are sufficient to represent the lines in the frequency range used during reflectometry . in the sequel we will suppose that this model can be used for all frequencies . however , it appears to be impossible to retrieve all these parameters uniquely through the information provided by reflectometry experiments . everywhere , through this paper , we will consider the simpler nonuniform lossless situation ( @xmath9 ) . as we will see later , the reflectometry measurement is still not enough to retrieve the both parameters @xmath5 and @xmath6 but rather an aggregate of these two parameters , the local characteristic impedance @xmath10 . following @xcite and @xcite , the presentation of the reflectometry experiment of this section , is based on a model derived from the `` telegrapher s equation '' and parameterized by @xmath11 . to cope with the network case , we have translated the kirchhoff rules at the nodes of the network within this new modeling framework . note that , in this paper , we restrict ourselves to the case of a simple star - shape network and therefore the only node of the graph where the kirchhoff rules need to be adapted is the central one . the faults , in which we are interested through this approach , are represented by the lengths of the branches ( hard faults ) and by the heterogeneities of @xmath11 along the branches ( soft faults ) . the considered reflectometry experiment model is based on a far - field method consisting in adding a uniform infinite wire joined to the network at its central node . in practice , connecting a matched charge to the extremity of a finite line , is sufficient to emulate the electrical propagation through an infinite line . the linearity of the transmission line model allows to replace any test by an equivalent test in harmonic regime . we can therefore start by stating the telegrapher s equation in the harmonic regime , i.e. the tension and the current intensity are respectively of the form @xmath12 and @xmath13 , where @xmath14 is the time frequency and @xmath4 the position . on each line , we have @xmath15 through this paper , we assume that * a1 * : : the distributed parameters @xmath16 and @xmath17 are twice continuously differentiable on the transmission lines ; * a2 * : : they are strictly positive , @xmath18 ; * a3 * : : the characteristic impedance @xmath10 is continuous at the central node of the star - shape network ; * a4 * : : the transmission lines are uniform in a neighborhood of the extremities of the branches . _ the liouville transformation . _ note that the reflectometry experiment leads to observing the tensions and currents along the time at some position : only the travelling times ( and amplitudes ) of waves are accessible by such experiments . a fault can only be localized in terms of the traveling time of the reflected test wave starting from the test point . this leads to a particular change of variables , the liouville transformation , allowing to work with the traveling time rather than spatial coordinates . let us recall this transformation : @xmath19 which corresponds to the wave traveling time from the position @xmath20 to the position @xmath4 . remark that after this transformation , @xmath14 is also the wave number on each branch . the inverse transformation being well defined , we will write @xmath21 , @xmath22 , @xmath23 and @xmath24 . the telegrapher s equation becomes @xmath25 _ the wave decomposition and equivalent forms of the telegrapher s equation . _ define @xmath26 and @xmath27 . we have the following decomposition of @xmath28 : @xmath29 in particular , in an interval where a branch is uniform , @xmath30 , and the solution is the sum of waves of opposite directions . for any @xmath31 and @xmath32 in this interval : @xmath33 define now @xmath34 and @xmath35 , with @xmath36 . the telegrapher s equation becomes a schrdinger equation : @xmath37 it can be seen from and , that , the knowledge of the potential @xmath38 or of @xmath39 and of the boundary conditions on @xmath28 and @xmath40 is sufficient to compute the solution of on the network . in our lossless situation , we have chosen @xmath41 as the parameter to be identified through the reflectometry experiment . a variant of , the zakharov - shabat equations , would be the good choice in the more general lossy case . remark that @xmath42 . _ the reflection coefficient . _ taking @xmath43 , oriented toward the increasing @xmath32 , as a reference forward wave , the reflection coefficient is the following ratio of backward over forward wave amplitudes : @xmath44 . in particular , @xmath45 : defined in this way , the reflection coefficient is constant in intervals where @xmath11 is constant . for an arbitrary @xmath32 , denoting by @xmath46 the ( possibly infinite ) apparent impedance at @xmath32 , we still define the reflection coefficient as being @xmath47 with this definition , in general , @xmath48 , and , using , it is easy to check that @xmath7 is solution of the following riccati equation : @xmath49 finally , note that , we will only consider `` positive real '' terminal impedances @xmath50 , in the sense that @xmath51 and @xmath52 . this together with the fact that @xmath53 is positive implies that @xmath54 . furthermore , experiments with @xmath55 are sufficient . _ the network under test . _ throughout this paper , @xmath56 represent the compact star - shape network consisting of the branches @xmath57 joining at the central node and @xmath58 is the extended graph where the test branch @xmath59 is also added to the graph . we have @xmath60 equations of the form @xmath61 where @xmath62 is the wave traveling time associated to the branch number @xmath63 ( @xmath64 as the added branch @xmath59 is assumed to be an infinite line ) . in particular note that , as the infinite branch @xmath59 is assumed to be a uniform transmission line , we have @xmath65 . _ the boundary condition for the reflectometer . _ consider now that a generator with a matched internal impedance @xmath11 and an electromotive force @xmath66 is connected in @xmath31 of an interval @xmath67 where @xmath11 is constant . we have @xmath68 which can be the boundary condition if the branch terminates at @xmath31 . we have @xmath69 and @xmath70 for all @xmath71 . as , @xmath45 , the reflection coefficient can then be determined from the measurement of the tension @xmath72 anywhere in @xmath67 . in the sequel , we will use a test branch @xmath59 with a constant @xmath11 connecting a matched generator to the central node . the measured reflection coefficient on this branch , will be simply written @xmath73 . as we will work with @xmath74 , we choose @xmath75 , so that @xmath76 . finally , it will be convenient to take the same positive orientation on all the branches , from the central node at @xmath77 toward the increasing @xmath32 . our reference forward wave on @xmath59 is then in the direction of the decreasing @xmath32 , so that , changing @xmath32 into @xmath78 , and supposing @xmath59 of infinite length , the boundary condition for the reflectometer is : @xmath79 _ the two sets of boundary conditions at the network extremities . _ in order to recover the potential of the star - shape network , we will need to consider two experimental settings , with open circuit or short circuit at the extremities of the branches . this will lead to a problem similar to solving an inverse spectral problem for the sturm - liouville operator when two spectra are known . the first setting corresponds to open circuit configuration at the extremities of the finite branches ( @xmath57 ) . this , together with the assumption * a4 * on the local uniformity of the lines around @xmath62 s , leads to boundary conditions of the form @xmath80 , or equivalently , we obtain the setting called , the neumann configuration : @xmath81 the second setting corresponds to the short circuit configuration at the extremities of the finite branches ( @xmath57 ) . this leads to boundary conditions of the form @xmath82 , or equivalently , we obtain the setting called , the dirichlet configuration : @xmath83 [ rem : riccati ] in some of the applications that we have in mind , the reflectometry experiment has to take place without perturbing significantly the normal utilization of the transmission network , so that using open or short circuits conditions would be impossible . there is a way to circumvent this problem by computing the results of the open or short circuit experiments from results of two less invasive experiments . the idea is to use nonlinear superposition properties of solutions of riccati equations as in @xcite , in order to get a closed - form representation of the reflection coefficient , solution of , as a function of a general load impedance ( value of @xmath84 at the extremity of a branch ) and of two particular solutions corresponding to two load impedances more compatible with the network utilization . this will be presented in a forthcoming paper . _ the boundary condition at the central node . _ it writes @xmath85 where @xmath86 and @xmath87 denote the spatial derivatives at the point @xmath77 and @xmath88 is the characteristic impedance of the branch number @xmath63 . note , in particular , that we have applied the continuity of @xmath88 s at the central node ( assumption * a3 * ) : @xmath89 , @xmath90 . _ formulation of the model . _ in conclusion , in order to study the @xmath91-transmission line equations on the graph @xmath58 , we can study the schrdinger operators @xmath92 where @xmath93 ( resp . @xmath94 ) denotes the space of infinitely differentiable functions @xmath95 defined on @xmath58 satisfying the boundary conditions @xmath96 the operators @xmath97 are essentially self - adjoint . to prove this fact we observe first that these operators are a compact perturbation of the operators @xmath98 with the same boundary conditions . now , we apply a general result by carlson @xcite on the self - adjointness of differential operators on graphs . indeed , following the theorem 3.4 of @xcite , we only need to show that at a node connecting @xmath99 edges , we have @xmath99 linearly independent linear boundary conditions . at the terminal nodes of @xmath100 this is trivially the case as there is one branch and one boundary condition ( dirichlet or neumann ) . at the central node it is not hard to verify that define @xmath60 linearly independent boundary conditions as well . this implies that the operators @xmath97 are essentially self - adjoint and therefore that they admit a unique self - adjoint extension on @xmath101 . + we are interested in the scattering solution where a signal of frequency @xmath14 is applied at the infinite extremity of the infinite branch . in such a case , we will be seeking a solution satisfying the asymptotic behavior @xmath102 [ prop : scattering ] under the assumptions * a1 * through * a4 * , there exists a unique solution @xmath103 continuous with respect to @xmath14 , of the scattering problem , satisfying * @xmath104 for @xmath105 ; * @xmath106 satisfy the boundary conditions and or ; * for each @xmath107 , there exist @xmath73 such that @xmath108 we will denote the reflection coefficient @xmath73 defined by in the neumann ( resp . dirichlet ) case by @xmath109 ( resp . @xmath110 ) . this coefficient appears to be unique . this proof gives us a concrete method for obtaining scattering solutions . indeed , we will propose a solution and we will show that it is the unique one . + in this aim , we need to use dirichlet / neumann fundamental solutions of a sturm - liouville boundary problem . [ def : fund ] consider the potentials @xmath111 as before and extend them by 0 on @xmath112 so that they are defined on the intervals @xmath113 $ ] . the dirichlet ( resp . neumann ) fundamental solution @xmath114 ( resp . @xmath115 ) , is a solution of the equation , @xmath116 consider , now , the function @xmath117 where @xmath118,~~j=1,\cdots , n.\ ] ] here the coefficients @xmath119 and @xmath120 are given by the boundary conditions at the central node : @xmath121one , easily , sees that this @xmath122 is in @xmath123 , the domain of the operator , and satisfies the conditions of the proposition as soon as the coefficients @xmath119 and @xmath124 are continuous with respect to @xmath14 . this , trivially , provides the existence of a scattering solution . here , we show that @xmath122 is actually the unique one . assume that there exists another @xmath125 solution of the scattering problem . by the compact injection of @xmath126 in @xmath127 , we now that @xmath128 and @xmath129 are @xmath127 functions of @xmath32 over @xmath130 $ ] . therefore , the wronskian @xmath131 is well - defined . moreover , as the potentials @xmath132 are continuous functions over @xmath130 $ ] and as @xmath133 and @xmath134 are solutions of the associated sturm - liouville equation , they are in fact @xmath135 over @xmath130 $ ] . thus , the derivative of the wronskian is also well defined over @xmath130 $ ] . through a simple computation and by noting that @xmath133 and @xmath134 are solutions of the same sturm - liouville equations , one has @xmath136,\ ] ] and so the wronskian remains constant over the interval @xmath130 $ ] . for the finite branches @xmath57 , applying the ( dirichlet or neumann ) boundary conditions at the terminal nodes , we easily have @xmath137 and therefore the wronskian is identically 0 over the whole branch . this implies that , @xmath133 and @xmath134 are co - linear : @xmath138,\quad j=1,\cdots , n.\ ] ] over the branch @xmath59 , as @xmath139 satisfies a homogenous sturm - liouville equation ( @xmath140 ) , it necessarily admits the following form @xmath141 what remains to be shown is that one necessarily has @xmath142 and similarly @xmath143 . indeed , the equations and provide @xmath60 linear relations for the @xmath60 unknown coefficients @xmath119 and @xmath124 . trivially , as soon as , the coefficients @xmath144 are non - zero , these linear relations are independent and there exists a unique solution for the unknowns @xmath119 and @xmath124 . however , the zeros of each one of the coefficients @xmath144 correspond to isolated values of @xmath14 ( square - root of the eigenvalues of the operator @xmath145 with dirichlet boundary condition at @xmath77 and dirichlet or neumann boundary condition at @xmath146 ) . therefore , the coefficients @xmath119 and @xmath124 are well - defined except for a set of isolated values of @xmath14 . however , as these coefficients need to be continuous with respect to @xmath14 , they will , also , be defined uniquely over these singular points . furthermore , dividing by @xmath147 and inserting , we find the explicit formula @xmath148 finally , inserting the value of @xmath149 found in into , we find @xmath150 as a first inverse problem , we consider the inversion of the geometry of the network . in fact , we will prove the well - posedness of the inverse problem of finding the number of branches @xmath0 and the lengthes @xmath151 of a star - shape graph through only one reflection coefficient @xmath152 ( the case of dirichlet reflection coefficient can be treated similarly ) . furthermore , as we will see through the proof of the theorem [ thm : geometry ] , the method is rather constructive and one can think of an algorithm to identify the lengthes , at least approximately . the proof is based on an asymptotic analysis in high - frequency regime of the reflection coefficient and some classical results from the theory of almost periodic functions ( in bohr sense ) . before , announcing the main theorem , we need a few lemmas . a first lemma precises the high frequency behavior of the neumann fundamental solutions @xmath153 : [ lem : fund ] consider a potential @xmath41 in @xmath154)$ ] and take the neumann fundamental solution , @xmath155 , defined as in definition [ def : fund ] . we have @xmath156 where @xmath157 denotes the spatial derivative @xmath158 at @xmath77 . we start by writing @xmath155 in its integral representation . indeed , the fundamental solution @xmath159 is given by @xcite : @xmath160 where @xmath161 is a real function with the same regularity as @xmath162 : here , it is @xmath127 with respect to both coordinates . we note that , as @xmath163 is in @xmath127 , by integrating by parts , one has : @xmath164 therefore @xmath165 as @xmath166 and we have the first relation . for the spatial derivative @xmath167 at the point @xmath77 , we have : @xmath168 the kernel @xmath169 being @xmath127 , we have @xmath170 thus , @xmath171 as @xmath166 and the second relation follows . as we see by lemma [ lem : fund ] , in the high - frequency regime , the fundamental solutions become asymptotically independent of the potential @xmath41 . the next lemma , provides an explicit method to identify the number @xmath0 and the lengthes @xmath151 of the branches for the homogenous case , where we know that the potentials @xmath172 are all zero . [ lem : homogenous ] consider a star - shape network @xmath56 composed of @xmath173 branches of length @xmath62 ( @xmath174 ) all joining at a central node so that the whole number of branches @xmath0 is given by @xmath175 . assume the potential @xmath41 on the network to be 0 ( @xmath176 ) . then the knowledge of the neumann reflection coefficient @xmath152 determines uniquely the parameters @xmath177 and @xmath178 . we need to apply the explicit computation of the reflection coefficient provided by . the fundamental solutions are given , simply , by @xmath179 . therefore : @xmath180 the knowledge of @xmath152 determines uniquely the signal : @xmath181 assuming , without loss of generality , that the lengthes @xmath62 are ordered increasingly @xmath182 , the first pole of the function @xmath183 coincides with @xmath184 and therefore determines @xmath185 . furthermore , @xmath186 and therefore one can also determine @xmath187 . now , considering the new signal @xmath188 , one removes the branches of length @xmath185 and exactly in the same manner , one can determine @xmath189 and @xmath190 . the proof of the lemma follows then by a simple induction . we are now ready to announce the main theorem of this section : [ thm : geometry ] consider a star - shape network @xmath56 composed of @xmath173 branches of length @xmath62 ( @xmath174 ) all joining at a central node so that the whole number of branches @xmath0 is given by @xmath175 . assume the potential @xmath41 on the network to be , simply , @xmath191 . then the knowledge of the neumann reflection coefficient @xmath152 determines uniquely the parameters @xmath177 and @xmath178 . assume that , there exists two graph settings @xmath192 and @xmath193 ( the lengthes @xmath62 are not necessarily different ) giving rise to the same neumann reflection coefficients : @xmath194 . by the explicit formula , we have @xmath195 this is equivalent to : @xmath196defining the function : @xmath197the asymptotic formulas of lemma [ lem : fund ] and the imply @xmath198 however , the function @xmath199 is a trigonometric polynomial and almost periodic in the bohr s sense @xcite . the function @xmath200 is , also , almost periodic and furthermore , we have @xmath201 this , trivially , implies that @xmath202 ( one only needs to apply the parseval s theorem to the generalized fourier series of the function @xmath203 ) . however , the relation @xmath204 is equivalent to @xmath205 and therefore , by lemma [ lem : homogenous ] , the two settings are equivalent and the theorem follows . a second inverse problem , related to the detection of soft faults in the network , can be formulated as the identification of the potentials on the branches . here , we consider the case of homogenous perfect wires . we will show that the measurement of the two reflection coefficients @xmath110 and @xmath109 , corresponding , respectively , to a short circuit and an open circuit experiment , is enough to identify uniquely the small changes in the potential remaining in a certain regularity class . indeed , we will prove that the inverse problem of finding the @xmath206 potentials in an @xmath207 @xmath208-neighborhood of the zero potential ( homogenous case ) , is well - posed for @xmath207 small enough . in this aim , we will need an additive assumption on the electrical lengths @xmath151 , in order to remove symmetries leading to degeneracy problems . however , as it will be discussed later , it seems that this assumption can be relaxed . in a first result ( theorem [ thm : equal ] ) , we prove that under some natural assumption on the electrical lengths , the knowledge of only one reflection coefficient ( here @xmath109 ) is sufficient to identify uniquely the values : @xmath209 if anyone of these quantities appear to be different from zero , we know that there has been a change of parameters in the corresponding branch . by performing classical inverse scattering techniques over this branch @xcite we can identify its soft faults . the question is therefore to identify the soft faults in the network which does not change the quantities @xmath210 . this is treated in the two theorems [ thm : borg ] and [ thm : half ] . for both these theorems , we will need some more restrictive assumptions on the electrical lengths @xmath211 . in theorem [ thm : half ] , we will show that the knowledge of only one reflection coefficient ( here @xmath109 ) is enough to identify the potential when we know that it has not changed on the first half of the branches . the theorem [ thm : borg ] provides a well - posedness result for the inversion of the potential over the whole graph but necessitates the knowledge of both reflection coefficients @xmath110 and @xmath109 . in the sequel , we note that the potential over the infinite branch @xmath59 is always given and is 0 . indeed , this homogenous line is added only for the reflectometry experiment . the following theorem provides a global inversion result concerning the quantities @xmath212 . [ thm : equal ] consider a star - shaped graph @xmath56 and assume that @xmath213 : : the electrical lengths @xmath214 are not entire multiples of each other , @xmath215 if there exist two potentials @xmath216 and @xmath217 in @xmath206 giving rise to the same reflection coefficient , @xmath218 , one necessarily has : @xmath219 [ rem : integral ] note that the result of the theorem is also valid if we have @xmath220 . this theorem allows us to identify the situations where the soft fault in the network causes a change of the quantities @xmath212 . as explained above , the inverse problem can then be considered on separate branches and solved through classical inverse scattering techniques . in the two following theorems , we assume that the soft faults in the network leave the quantities @xmath212 unchanged and as for the perfect situation we are dealing with homogenous lines we will assume that @xmath221 : : @xmath222 for @xmath223 . [ thm : half ] consider a star - shaped graph @xmath56 and assume that @xmath224 : : for any @xmath225 such that @xmath226 , @xmath227 is an algebraic irrational number . then , there exists @xmath228 small enough such that , if the potentials @xmath41 and @xmath229 belonging to @xmath206 and satisfying @xmath221 , @xmath230 and @xmath231 for @xmath232 $ ] , @xmath223 , give rise to the same neumann reflection coefficients , @xmath218 , then @xmath233 . [ rem : half ] once again , the result remains valid if we replace the neumann reflection coefficient by the dirichlet one . [ rem : b1 ] the assumption @xmath224 seems very restrictive . however , as it will be seen through the proof , the only thing we need is that for any of fractions @xmath227 , there exists at most a finite number of co - prime factors @xmath234 , such that the diophantine approximation @xmath235 holds true . however , this is a classical result of the borel - cantelli lemma that for almost all ( with respect to lebesgue measure ) positive real @xmath236 s this diophantine approximation has finite number of solutions . [ thm : borg ] consider a star - shaped graph @xmath56 satisfying @xmath224 . there exists @xmath228 small enough such that , if the potentials @xmath41 and @xmath229 belonging to @xmath206 and satisfying @xmath221 and @xmath237 , give rise to the same neumann and dirichlet reflection coefficients , @xmath238 then @xmath233 . the theorem [ thm : borg ] is a natural generalization to the case of a graph of the theorem of two boundary spectra on an interval @xcite ( see , for instance , theorem 1.4.4 , page 24 , @xcite ) . in @xcite , borg proved that the knowledge of two spectral data corresponding to two boundary conditions , determine uniquely the potential on an interval . here the reflection coefficients @xmath109 and @xmath110 play the role of this spectral data ( see the subsection [ ssec : sturm - liouville ] ) . throughout this subsection , we will consider a general star - shaped graph @xmath56 ( of @xmath0 finite branches ) and a potential @xmath216 belonging to @xmath206 . furthermore we assume for the potential @xmath41 that , the norm @xmath239 is sufficiently small . the main objective of this subsection is to show that the knowledge of the reflection coefficient @xmath109 for @xmath240 ( resp . @xmath110 for @xmath241 ) is equivalent to the knowledge of different positive spectra of sturm - liouville operators defined on @xmath56 with neumann ( resp . dirichlet ) boundary conditions at terminal nodes and for various boundary conditions at the central node . in fact , defining the function @xmath242 where @xmath243 is given by , we have the following result . [ prop : sturm ] fix @xmath107 and define the schrdinger operators @xmath244 on the compact graph @xmath56 as follows : @xmath245 where @xmath246 ( resp . @xmath247 ) denotes the space of infinitely differentiable functions @xmath248 defined on @xmath56 satisfying the boundary conditions @xmath249 then we are able to characterize the positive spectrum of @xmath244 as a level set of the function @xmath250 : @xmath251 we prove the proposition for the case of neumann boundary conditions . the dirichlet case can be treated exactly in the same manner . we start by proving the inclusion @xmath252 let @xmath253 , then there exists @xmath254 eigenfunction of the operator @xmath255 associated to @xmath256 . in particular , it satisfies @xmath257 where @xmath258 is the common value of @xmath254 at the central node . now we extend @xmath254 to the extended graph @xmath58 , such that @xmath259 is a scattering solution for @xmath240 ( see the proposition [ prop : scattering ] ) . in particular , the function @xmath259 must satisfy , at the central node , @xmath260 noting that @xmath254 is an eigenfunction of @xmath261 , we have @xmath262 now , noting that @xmath259 over the infinite branch admits the following form @xmath263 the relation yields to @xmath264 or equivalently @xmath265 this proves the first inclusion . now , we prove that @xmath266 let @xmath267 be such that @xmath268 . we consider a scattering solution @xmath259 of the extended operator @xmath240 ( defined by ) associated to the frequency @xmath256 . we , then , prove that the restriction of @xmath259 to the compact graph @xmath56 is an eigenfunction of @xmath255 associated to the eigenvalue @xmath256 . this trivially implies that @xmath253 . in this aim , we only need to show that this restriction of @xmath259 to @xmath56 is in the domain @xmath269 . indeed , this is equivalent to proving that the boundary condition : @xmath270 is satisfied . as @xmath259 is a scattering solution of @xmath240 , it satisfies @xmath271 furthermore , @xmath272 and so @xmath273 this proves and finishes the proof of the proposition . we have shown that , the knowledge of the reflection coefficient @xmath274 for @xmath58 is equivalent to the knowledge of the positive part of the spectra for @xmath275 for all values @xmath276 at the central node . the following proposition provides the characteristic equation permitting to identify the eigenvalues of the operator @xmath244 : [ prop : char ] the real @xmath277 is an eigenvalue of the operator @xmath244 if and only if @xmath278 where @xmath279 @xmath280 being the fundamental solutions on different branches . we give the proof for the neumann boundary conditions , noting that the dirichlet case can be treated , exactly , in the same manner . assume @xmath281 to be a positive eigenvalue of @xmath255 . the associated eigenfunction , @xmath282 , has necessarily the following form : @xmath283 where @xmath284 s are real constants and the vector @xmath285 is different from zero . the function @xmath286 , being in the domain @xmath269 , it should satisfy the associated boundary condition at the central node . this implies that the vector @xmath285 is in the kernel of the matrix : @xmath287 m:= ( [ cols= " < , < , < , < , < " , ] ) @xmath287where @xmath288 denotes @xmath289 . this means that the determinant @xmath290 is necessarily 0 . developing this determinant we find : @xmath291 [ cor : char ] consider two potentials @xmath216 and @xmath292 and denote by @xmath240 and @xmath293 , the associated neumann schrdinger operators defined on the extended graph @xmath58 . assuming that the of the reflection coefficients @xmath109 and @xmath294 are equivalent @xmath218 , we have @xmath295 where @xmath296 , @xmath297 , @xmath298 and @xmath299 are defined through for the potentials @xmath41 and @xmath229 . by proposition [ prop : sturm ] , @xmath300 is an eigenvalue of the operator @xmath301 and @xmath302 . applying the proposition [ prop : char ] , this means that @xmath303 as @xmath194 , we have @xmath304 and thus the above equation yields to . the above corollary is also valid when we replace the neumann by dirichlet boundary conditions . finally , this corollary yields to the following proposition on the difference between the two potentials @xmath41 and @xmath229 . [ prop : char2 ] consider two potentials @xmath216 and @xmath292 and denote by @xmath240 and @xmath293 , the associated neumann schrdinger operators defined on the extended graph @xmath58 . assuming that the of the reflection coefficients @xmath109 and @xmath294 are equivalent @xmath218 , we have @xmath305 where @xmath306 . for @xmath223 , we have : @xmath307 here the second line has been obtained from the first one , replacing @xmath308 by @xmath309 and integrating by parts . using and the above equation , we have : @xmath310 before finishing this subsection , note that , once more , the above proposition is also valid for the case of dirichlet boundary conditions and @xmath311 implies : @xmath312 we are now ready to prove the theorems [ thm : equal ] , [ thm : half ] and [ thm : borg ] . we prove the theorem [ thm : equal ] applying the characteristic equation and the integral representation of the functions @xmath313 . again , for simplicity sakes , we give the proof only for the case of neumann boundary conditions , noting that the dirichlet case can be done in a similar way . similarly to the section [ sec : geometry ] , the fundamental solutions @xmath115 are given by @xcite : @xmath314 where @xmath315 are @xmath127 with respect to the both coordinates . the above equation yields to the following representation for the product functions : @xmath316 where @xmath317 is a volterra kernel , i.e. @xmath318 if @xmath319 and @xmath320+\\ + \int_{2y - x}^x\tilde{g}_{{\mathcal n}}^j(x , s)g_{{\mathcal n}}^j(x , s-2y)ds+\int_{-x}^{x-2y}\tilde{g}_{{\mathcal n}}^j(x , s)g_{{\mathcal n}}^j(x , s+2y)ds.\end{aligned}\ ] ] at this point , we note that , as @xmath321 is a @xmath127 function , we have @xmath322.\ ] ] applying the characteristic equation and developing the products @xmath115 @xmath323 by , and finally using , we have : @xmath324 in the last passage , we applied the fact that @xmath325 , since @xmath326 is in @xmath206 . + the left side of is an almost periodic function with respect to @xmath14 , in the bohr s sense . following the same arguments as those of the theorem [ thm : geometry ] we obtain @xmath327 choosing @xmath328 and noting that , as the parameters @xmath151 are not entire multiples of each other i.e. @xmath329 : @xmath330 for each value @xmath331 , we have @xmath332 and finishes the proof of the theorem [ thm : equal ] . in this subsection , we consider two potentials @xmath216 and @xmath217 , satisfying the assumptions of the theorem [ thm : half ] . assuming that they give rise to the same neumann reflection coefficients , @xmath194 , we have the characteristic equation . let us define the operator @xmath333 to be the operator @xmath145 over @xmath130 $ ] with the domain @xmath334 where @xmath335 denotes the space of infinitely differentiable functions @xmath336 defined on @xmath130 $ ] satisfying dirichlet boundary condition at 0 and neumann boundary condition at @xmath62 . by the perturbation theory for linear operators @xcite , there exists @xmath337 small enough such that , if @xmath338 then the eigenvalues of @xmath333 remain positive . considering @xmath339 ( @xmath340 ) the sequence of eigenvalues of @xmath333 , implies for each @xmath341 @xmath342 where we have applied the fact that @xmath343 . at this point , we will use the assumption @xmath224 on the lengthes @xmath62 to obtain a lemma on the non - overlapping of the eigenvalues for different branches : [ lem : separation ] under the assumptions of the theorem [ thm : half ] , there exists @xmath344 small enough such that , if @xmath345 , then @xmath346 in order to prove this lemma , we only need to show that @xmath347 is not an eigenvalue of @xmath348 nor @xmath349 for @xmath350 . in this aim , we first show that there exists @xmath351 such that for integers @xmath352 , @xmath353 is different from @xmath354 and @xmath355 the eigenvalues of @xmath348 and @xmath349 . assume , contrarily , that , for all @xmath351 , there exists @xmath352 , such that @xmath356 as the potentials @xmath41 and @xmath229 are @xmath357 , we have the following asymptotic formula s for the eigenvalues ( see , for instance , remark 1.1.1 , page 7 , @xcite ) : @xmath358 however , as by assumption @xmath221 , the integrals @xmath359 and @xmath360 are zero , this implies : @xmath361 therefore , the relation implies that , for all @xmath351 there exists @xmath352 such that @xmath362 assuming , without loss of generality , that @xmath363 and dividing the above inequality by @xmath364 , we have @xmath365 therefore , we must have the existence of an infinite number of integer couples @xmath366 satisfying the above inequality . however , by thue - siegel - roth theorem @xcite , for any irrational algebraic number @xmath236 , and for any @xmath367 , the inequality @xmath368 has only a finite number of integer solutions @xmath369 ( @xmath370 ) . this , trivially , leads to a contradiction and therefore there exists @xmath351 such that for @xmath352 and @xmath371 , @xmath372 and @xmath373 . for the @xmath374 first eigenvalues on each branch , we apply the perturbation theory for linear operators @xcite . having @xmath63 , the branch index , fixed , we will show that for @xmath375 small enough , if @xmath376 then the quantities @xmath377 do note coincide with the quantities @xmath378 , where @xmath350 . in fact , for the case of @xmath379 , this is , trivially , a consequence of the fact that the branch lengthes @xmath62 and @xmath380 are 2-by-2 @xmath381-linearly independent . now , adding a small perturbation , @xmath41 or @xmath229 , in the generalized sense ( see page 206 @xcite ) , by the continuity of a finite system of eigenvalues ( see page 213 @xcite ) , this claim remains valid . thus , there exists a small enough @xmath382 such that , if @xmath383 then @xmath377 do not coincide with @xmath384 , for all @xmath385 verifying @xmath371 . applying lemma [ lem : separation ] to the , for @xmath386 , we have : @xmath387 using the integral representation with a volterra kernel , we rewrite@xmath388where we have applied the assumption @xmath221 , @xmath389 , for the passage from the second to the third line and the fubini theorem for the last passage . this implies that @xmath390 where @xmath391 by the assumption of theorem [ thm : half ] , the potentials @xmath392 are zero on @xmath393 $ ] and consequently @xmath394 on @xmath393 $ ] . therefore , we have @xmath395 however , by the proposition 1.8.6 of @xcite , the system @xmath396 provides a riesz basis over @xmath397 $ ] . consequently , @xmath398.\ ] ] this homogenous volterra integral equation has only the trivial solution @xmath399 on @xmath130 $ ] . this , trivially , implies that for @xmath400 , if @xmath401 then @xmath402 on @xmath56 . we consider two potentials @xmath216 and @xmath217 , satisfying the assumptions of the theorem [ thm : borg ] . assuming that they give rise to the same neumann and dirichlet reflection coefficients , @xmath194 and @xmath311 , we have the characteristic equations and . we define the operator @xmath403 exactly as @xmath333 ( defined in the previous subsection ) with dirichlet boundary conditions at 0 and at @xmath62 . still , by the perturbation theory for linear operators @xcite , there exists @xmath337 small enough such that , if @xmath338 then the eigenvalues of @xmath333 and @xmath403 are all positive . considering @xmath339 ( @xmath340 ) the sequence of eigenvalues of @xmath333 , and @xmath404 ( @xmath405 ) the sequence of eigenvalues of @xmath403 , the equations and imply:@xmath406where we have applied the fact that @xmath407 . following the same arguments as those of the lemma [ lem : separation ] , there exists @xmath344 small enough , such that @xmath408 therefore , for each @xmath223 we have , @xmath409 at this point we define the sequence @xmath410 as follows @xmath411 as @xmath412 by assumption @xmath221 , implies @xmath413 where @xmath414 however , by lemma [ lem : riesz2 ] proved in the appendix , the sequence @xmath410 forms a riesz basis over @xmath415 and therefore @xmath416 , which finishes the proof of theorem [ thm : borg ] . this work was supported by grants from the french national research agency ( anr project 0-defect ) and from digiteo ( project diags ) . the goal of this appendix is to provide a lemma [ lem : riesz2 ] on the riesz basis property for the sequences @xmath410 defined in subsection [ ssec : borg ] . note that this is a direct consequence of a result already proved in @xcite . however , we provide a proof for the sake of completeness . we first provide a classical result on riesz sequences @xcite : [ lem : bar ] let @xmath417 be a riesz basis for the banach space @xmath418 . let @xmath419 be quadratically close to @xmath417 , i.e. @xmath420 then if the sequence @xmath421 is @xmath14-linearly independent or if it is complete in @xmath418 , then it is , also , a riesz basis for @xmath418 . throughout the appendix , we consider the sequence @xmath422 to be defined as in subsection [ ssec : borg ] . we therefore have the following lemma : [ lem : riesz2 ] assuming @xmath41 and @xmath229 in @xmath206 , the sequence @xmath410 provides a riesz basis over the hilbert space @xmath415 . fixing the branch index @xmath63 , we prove that @xmath423 is a riesz basis over @xmath424 . here , for simplicity sakes , we have identified the vector @xmath425 with its @xmath63-th component and therefore @xmath425 is not a vector anymore but rather a function in @xmath426 . we also remove the branch index @xmath63 . first , we prove that @xmath427 is a riesz basis over @xmath428 . note that @xmath429 in fact , one can easily see that the functions @xmath430 are solutions of @xmath431 where @xmath432 in particular , we obtain that the functions @xmath433 are the eigenfunctions of with boundary conditions @xmath434 . then there exists a bi - orthonormal sequence @xmath435 in @xmath436 $ ] , eigenfunctions of the adjoint operator @xmath437 thus @xmath433 are @xmath14-linearly independent . + from the integral representation of fundamental solutions and from lemma [ lem : fund ] , we have @xmath438 so @xmath439 i.e. @xmath433 is quadratically close to @xmath440 . by the virtue of the lemma [ lem : bar ] , this gives that @xmath433 is a riesz basis over @xmath441 . now let us show that the sequence of the functions @xmath442 is complete in @xmath441 . indeed , for any @xmath443 , suppose that @xmath444 in particular , considering @xmath445 , we have @xmath446 and therefore , @xmath447 integrating by part , we have @xmath448 and therefore using the boundary condition @xmath449 , @xmath450 the sequence @xmath433 is complete , hence @xmath451 for @xmath452 $ ] . therefore @xmath453 is complete in @xmath441 . since @xmath454 ( where @xmath455 s are appropriate normalizing constants ) is quadratically close to @xmath456 , it follows from lemma [ lem : bar ] that @xmath457 is a riesz basis for @xmath441 .	the frequency domain reflectometry ( fdr ) is studied as a powerful tool to detect hard or soft faults in star - shaped networks of nonuniform lossless transmission lines . processing the fdr measurements leads to solve an inverse scattering problem for a schrdinger operator on a star - shaped graph . throughout this paper , we restrict ourselves to the case of minimal experimental setup corresponding to only one diagnostic port plug . first , by studying the asymptotic behavior of the reflection coefficient in the high - frequency limit , we prove the identifiability of the geometry of this star - shaped graph : the number of edges and their lengths . the proof being rather constructive , it provides a method to detect the hard faults in the network . next , we study the potential identification problem by inverse scattering , noting that the potentials represent the inhomogeneities due to the soft faults in the network wirings . here , the main result states that the measurement of two reflection coefficients , associated to two different sets of boundary conditions at the extremities of the tree , determines uniquely the potentials ; it is a generalization of the theorem of the two boundary spectra on an interval @xcite . + _ inverse scattering , telegrapher s equation , inverse sturm - liouville problem _
several fields of computational science ( nanotechnology , materials science or biochemistry just to name a few ) would greatly benefit from the possibility of performing simulations of large systems , containing hundreds of thousands of atoms . conventional electronic structure calculations require the diagonalization of matrices whose size @xmath0 is of the order of the number of electrons in the system . the @xmath1-scaling cost of this step greatly limits the range of systems which can be tackled by _ ab - initio _ techniques , despite the fast - paced progress in the computational power of modern processors . based on the theoretical foundations of the nearsightedness principle of electronic matter@xcite , several techniques have been developed in the last years to avoid the diagonalization step , by directly computing the density matrix of the system using linear scaling algorithms@xcite . one of the earliest approaches have been to compute the finite temperature density matrix , i.e. the fermi function of the system s hamiltonian , @xmath2 , by decomposing it into easier - to - compute functions of the hamiltonian@xcite , for instance chebyshev polynomials or rational functions . in a recent paper@xcite we have discussed an exact decomposition of the fermi operator which can be efficiently computed by a combination of polynomial expansion and iterative inversion techniques . in this way , we achieved an efficient scaling with @xmath3 , the width of the spectrum of the hamiltonian in units of @xmath4 . this makes the method attractive for applications to metals or low - band gap semiconductors at low electronic temperature . in this short paper we discuss a number of improvements to this scheme , which further lower the operation count , leading to a scaling @xmath5{\delta\epsilon}$ ] , which is , to the best of our knowledge , the lowest so far reported in literature . we will follow closely the scheme of our previous work@xcite , obtaining analytical estimates for the operation count of the different steps which compose our algorithm , so as to optimize them in order to achieve optimal performance . our decomposition scheme is based on an exact expansion of the fermi operator , which can be elegantly derived using the grand - canonical formalism@xcite . here we only report the final result , namely the fact that the finite - temperature density matrix @xmath6 can be written in terms of a sum of complex - valued matrices @xmath7@xcite : @xmath8 equation ( [ eq : expansion ] ) is exact for any value of @xmath9 , but large values should be chosen so as to allow for simple and economical evaluation of the matrix exponential @xmath10 . here and in the rest of the paper , with no loss of generality , we use a shifted and scaled hamiltonian , i.e. we set the zero of energies at @xmath11 , and measure energy in units of @xmath12 . the expensive step in applying eq . ( [ eq : expansion ] ) is the inversion of the @xmath7 matrices . in ref.@xcite we have shown that , for large values of @xmath9 , a @xmath13 exists such that all of the matrices with @xmath14 are almost optimally conditioned , and can be easily inverted by a polynomial expansion . we have also shown how their overall contribution to the fermi operator can be computed at once , at the same cost of computing a single term , making the computational burden of the method virtually independent of @xmath9 . we will exploit this property and take often the large @xmath9 limit to derive analytical results . the remaining , low-@xmath15 matrices are more efficiently treated with an iterative newton inversion scheme@xcite . overall , the scaling in terms of matrix - matrix multiplications count , which is @xmath16 for standard chebyshev polynomials expansion@xcite , reduces to @xmath17 , and is even lower when fast polynomial summation methods@xcite are used . a first improvement over the initial formulation of our scheme can be seeked by noting that only the real part of @xmath18 enters the expression for the fermi operator , so that it can be more efficient to write @xmath19,\text { where } \quad { { \ensuremath{{\bf \bf n}}}_{{l}}}={{\ensuremath{{\bf 1}}}}+ e^{{\ensuremath{{\bf h}}}/p}-2 e^ { { \ensuremath{{\bf h}}}/2p}\cos \pi\frac{2{l}-1}{2p}. \label{eq : def - enne}\ ] ] only real matrices are involved in eq . ( [ eq : def - enne ] ) , leading to substantial savings in memory requirements and computation time . the inversion of the @xmath20s is the computationally demanding part of this approach ; in fact , if one performs an analysis similar to the one we have carried out in ref.@xcite , the condition number of @xmath20 is found to be approximately @xmath21 . this is higher than @xmath22 , but decreases more rapidly with @xmath15 . the low-@xmath15 inverse matrices , which are worse conditioned , can be computed by newton inversion , whose performances are only weakly affected by high condition numbers . moreover , we take profit of the form of eq . ( [ eq : def - enne ] ) to improve the evaluation of the contribution of the high-@xmath15 matrices , which in our previous work@xcite had a computational cost scaling as @xmath23 . since we also want to set up an expansion in chebyshev polynomials for @xmath24 , we rewrite @xmath20 it in terms of an auxiliary matrix @xmath25 whose spectrum lies between @xmath26 and @xmath27 : @xmath28 here we have introduced the shifting and scaling parameters @xmath29 and @xmath30 , computed in terms of the extremal values of the hamiltonian spectrum , @xmath31 and @xmath32 . these parameters can be estimated with a lanczos procedure . alternatively , since only a rough estimate is required , one can also perform a test calculation on a smaller , similar system , or use a matrix norm as an upper bound to the spectral radius of @xmath2 . the inverse @xmath24 can be therefore approximated as a sum of chebyshev polynomials of @xmath25 , @xmath33 , where @xmath34 is the @xmath35-th chebyshev polynomial , and the coefficients @xmath36 can be computed , by straightforward if tedious algebra , and take the values @xmath37 , \text { where } \quad b=\left(e^{{\mathrm{i}}\pi\frac{2{l}-1}{2p}}-z_0\right)/\zeta . \label{eq : cheb - coefficients}\ ] ] an upper bound to the error due to truncation of the chebyshev expansion can be estimated from @xmath38 . this estimate leads to a complex expression , which simplifies considerably if we assume @xmath39 . in this case , by taking the large @xmath9 limit , that the number of terms to be included in order do reach @xmath40 relative accuracy on @xmath24 reads @xmath41 this can be used as a first guess , to be refined by explicitly summing up the @xmath42 as computed from eq . ( [ eq : cheb - coefficients ] ) . at variance with the polynomial expansion of ref.@xcite , the operation count is linear with @xmath3 . this is the same scaling observed when the fermi operator is directly expanded in chebyshev polynomials@xcite . this analogy is not surprising , since the matrix powers @xmath43 entering the expansion are all independent of @xmath15 , and one could in principle obtain the whole density matrix from a single chebyshev polynomial evaluation , computing the expansion coefficients by summing over all the @xmath36s . estimated number of required matrix - matrix multiplications needed in order to obtain @xmath40 relative accuracy on the finite temperature density matrix , as a function of the hamiltonian spectum range @xmath3 . the grid lines correspond to a scaling @xmath44{{\delta\epsilon}}$ ] . we also plot the operation count expected when applying the fast summation methods of ref.@xcite to the standard chebyshev polynomial expansion@xcite ( @xmath45 ) . data points correspond to a test performed on the _ ab - initio _ hamiltonian for a @xmath46 alloy . details can be found in ref.@xcite : @xmath3 for the system is @xmath47 a.u . , and we have performed density matrix calculations at different temperatures , setting the target accuracy to @xmath48 . next to each point , we also report the relative error on the band structure energy , as computed with our algorithm , taking as reference the value obtained by diagonalization at the same electronic temperature . , scaledwidth=62.4% ] however , it is more convenient to stop the summation at @xmath49 , so as to reduce @xmath50 , and tackle the remaining terms by iterative inversion . the practical recipe is therefore to compute the coefficients @xmath51 and @xmath52 , using them to obtain @xmath53 and the contribution to the fermi operator from the `` tail '' of matrices with @xmath54 , respectively . the next step is to compute the terms up to @xmath55 by iterative inversion , a task which becomes much more efficient when a good initial guess for the inverse matrix is available . such a guess can be obtained by using a finite number of terms in the extrapolation @xmath56^i,\text { with } \quad \eta=\cos \frac{\pi\delta{l}}{p}+\sin\frac{\pi\delta{l}}{p } \tan\frac{\pi\left(2{l}-1\right)}{2p}. \label{eq : newt - extrapolant}\ ] ] the simplest approximation @xmath57 is already sufficient to guarantee convergence , which is fast and almost independent of the condition number of the @xmath20 s . the number of multiplications needed to obtain a relative accuracy of @xmath40 on @xmath24 starting from @xmath58 is in fact , in the large @xmath9 limit , @xmath59 one can use @xmath60 , which has been computed with the chebyshev expansion of the tail contribution , as the initial guess for the newton iteration which gives @xmath61 . the converged @xmath62 is in turn used to evaluate @xmath63 , and so on . the total number of matrix multiplications involved is easily obtained by combining eq . ( [ eq : cheb - mult ] ) and ( [ eq : newt - mult ] ) , @xmath64 . by minimizing this quantity the optimal value of @xmath13 is readily found . the chebyshev polynomials can be computed with fast summation techniques , in which case the operation count for that part becomes @xmath65 . in fig . [ fig : scaling ] we report our theoretical estimate for @xmath66 as a function of @xmath3 . the scaling is @xmath5{{\delta\epsilon}}$ ] , and the crossover point with fast - summed chebyshev expansion is around @xmath67 , which is easily reached if electronic temperatures of a few hundred k are necessary . in the same figure we also report the results from some test calculations performed for the hamiltonian of a semimetallic @xmath46 alloy , together with the measured relative error on the total energy , which is always one order of magnitude smaller than the target value enforced when choosing a particular chebyshev expansion length and convergence threshold for the matrix inversion . in order to keep the scheme as simple as possible , we have not considered the option of fine - tuning the procedure , which could further reduce the operation count . first , the operations count for newton inversion ( [ eq : newt - mult ] ) turns out to be often overestimated , so that hand - tuning @xmath13 can save several multiplications . higher - order extrapolations are easily obtained ( see appendix b of ref.@xcite ) , which provide a better starting point for iterative inversion . some preliminary results we obtained performing molecular dynamics with forces computed on the fly with this scheme show that the number of multiplies needed for iterative inversion can be halved if one uses the inverse _ saved from the previous timestep _ as the initial guess . in this short paper we have introduced significant improvements to an algorithm that tackles the problem of fermi operator expansion by an hybrid approach based on a combination of polynomial expansion and iterative matrix inversion , which we have presented in a recent paper . with this improved implementation , only real - valued matrices are involved , leading to significant savings with respect to the previous , complex - valued formalism . even more important , the scaling of the operation count with the hamiltonian spectrum width is lowered from @xmath68 to @xmath69{{\delta\epsilon}}$ ] , which is extremely appealing for broad - spectrum or low electronic temperature problems . work in the direction of a practical , linear - scaling implementation of these ideas is in progress@xcite . beside the use of an efficient sparse - matrix algebra library , subtle issues regarding matrix truncation must be addressed , and several optimizations , such as the extrapolation of @xmath18 from previous timesteps discussed above , can be used to improve the performances of the algorithm . we would like to thank clotilde cucinotta and giacomo miceli for discussion on @xmath46 system , and luca ferraro and simone meloni for helping us in a preliminary implementation of our method in a tight binding molecular dynamics code . v. pan , and j. reif , `` efficient parallel solution of linear systems , '' in _ stoc 85 : proceedings of the seventeenth annual acm symposium on theory of computing _ , acm press , new york , ny , usa , 1985 , p. 143 , isbn 0 - 89791 - 151 - 2 .	in a recent paper we have suggested that the finite temperature density matrix can be computed efficiently by a combination of polynomial expansion and iterative inversion techniques . we present here significant improvements over this scheme . the original complex - valued formalism is turned into a purely real one . in addition , we use chebyshev polynomials expansion and fast summation techniques . this drastically reduces the scaling of the algorithm with the width of the hamiltonian spectrum , which is now of the order of the cubic root of such parameter . this makes our method very competitive for applications to _ ab - initio _ simulations , when high energy resolution is required . address = computational science , department of chemistry and applied biosciences , eth zurich , usi campus , via giuseppe buffi 13 , ch-6900 lugano , switzerland , , email=michele.ceriotti@phys.chem.ethz.ch address = computational science , department of chemistry and applied biosciences , eth zurich , usi campus , via giuseppe buffi 13 , ch-6900 lugano , switzerland address = computational science , department of chemistry and applied biosciences , eth zurich , usi campus , via giuseppe buffi 13 , ch-6900 lugano , switzerland
so far , grbs have only been identified at cosmological distances . it is believed that grbs are caused by highly relativistic outflows with bulk lorentz factors @xmath1 that form a pair of opposite jets . even after correcting for narrow beaming , the energy radiated by long - duration ( @xmath2 ) grbs typically is @xmath3 , and the estimated kinetic energy of the jets reaches values of @xmath4 @xcite . relativistic shocks convert ( accelerate ) _ all _ particles from the incoming plasma they encounter to relativistic energies by randomizing their velocities in the comoving frame . in the stationary frame the mean energy per particle is @xmath5 @xcite . the total energy of cosmic rays ( crs ) produced by a grb can thus be up to @xmath6 times higher than the @xmath7 produced by nonrelativistic shocks of typical supernova remnants ( snr ) @xcite . grbs are prime candidate sources for the extragalactic component of crs , and they may also be sources of ultrahigh energy crs in our galaxy @xcite . calculations taking into account the star formation history and the grb beaming factor predict a galactic rate of one long grb every 3000 to @xmath8 years @xcite . a grb that occurred @xmath9 ago at @xmath10kpc from us could explain the knee in the observed cr spectrum at @xmath11 @xcite . ohers have suggested that giant radio shell objects in the milky way , such as the snr w49b @xcite or kpc - size hi super - shells @xcite could be remnants of grbs . the hess ( high energy stereoscopic system ) collaboration has recently discovered a population of @xmath12-ray sources in the galactic plane @xcite that are tev - bright but are often quiet in all other wavelengths and remain unidentified . all previously detected tev sources show strong x - ray emission which is the synchrotron counterpart of the inverse compton ( ic ) tev radiation from multi - tev electrons . the suppression of low - frequency flux is unexpected within conventional models . @xcite has proposed that this feature points to radiation production in the deep gravitational potential wells of black holes formed by galactic grbs . however , the extended appearance of these sources seems to contradict this explanation . the estimated rate of galactic grbs implies a likelihood of between one and several `` young '' ( age @xmath13yr ) , grb remnants in the interstellar medium ( ism ) at distances up to @xmath14 - 20kpc from us . some @xmath15 years after their acceleration by the relativistic shocks of grbs , multi - tev particles would diffuse over distances of about @xmath16pc . therefore , if detected in tev @xmath12-rays , galactic grbrs should appear as large plerionic ( ` center - filled ' ) nebulae in the ism . while probably most of the unidentified tev sources would eventually be associated with conventional galactic objects , here we argue that at least one source , hess j1303 - 631 @xcite , shows features that strongly suggest its identification with the remnant of a grb that happened @xmath17 years ago in our galaxy at a distance @xmath18 - 15)kpc from us . we show that this can explain both the total cr energy in the source , and its non - detection below tev energies . furthermore , we discuss observational signatures that can be used to identify grbrs unambiguously . among the unidentified tev sources discovered by hess , j1303 - 631 is the brightest source with the most detailed data @xcite . the measured @xmath19-@xmath20tev differential spectrum , approximated as a power - law @xmath21 , has a mean spectral index @xmath22 and an integrated energy flux @xmath23 . so far , no emission has been detected in any other wave band , including the x - ray band @xcite . the source is extended and is approximated , within current statistical errors , by a spherically symmetric 2-dimensional gaussian distribution with inherent width @xmath24 @xcite . the center - filled appearance of the source and the gradual steepening of the tev spectrum suggest the emission origin from a population of high - energy particles injected into the ism from a point source and evolved by energy - dependent diffusion . absence of a detectable synchrotron flux implicitly suggests a hadronic origin of the tev radiation . since the two @xmath12-rays from a @xmath25-decay each carry a fraction of the incident proton energy of about @xmath26 , the observed flux requires @xmath27 total energy in protons with @xmath28 . here , @xmath29 is the @xmath30-cooling time , and @xmath31 is the fraction of the lost proton energy that is converted into @xmath25-mesons . substitution of the cross - section @xmath32 and the inelasticity @xmath33 gives @xmath34 here @xmath35 , and @xmath36 accounts for protons with @xmath37tev . for standard snr acceleration , a broken power - law spectrum with @xmath38 above and @xmath39 below 5tev gives the smallest possible @xmath40 . a more plausible snr spectrum with @xmath41 predicts @xmath42 . integration of the proton energy distribution formed by diffusion from a point ` impulsive ' source @xcite along the line of sight gives @xmath43 $ ] , where @xmath44 is the mean diffusion length ( along each of the axes ) for a particle with energy @xmath45 , diffusion coefficient @xmath46 in a power - law approximation , and @xmath47 . integrating over a circle of angular radius @xmath48 , we derive a proton energy spectrum @xmath49 $ ] . the observed spectral index @xmath50 is significantly steeper than @xmath51 - 2.2 expected for a typical snr . we explain this steepening as an unavoidable result of energy dependent diffusion . given the hess point spread function ( psf ) , the derived @xmath52 suggests @xmath53 . calculations nicely fit the data for @xmath54 , and exclude @xmath55 or @xmath56 ( see fig.1 ) . this defines the following key - relation between the source distance and its age @xmath57 ( we count @xmath58 from the time when the first light from the source reached us , and neglect the source expansion during the light crossing time @xmath59 ) : @xmath60 this constrains the energy required to explain both the spectral and the spatial appearance of the source : @xmath61 the diffusion coefficient in the ism should have @xmath62 to explain the cr spectral index @xmath63 . a typical value of @xmath64 at @xmath65 is @xmath66 @xcite . a lower limit of @xmath67 can be derived from the grammage @xmath68 that explains the observed composition of cr nuclei @xcite , and does not allow crs to stay for more than @xmath69 in the galactic disk at heights @xmath70 where @xmath71 @xcite . values of @xmath72 much smaller than the typical ones for the ism would imply an enhanced level of turbulence and can be excluded , as the associated enhanced magnetic field would result in radio emission exceeding the flux upper limit derived below . with these results , we are now in the position to argue that the source is a grbr and not a typical snr . using @xmath67 , we constrain the source age to @xmath73 . even for a snr with @xmath74 , an age @xmath75 , and location at @xmath76 in the ism with @xmath77 , equation ( 1 ) predicts @xmath78 . for a typical 10% cr injection efficiency by a snr , this requires a kinetic energy @xmath79 . in principle , such energies can not be excluded for some extremely powerful supernovae ( sne ) of type ib / c or iin ( which in fact might be grbs ) that constitute @xmath80 of all sne @xcite . this implies a total of @xmath81-@xmath82 such events in our galaxy . the expected number of such snrs at a distance @xmath83 is only @xmath84-@xmath85 . furthermore , the non - detection of the shell of such powerful , close and young snr in the radio and x - ray bands is difficult to explain . an assumption that the source is an snr inside a molecular cloud with @xmath86 at @xmath87 @xcite relaxes the extremely high snr energy requirement and increases @xmath88 to acceptable values . however , the problem of the missing shell of this still powerful and young snr in such a dense medium persists . even for a total energy in relativistic electrons of about @xmath89 of the protons , which is a typical lower limit for an snr , the synchrotron flux would significantly exceed the flux upper limit at 5 ghz ( see fig . 3 ) . for a source age of @xmath90 , equation ( [ distance ] ) moves the source to @xmath91kpc and increases dramatically the energy requirements for an snr that accelerates particles by non - relativistic shocks . the full radius of the source ( detected up to @xmath92 ) increases to @xmath93 implying that it is located either in ordinary ism or _ entirely _ inside ( to have a quasi-_uniform _ profile for @xmath94 ) a giant molecular cloud . figure 2 shows the fluxes calculated for @xmath95 , @xmath96 and @xmath97yr . two different proton energy spectra are considered . one implies an ` extreme snr ' , with @xmath98 at @xmath99 but @xmath100 at @xmath101 . the total energy derived is @xmath102 @xmath103 . the ` grbr ' spectrum assumes @xmath104 implying protons accelerated by a relativistic shock . for qualitative estimates the break energy could be defined as @xmath105gev , where @xmath106 is the shock lorentz - factor at the time when most of the initial shock energy is transferred to relativistic protons ( see sec.3 ) . fitting the data we find @xmath107 . thus @xmath106 can be as high as the _ initial _ @xmath108 of grb jets @xcite . figure 2 assumes @xmath109 . this reduces the total energy required for the grbr model to @xmath110 . values of @xmath111 , which are not excluded , would lead to an increase @xmath112 ( for @xmath113 ) , and even slower for @xmath114 . the hess collaboration discussed the giant ( @xmath115 ) molecular cloud g303.9 - 0.4 @xcite with @xmath116-@xmath117 located at 12 kpc from us as a possible host of the source . for the snr - type spectrum , the minimum cr energy @xmath118(4 - 8)@xmath119 is problematic even for a very powerful snr . furthermore , sub - relativistic protons alone would contain sufficient energy for heating ( at a rate @xmath120 ) , and possibly also disintegration , of the cloud via coulomb losses . because of the `` low - energy cutoff '' expected below @xmath121 , these constraints do not apply if hess j1303 - 631 is a grbr . it is also possible that the cloud is merely a line of sight coincidence . even at a distance of @xmath12215kpc a grbr would still be able to power the inferred proton energy density in the ism . leptonic models require less power than hadronic models . ic radiation of multi - tev electrons on the cmb and galactic far - infrared photons could reproduce the observed spectrum . a broken power - law electron spectrum with @xmath123 below 10 tev , and @xmath124 above leads to the minimum required energy @xmath125 . however , the synchrotron radio flux of the leptonic model exceeds the flux upper limit that we derived from the parkes - mit - nrao southern 4850mhz survey @xcite even for an ism magnetic field @xmath126 ( see figure 3 ) . we estimate the limit by measuring a mean flux of 0.0032 jy / beam in the @xmath127 radius region centered on the source . calculating the flux in background regions of the same aperture , we estimate a standard deviation of 0.0033 jy / beam . using a beam width of @xmath128 fwhm , we then obtain a 5@xmath129 upper limit of 0.12jy on the flux from within @xmath127 from the source centroid . we have also computed the ic fluxes from the secondary @xmath130-decay electrons and from @xmath131-decay electrons @xcite , accounting also for their radiative energy losses . both fluxes are almost two orders of magnitude lower than the @xmath132-decay flux . the ratio of synchrotron to ic fluxes depends only on the ratio of the magnetic field to the target photon energy densities . hence the radio flux from the secondary electrons for the grbr model in figure 2 does not violate the derived upper limit in so far as their ic emission is well below the observed tev flux . a grbr should differ from a snr not only by the energetics , but also by three characteristics of its relativistic particle population . first , the prompt differential spectrum of protons ( and electrons ) behind the relativistic shock can not be a single power - law @xmath133 with @xmath134 as in the case of nonrelativistic shocks . as shocks with lorentz - factor @xmath135 boost _ all _ particles they encounter to relativistic energies , the prompt spectrum should break to an index @xmath136 below some energy @xmath137 ( observer frame ) to satisfy the laws of energy and particle number conservation across the shock @xcite . the final power - law index @xmath138 ( presumably @xmath139 ) and the break energy @xmath121 are formed while the shock decelerates , and will depend on a number of processes , such as diffusive escape of relativistic protons from the shell , their adiabatic cooling and reacceleration , etc . at @xmath140 a spectral index @xmath141 is expected to form in fermi - type processes @xcite . for the current data , @xmath121 could be as high as 5tev ( in figure 2 ) , but it could also be smaller , leading to gev @xmath12-ray fluxes detectable by glast . the second feature of grbrs that we predict is that a substantial fraction of the initial shock energy is eventually transferred to high - energy protons , while relativistic electrons would contain only a tiny fraction of it at times of non - relativistic evolution of the grbr shell . a grb shock becomes non - relativistic , and also quasi - spherical , on timescales @xmath142 at distances @xmath143 @xcite . at this stage , magnetic fields of @xmath144 @xcite are still high and will ( synchrotron ) cool electrons to a few gev within @xmath1451yr . as the initial kinetic energy ( 5 - 10)@xmath146 of the jets is much larger than the radiated ( and the magnetic field s ) energy , by then most of it should be in relativistic protons . the evolution of a non - relativistic shell that is dominated by the relativistic energy , and not by the kinetic energy of the shock , requires a fully relativistic treatment and remains unexplored . diffusion of relativistic protons ( thus escaping adiabatic losses ) upstream may broaden the shock precursor . because of the preferential acceleration of particles with lorentz factors larger than the precursor @xcite , energetic protons will readily be accelerated while acceleration of electrons will be suppressed . this would also apply to the spherical `` supernova''-type component ( which , however , may not be developed by grbs at all , see @xcite ) as this non - relativistic shock has to pass through the central @xmath147 region dominated by the energy of relativistic protons from the ( quasi - spherically opened up ) grb jets . we also note that in the presence of relativistic protons , nonrelativistic shocks may evolve ( and dissipate ) much faster than predicted by the standard sedov - taylor solution for snr shocks propagating in `` cold '' plasmas @xcite . third , considerable energy , @xmath148-@xmath149 , of the accelerated protons is channeled into collimated beams of neutrons @xcite from interactions @xmath150 with x - ray photons from the grb . first , this mechanism assures a minimum level of total energy of the protons that avoid adiabatic losses in the shell , @xmath151 . furthermore , neutrons with energies @xmath152 will @xmath131-decay into protons on length scales of @xmath153 . a characteristic signature of a grbr might then be an elongation of the tev @xmath12-ray emission due to conversion of the initial double - sided neutron beam into protons , which will subsequently be injected into the diffusion process along the jet path on length scales @xmath154-@xmath155 ( for @xmath156 ) . depending on the contribution of neutrons to the high - energy proton population , the two - dimensional gaussian profile may have an aspect ratio of up to @xmath157 with @xmath158 and @xmath159 . for a source size @xmath160 this would predict @xmath161 - 0.1 ) . however , neutron beams can also affect the shape of the grbr nebula indirectly via driving ( after @xmath131-decay ) the ism , and stretching the ambient magnetic field , along the jet direction @xcite . a faster diffusion of the bulk of protons in this direction could result in @xmath162 . remarkably , in both cases the aspect ratio should increase at higher energies . the hess collaboration determined the source spectrum with the photons from the bright central part of the source ( within 13.4@xmath163 from the centre ) . while our model explains this spectrum , it predicts that the energy spectrum will be significantly harder when determined for the entire source ( see figure 2 ) . another specific signature of a grbr is the unusually hard energy spectrum at @xmath164 - 100gev . glast will be able to verify or falsify this prediction . if hess j1303 - 631 is a snr , glast will easily detect the source at the flux level close to the upper limit of egret @xcite . however , a weak detection ( confirming the hard spectrum ) or non - detection by glast would signify the tell - tale detection of the signatures of proton acceleration by relativistic shocks of a grb . confirmation of hess j1303 - 631 and at least one other extended unidentified tev source , such as tev j2032 + 4130 @xcite , as grbrs would confirm the high rate , of order @xmath165 , of galactic grbs , and would make it more likely that grbs had a direct impact on life on earth @xcite . we thank dr . c. dermer , dr . p. mszros , and the referee for helpful comments . aa acknowledges the hospitality of the physics department and of the mcdonnell center for space sciences of the washington university . aharonian , f. a. , et al . 2002 , a&a , 393 , l37 aharonian , f. a. , et al . 2005 , a&a , 439 , 1013 aharonian , f. a. , et al . 2005 , science , 307 , 1938 aharonian , f. a. , et al . 2006 , , 636 , 777 atoyan , a. m. , aharonian , f. a. & vlk , h. j. 1995 , phys . d , 52 , 3265 atoyan , a. & dermer , c. d. 2003 , , 586 , 79 berezhko , e. g. , & ellison , d. c. 1999 , , 526 , 385 berezinsky , v. s. , et al . 1990 , astrophysics of cosmic rays ( amsterdam ) berger , e. , kulkarni , s. r. & frail , d. a. 2004 , , 612 , 966 blandford , r. d. & mckee , c. f. 1976 , phys . fluids , 19 , 1130 bloom , j. s. , frail , d. a. & kulkarni , s. r. 2003 , , 594 , 674 buckley , j. , et al . , 1994 , , 429 , 736 condon , j. j. , griffith , m r. , & wright , a. e. , 106 , 1095 dermer , c. d. & atoyan , a. 2003 , , 91 , 071102 1 - 4 dermer , c. d. 2002 , , 574 , 65 dermer , c. d. & holmes , j. m. 2005 , , 628 , l21 efremov , y.n.,elmegreen , b . & hodge , p.w.1998,,501,l163 grabelsky , d. a. , et al . 1998 , , 331 , 181 ioka , k. , kobayashi , s. & mszros , p. 2004 , , 613 , l17 katz , j. i. 1994 , , 432 , l107 kirk , j. g. & duffy , p. 1999 , j. phys . g , 25 , r163 lang , m. j. , et al . 2004 , a&a , 423 , 415 livio , m. & waxman , e. 2000 , , 538 , 187 macfadyen , a. i. & woosley , s. e. 1999 , , 524 , 262 melott , a. l. et al . 2004 , int . j. astrobiology , 3 , 55 mszros , p. 2002 , , 40 , 137 mitra , a. 2005 , proc . 29th icrc , to appear ; astro - ph/0507697 mukherjee , r. & halpern , j. p. 2005 , , 629 , 1017 ptuskin , v. s. & soutoul , a. 1998 , a&a 337 , 859 sveshnikova , l. g. 2003 , a&a , 409,799 toptygin , 2000 , astron . lett . , 26 , 356 wick , s . , dermer , c.d . & atoyan , a . 2004 , astropart.phys.21,125	we present an investigation of the multiwavelength data on hess j1303 - 631 , an unidentified tev source serendipitously discovered in the galactic plane by the hess collaboration . our results strongly suggest the identification of this particular source as the remnant of a gamma - ray burst ( grb ) that happened some few tens of thousands years ago in our galaxy at a distance on the order of @xmath0kpc from us . we show through detailed calculations of particle diffusion , interaction and radiation processes of relativistic particles in the interstellar medium , that it is possible for a grb remnant ( grbr ) to be a strong tev emitter with no observable synchrotron emission . we predict spectral and spatial signatures that would unambiguously distinguish grbrs from ordinary supernova remnants , including : ( 1 ) large energy budgets inferred from their tev emission , but at the same time ( 2 ) suppressed fluxes in the radio through gev wavebands ; ( 3 ) extended center - filled emission with an energy - dependent spatial profile ; and ( 4 ) a possible elongation in the direction of the past pair of grb jets . while grbrs can best be detected by ground - based gamma - ray detectors , the future glast mission will play a crucial role in confirming the predicted low level of gev emission .
a schematic qcd phase diagram in the plane of temperature @xmath0 and baryon chemical potential @xmath1 . solid lines indicate the first order phase boundaries . several critical points at which the first order lines terminate may exit , e.g. the high @xmath0 critical point @xcite and high @xmath1 critical point @xcite.,width=302 ] one of the main goals of the lattice qcd studies is to make first principle analysis of hot / dense qcd and to supply reliable inputs to quark - gluon plasma ( qgp ) phenomenologies @xcite . shown in fig.[t - mu ] is a schematic qcd phase diagram indicating three basic phases ; the chiral symmetry broken ( @xmath2sb ) phase , the color superconducting ( csc ) phase and the qgp phase . precise locations of the phase boundaries and the critical points as well as the dynamics in each phase should be determined by non - perturbative method such as the lattice qcd . qcd partition function in a finite box with a spatial volume @xmath3 and the lattice spacing @xmath4 can be written as @xmath5 = \int [ du]\ f(u)\ { \rm e}^{-s_g(u ) } , \label{eq : z}\end{aligned}\ ] ] where @xmath6 is the gauge field defined as an element of @xmath7 and @xmath8 denotes the quark contribution to @xmath9 . eventually , we need to take the continuum limit ( @xmath10 ) and the thermodynamic limit ( @xmath11 to extract physical observables in the real world . in the past few years , there have been considerable progress in lattice qcd approach : ( i ) available computer speed becomes as fast as 50 tflops , ( ii ) improved fermion formulations have been tested extensively ( such as staut , asqtad and p4 improved actions for staggered quarks and the clover action for wilson quark ) , ( iii ) new fermion formulations with good chiral properties are started to be used ( such as domain wall and overlap fermions ) , and ( iv ) new algorithms for full qcd simulations are proposed and implemented ( such as the rational hybrid monte carlo method ( rhmc ) and the domain - decomposition hybrid monte carlo method ( ddhmc ) @xcite ) . shown in fig.[hmc ] is so called the berlin wall plot where the computational cost in unit of tflops@xmath12year is plotted against @xmath13 . the quark mass dependence of the cost becomes weaker than before , and simulations with realistic quark masses may be done in a reasonable amount of time by using o(50 ) tflops machines . in the following , we will focus on the bulk and spectral properties associated with the qcd phase transition at finite @xmath0 with @xmath14 . for the developments in lattice qcd at finite chemical potential , see the recent review @xcite . updated berlin wall plot which shows the cost to generate 1000 independent gauge configurations on @xmath15 lattice with @xmath16 fm in various fermion formulations with modern algorithms @xcite . , width=321 ] the equation of state at finite @xmath0 characterizes the bulk qcd matter . also it becomes an fundamental input together with the transport coefficients @xcite to the relativistic hydrodynamics for qcd fluid @xcite . shown in fig.[milc ] is the state - fo - the - art calculation of the energy density @xmath17 as a function of @xmath0 in ( 2 + 1)-flavor qcd with asktad improved staggered quarks . ( see also the recent results for @xmath18-flavor qcd with the p4 action @xcite and for @xmath19-flavor qcd with the stout action @xcite . ) the figure shows that @xmath20 increases rapidly at @xmath21 mev and approaches the stefan - boltzmann ( sb ) limit from below . the deviation from the sb limit at high @xmath0 is due to the quark - gluon interaction which decreases only logarithmically as @xmath0 increases . quantitative understanding of the deviation is a non - trivial issue and can not be treated in naive thermal perturbation @xcite . energy density @xmath17 as a function of @xmath0 for ( 2 + 1)-flavor qcd with asqtad improved action on the lattice , @xmath22 and @xmath23 @xcite . the strange quark mass @xmath24 is about the physical value , while the light quark mass @xmath25 corresponding to @xmath26 . the sommer scale @xmath27 fm is used to set the scale . lattice spacing at @xmath28 mev is 0.247 ( 0.165 ) fm for @xmath29.,width=340 ] using the pressure @xmath30 and a vector @xmath31 , one may define the @xmath32-th order transition as the one where non - analyticity of @xmath33 in the thermodynamic limit ( @xmath34 ) first appears in the @xmath32-th derivative of @xmath33 with respect to @xmath35 . for example , if @xmath36 is discontinuous at certain @xmath0 , it is the first order transition . on the other hand , if @xmath37 is continuous and @xmath38 diverges at certain @xmath0 , it is an example of the second order transition . if smooth phase change takes place without the non - analyticity of @xmath33 , it is the crossover . to make a distinction between the phase transition and the crossover from the numerical simulations , the scaling behavior of certain observables as a function of @xmath3 needs to be studied ( the finite size scaling analysis ) @xcite . in the ( 2 + 1)-flavor qcd , the relevant observable is the chiral susceptibility whose volume dependence at its maximum reads @xmath39 where @xmath40 for the first ( second ) order transition , and @xmath41 for crossover . shown in the left panel of fig.[aoki_sus ] is the dimensionless chiral susceptibility @xmath42 around its peak for different values of the dimensionless volume @xmath43 for ( 2 + 1)-flavor qcd with realistic masses of u , d , s quarks @xcite . there is no appreciable volume dependence for the height of the susceptibility , which suggests the crossover nature . the right panel of fig.[aoki_sus ] shows the subtracted chiral susceptibility in the continuum limit @xmath44 ( = the peak of @xmath45 ) in a renormalization group invariant combination @xmath46 as a function of the dimensionless inverse - volume , @xmath47 . the height of the susceptibility approaches to a non - vanishing constant in the thermodynamic limit and has different scaling behavior from those of first or second order transitions . this gives a strong evidence that the hadron - qgp transition at finite @xmath0 with zero chemical potential is a crossover in the real world . , for different values of the spatial volume , @xmath48 . the stout improved staggered fermion action is employed and the quark masses are chosen to reproduce @xmath49=3.689 and @xmath50=1.185 . right : the volume dependence of the peak of the subtracted chiral susceptibility in the continuum limit . dotted ( dashed ) line is an expectation from the first order ( second order ) phase transition . figures are taken from @xcite.,title="fig:",width=245 ] , for different values of the spatial volume , @xmath48 . the stout improved staggered fermion action is employed and the quark masses are chosen to reproduce @xmath49=3.689 and @xmath50=1.185 . right : the volume dependence of the peak of the subtracted chiral susceptibility in the continuum limit . dotted ( dashed ) line is an expectation from the first order ( second order ) phase transition . figures are taken from @xcite.,title="fig:",width=283 ] recent determinations of the ( pseudo-)critical temperature for ( 2 + 1)-flavor qcd with staggered quarks . as for the lowerst bar , we add 9 mev to make a comparison to other results obtained from @xmath42 @xcite . , width=453 ] since the hadron - qgp transition in the real world is crossover , the critical " temperature @xmath51 is not a well - defined concept . nevertheless , one may introduce a pseudo - critical " temperature @xmath52 as a peak position of certain susceptibility . the actual value of @xmath52 is different for different choice of susceptibility , e.g. @xmath53 , @xmath54 , and @xmath55 . note also that multiplying an arbitrary function of @xmath56 and @xmath0 to the susceptibilities can change the value of @xmath52 . shown in fig.[t_pc ] is a summary of ( pseudo-)critical temperature recently calculated from the dimensionless chiral susceptibility , @xmath42 , for ( 2 + 1)-flavor qcd . extrapolation to the continuum limit is taken by using the data for several values of @xmath57 . the upper two bars are @xmath51 with different extrapolations to the chiral limit @xmath58 by milc collaboration @xcite . the middle bar is @xmath52 extrapolated to the realistic @xmath56 by rbc - bielefeld collaboration @xcite . the lower bar is @xmath52 calculated at the realistic @xmath56 by wuppertal - budapest collaboration @xcite . note that different improved actions and different scale determinations are adopted in three groups . when extracting @xmath52 in mev unit , we need to use the chain rule : @xmath59 here @xmath60 , which has a dimension of mass , is chosen either to be a direct physical observable ( such as @xmath61 , @xmath62 and @xmath63 ) or to be a semi - empirical quantity ( such as the sommer scales @xmath64 and @xmath65 defined from the heavy - quark potential through @xmath66 and @xmath67 ) . there are a number of issues to be studied further before drawing definite conclusion on the value of @xmath52 in the real world . in particular , ( i ) careful continuum extrapolation ( @xmath68 ) by the data with larger @xmath57 , ( ii ) determination of @xmath52 in the case of wilson fermion @xcite , and ( iii ) resolving the difference between the sommer scale in the continuum limit obtained from @xmath4 being determined by the 2s-1s mass splitting of the bottomonium with staggered quark ( @xmath69 fm ) @xcite and that obtained from @xmath4 being determined by the kaon mass with wilson quark ( @xmath70 fm ) @xcite . heavy quarks in the real world ( such as the charm and bottom ) have finite masses and may receive substantial kick from the light plasma - constituents which have typical energy of about @xmath71 . therefore , it is important to examine the dynamical correlations ( both spatial and temporal ) of the heavy quarks inside qgp . it is also relevant to study the modification of the heavy quarkonium properties in hot qcd matter @xcite . the equal - time bethe - salpeter wave function normalized at the origin as a function of the spatial separation @xmath72 between the charm and the anti - charm . blue ( red ) points correspond to free ( interacting ) quarks @xcite . the black dashed line is the wave function initially prepared at @xmath73 . note that @xmath74 in this figure . , width=321 ] an attempt to study the @xmath75 wave function at finite temperature in quenched qcd ( @xmath8 is take to be unity in eq.([eq : z ] ) ) on an anisotropic lattice was performed by umeda et al . they have studied the equal - time bethe - salpeter wave function @xmath76 at finite @xmath0 in the coulomb gauge . in fig.[umeda ] , the normalized wave function , @xmath77 is plotted below and above @xmath78 for various time slices . the black dashed line is the wave function prepared at @xmath73 . for free quarks without gauge interactions , the wave function becomes broader as the imaginary time @xmath79 increases . on the other hand , the wave function for the interacting system is almost unchanged even at @xmath80 , which suggests that @xmath75 may survive as a bound state in the deconfined phase . the spectral functions of hadronic correlators give us another information on hadronic modes at finite @xmath0 as originally suggested in @xcite . the spectral function @xmath81 for hadronic correlation may be defined through the spectral decomposition : @xmath82 the maximum entropy method ( mem ) provides an efficient and powerful way to obtain a unique @xmath83 from the lattice data @xmath84 without making a priori parameterization @xcite . applications of mem to the @xmath85 mesons at finite @xmath0 @xcite and to the charmoniums at finite @xmath0 @xcite within the quenched qcd have been carried out and it was shown that the @xmath75 and @xmath86 survive even above @xmath78 as distinct peaks . shown in the left ( right ) panel of fig.[asakawa ] is the dimensionless spectral function for @xmath75 ( @xmath87 ) below and above @xmath78 on @xmath88 ( @xmath89 ) anisotropic lattices with the spatial volume @xmath90 ( @xmath91 ) taken from ref.@xcite ( ref.@xcite ) . the figures indicate that the low - lying s - wave resonances still survive even at 1.5 @xmath78 , which is consistent with the result in fig.[umeda ] . note that this result is not an artifact of the small spatial volume , which was carefully checked in @xcite . there are several topics which are interesting to be studied within the quenched qcd , such as the spectral functions of the charmoniums moving inside the hot medium @xcite and the spectral functions of bottomoniums above @xmath51 @xcite . for the @xmath75 channel @xcite . right : spectral functions measured in quenched qcd simulations on an anisotropic lattice with the spatial volume ( 1.34 fm)@xmath92 for the @xmath86 channel @xcite.,title="fig:",width=283 ] for the @xmath75 channel @xcite . right : spectral functions measured in quenched qcd simulations on an anisotropic lattice with the spatial volume ( 1.34 fm)@xmath92 for the @xmath86 channel @xcite.,title="fig:",width=257 ] anisotropic lattices . the spatial lattice volume is @xmath93 ( 1.2 fm)@xmath92 and the light quark mass corresponds to @xmath94 . the figures are adapted from @xcite . , title="fig:",width=264 ] anisotropic lattices . the spatial lattice volume is @xmath93 ( 1.2 fm)@xmath92 and the light quark mass corresponds to @xmath94 . the figures are adapted from @xcite . , title="fig:",width=264 ] in quenched qcd , only gluons are thermally excited in the plasma , while , in full qcd , both quarks and gluons are thermally active . because the number of degrees of freedom increase in full qcd , the dissociation rate of the chamoniums will be larger . on the other hand , the pseudo - critical temperature @xmath52 in full qcd ( 160 - 200 mev ) is substantially smaller than @xmath95 mev in quenched qcd . due to the compensation of these effects , the ratio of the dissociation temperature @xmath96 and the ( pseudo-)critical temperature in full qcd is not much different from the ratio in quenched qcd and could take a value of around 2 @xcite . aarts et al . has recently reported an exploratory study of 2-flavor qcd on @xmath97 anisotropic lattices @xcite . shown in fig.[aarts ] are the spectral functions of the @xmath75 and @xmath87 . the results suggest that , even with the thermal excitations of light quarks , the low - lying charmoniums may survive inside the quark - gluon plasma above @xmath51 until they dissociate away around @xmath98 . we reviewed recent developments in lattice simulations on the equation of state , crossover nature of the thermal phase transition and the determination of the pseudo - critical temperature in ( 2 + 1)-flavor qcd . recent lattice results on the spectral properties of the heavy quarkoniums inside the quark - gluon plasma are also summarized . [ [ acknowledgements ] ] * acknowledgements : * + + + + + + + + + + + + + + + + + + + this work was partially supported by japanese mext grant no . 18540253 . the author thanks t. umeda , s. ejiri , f. karsch and z. fodor for useful information and discussions . 10 yagi k , hatsuda t and miake y 2005 _ quark - gluon plasma _ , ( cambridge univ . press , cambridge ) .	we review recent developments in lattice siumulations of the equation of state , order of the thermal phase transition and the determination of the pseudo - critical temperature in ( 2 + 1)-flavor qcd . owing to the increasing computer power , new argothithms , and improved fermion formulations , studies of bulk qcd matter are approaching to the stage of precision science . we also review recent lattice studies on the spectral properties of heavy quarkoniums inside the quark - gluon plasma ( qgp ) . although they are still in an exploratory stage , interesting physics in relation to the strongly correlated qgp is being extracted .
in the first paper ( manset & bastien 2000 , hereafter referred to as paper i ) , we presented numerical simulations of the periodic polarimetric variations produced by a binary star surrounded by a circumbinary envelope composed of electrons . thomson scattering was considered in an optically thin situation , along with pre- and post - scattering extinction factors which produce a time - varying ( small ) optical depth and affect the morphology of the periodic variations by adding an additional harmonic to the variations . we studied the effects that orbital inclination , optical depth , geometry of the envelope and cavity , size and eccentricity of the orbits , and non - equal mass stars had on the average polarization level and the amplitude of the polarimetric variations . we found that high polarization levels will result from a high inclination , a high optical depth , a flat envelope , or a big central cavity . polarimetric variations are more apparent for a low inclination , a high optical depth , a flat envelope , a small cavity , or an orbit which brings the stars close to the inner edge of the cavity . using the formalism developed by brown , mclean , & emslie ( 1978 , hereafter bme ) the observations , or in this case the results of the numerical simulations , are fitted with a fourier series with coefficients of orders zero , one , and two ( see equations 3 and 4 below ) . the first order coefficients from the fit ( terms in @xmath8 ) correspond to single - periodic variations ( variations seen once per orbit ) , and those of second order ( terms in @xmath9 ) to double - periodic variations ( seen twice per orbit ) . those coefficients are used to find the orbital inclination ( see equations 5 and 6 below ) . it was then shown that the bme formalism can be used to find the orbital inclination if it is @xmath10 , even though this formalism does not include variable absorption effects as in our simulations . the geometry ( flatness of the envelope , size of the central cavity ) and size of the orbit have no significant influence on the inclination found by the bme formalism . an extension of the bme formalism for eccentric orbits has been performed by brown , aspin , simmons , & mclean ( 1982 ) ; however , their case considers a localized scattering region in eccentric orbit about a star , which is not a suitable geometry for the cases we want to study here . therefore , we performed our own study of eccentric orbits . for eccentric orbits , single - periodic variations appear for eccentricities as low as 0.10 . as the eccentricity increases , these single - periodic variations dominate over the double - periodic ones . for low eccentricities , @xmath11 , the inclinations can be found with the first or second - order fourier coefficients , if @xmath12 and @xmath13 respectively . for the high eccentricities , @xmath14 , only the first - order coefficients should be used , if @xmath15 . orbital eccentricity is not the only factor that can introduce single - periodic variations : variable absorption effects , non equal mass stars , and asymmetric envelopes may also produce those . in this second paper , we investigate the effects of scattering by spherical dust grains ( mie scattering ) . in addition to studying the influence of the optical depth and true orbital inclination already considered in paper i , we also consider the effects of the composition and size of the dust grains on the level of polarization and the amplitude of the polarimetric variations . our goal is to study the periodic polarimetric variations of binary young stars surrounded by envelopes in which dust grains are responsible for the scattering and polarization , and compare polarimetric observations with our numerical simulations . as we have shown in paper i that the bme formalism is still valid beyond its original limits ( circular orbits , no extinction effects ) , now we also want to see if the bme formalism can be extended to the case of mie scattering to find the orbital inclination , even though electrons and spherical dust grains have different scattering properties , and the bme formalism has been developed for thomson scattering . the simulations presented here are akin to the simulations where single - periodic polarimetric variations , usually double - peaked , are produced by a spotted star surrounded by a dusty circumstellar disk ( see for e.g. stassun & wood 1999 ) . whereas in the latter case the polarimetric variations are produced by a pair of intrinsically polarized hot spots illuminating ( and scattering off ) the disk , here the variability is produced by the binarity of the source of light . the polarization produced by a single star surrounded by a circumstellar disk has been studied with monte carlo simulations , by , for e.g. , wood et al . the scattering model used to compute the polarization and its position angle produced by two stars in orbit at the center of an ellipsoidal envelope was presented in paper i. in addition to the parameters presented there ( geometry of the envelope , orbit characteristics , optical depth , grid size ) , here we have to choose a grain composition ( which will give a specific complex refractive index ) and the size of the grains . table [ tab - grains ] presents the characteristics of the grains studied in this paper . a uniform size distribution was used for most of our simulations , but we also studied a distribution of grain sizes . the calculations for mie scattering use the following formulas : @xmath16 @xmath17 where : + @xmath18 : optical depth in the equatorial plane of the envelope ; + @xmath19 : optical depth before scattering ( between the light source and the scatterer ) ; + @xmath20 : optical depth after scattering ( between the scatterer and the observer ) ; + @xmath21 : angle between the east west direction and the scattering plane ( which includes the light source , the scatterer and the observer ) ; + @xmath22 : scattering angle , between the light source , the scatterer and the observer ; + @xmath23 : distance between the light source and the scatterer ; + @xmath24 : volume element ( computed in relation to the grid size ) ; + @xmath25 : weight which equals 1 if there is a scatterer , 0 otherwise ; + @xmath26 : optical depth per grid step ; + @xmath27 : radius of the spherical dust grain ; + @xmath28 : wavelength of observation ; + @xmath29 : extinction cross section ; + @xmath30 and @xmath31 : van de hulst intensities . + variable absorption effects are considered by using @xmath19 and @xmath20 . depending on where the binary star is located inside the envelope , the optical depths between the source and the scatterer ( @xmath19 ) , and between the scatterer and the observer ( @xmath20 ) will vary , and with them , the polarization . the van de hulst intensities and @xmath29 are calculated with the usual formulas that can be found in van de hulst ( 1981 , chap . 9 ) and make use of the complex refractive index of the grains . the polarization @xmath32 and its position angle @xmath33 are calculated from the stokes parameters @xmath34 and @xmath35 . when the polarization is produced by scattering , the sign used for @xmath32 is related to its orientation with respect to the scattering plane ( the plane containing the source of light , the scatterer , and the observer ) . a positive polarization is perpendicular to the scattering plane , whereas a negative one is parallel to it . thomson scattering always produces positive polarization , but both positive and negative polarization are produced by mie scattering . as in paper i , a `` canonical simulation '' is a simulation which has the following parameters : an envelope with axis ratios 1.0 , 1.0 and 0.25 ( also referred to as a 25% flat envelope ) , a spherical cavity with radius 0.20 ( also referred to as a 20% cavity ) , an optical depth of 0.1 , and a grid size of 65 . in addition , calculations are made for a wavelength of @xmath36 . figure [ fig - cb_2_05_05 ] presents a schematic view of this canonical geometry . for further details on the calculations , see paper i. the simulated polarimetric curves produced by the scattering model presented above are analyzed with the bme formalism , as in paper i. for reference , we give the main formulas that are used . observations are represented as first and second harmonics of @xmath37 , where @xmath38 is the orbital phase ( @xmath39 ) : + @xmath40 the inclination can be found with the first ( equation [ eq - io1-p2 ] ) or second ( equation [ eq - io2-p2 ] ) order fourier coefficients , although it is expected for circular orbits that second order variations will dominate : + @xmath41 ^ 2 & = & \frac{(u_1+q_2)^2 + ( u_2-q_1)^2}{(u_2+q_1)^2 + ( u_1-q_2)^2 } \label{eq - io1-p2},\\ \left [ \frac{1-\cos i}{1+\cos i } \right]^4 & = & \frac{(u_3+q_4)^2 + ( u_4-q_3)^2}{(u_4+q_3)^2 + ( u_3-q_4)^2}. \label{eq - io2-p2}\end{aligned}\ ] ] in an alternative representation , the eccentricity of the ellipse in the @xmath42 plane is related to the inclination . for the @xmath43 ellipse : + @xmath44 for the @xmath45 ellipse : + @xmath46 an important difference between this bme formalism and our numerical simulations is the inclusion of variable absorption effects in our work . as in paper i , a preliminary step to the numerical calculations was to determine a suitable grid size , one with which the bme formalism would find an orbital inclination close enough to the real inclination ( say , within @xmath47 ) . using astronomical silicates grains with radii of @xmath48 , we calculated models with grid sizes ( radius ) 25 to 95 , by steps of 10 . the results of the simulations , in which no noise was introduced , were used as input data for the bme equations . the orbital inclinations found in this manner are almost identical ( within a few tenths of a degree ) to those we found for thomson scattering in paper i. as in paper i , we chose a grid size of 65 , for which the inclination is within @xmath49 of the true value , and the calculations can be performed in reasonable times . we made calculations for the canonical simulation ( 25% flat envelope , 20% central cavity , @xmath50 , grid radius=65 ) , the four grain compositions listed in table [ tab - grains ] , and five grain radii ( 0.02 , 0.05 , 0.10 , 0.2 , and @xmath51 ) , at a wavelength of @xmath36 . since the polarization produced by mie scattering depends on the size of the grain @xmath27 and the wavelength of observation @xmath28 , it is common to introduce the parameter @xmath52 ; the values of @xmath53 for the previously mentioned simulations are 0.18 , 0.45 , 0.90 , 1.8 , and 4.5 . for each simulation , we used only one grain size and one grain composition . if we compare the polarizing properties of dust grains and electrons , we note that dust grains ( irrespective of composition and size ) are less efficient polarizers than electrons . grains also produce smaller absolute polarimetric variations per particle than electrons , although both electrons and grains produce about the same @xmath54 ratios ; see table [ tab - grains - e ] . that important observation shows that for binary stars embedded in dust envelopes , periodic polarimetric variations should be harder to observe and detect . for grain sizes of 0.02 , 0.05 or @xmath55 ( @xmath56 ) , the highest polarization and polarimetric variations are found for astronomical silicates , followed by graphite and amorphous carbon , and finally dirty ice ; grains made of dirty ice are 510 times less polarizing and produce variations about 510 times smaller than grains made of astronomical silicates . for grains with @xmath57 ( @xmath58 ) , the dirty ice produces the highest polarization and variations . for @xmath59 ( @xmath60 ) , astronomical silicates , graphite , and amorphous carbon have similar polarimetric properties , and dirty ices are less efficient polarizers . the highest polarizations are produced by grains with sizes in the range @xmath61@xmath62 ( @xmath632.0 ) . for astronomical silicates , we also did simulations for a wider range of grain sizes , going from @xmath64 to @xmath65 . the results are presented in figure [ fig - p_vs_x ] , where it is seen that the maximum polarization is reached for @xmath66 . these results follow what is presented in simmons ( 1982 ) . see for example his figure 4 ( polarization as a function of @xmath53 for different grain compositions ) , where it can be seen that for silicate grains and ice , polarization depends on @xmath53 with a very different behavior for these 2 grain compositions ; a grain that is the most polarizing at one value of @xmath53 will not necessarily be the most polarizing grain at other values of @xmath53 . figure 6 in simmons ( 1982 ) shows for silicate grains in particular how a different absorption coefficient changes the behavior of the polarization as a function of @xmath53 . finally , figure 5a shows that @xmath67 , the value of @xmath53 for which the polarization is maximum , depends on both the real and imaginary parts of the refractive index ( and thus on the grain composition ) . we also made some more calculations for astronomical silicates , adding grain sizes of @xmath68 and @xmath69 . big grains ( @xmath70 ) are less efficient polarizers , as was already known ( see for example daniel 1978 , simmons 1983 ) . as shown by our simulations and by simmons ( 1982 ) , the average polarization produced by dust grains thus depends on the grain composition and size . polarization reversal ( when the polarization goes from positive , or perpendicular to the scattering plane , to negative , or parallel to the scattering plane ) seen in our simulations also occurs for the @xmath53 values presented in simmons ( 1982 ) , figure 5b . for astronomical silicates , polarization reversal occurs at a value somewhere between @xmath71 and @xmath72 , or @xmath73 ( see figure [ fig - p_vs_x ] ) , in agreement with figure 5b of simmons ( 1982 ) . for dirty ice , we do not see any polarization reversal , which follows the figure 5b of simmons ( 1982 ) , except for our biggest grains ( @xmath74 , @xmath75 ) at low inclinations ( @xmath76 ) . for the grains with large imaginary refractive index values ( graphite and amorphous carbon ) , we see from table 1 of daniels ( 1980 ) that polarization reversal is not expected at all , which is what is seen here , except in 2 cases for intermediate inclinations . we also studied a distribution of grain sizes , using the mrn size distribution @xmath77 ( mathis , rumpel , & nordsiek 1977 ) for astronomical silicates with sizes between @xmath78 and @xmath79 . the results are not very different from the case where only one grain size was considered . since all polarimetric curves have the same morphology for a given set of parameters except grain size , adding many of those curves only changes the polarization level , not the morphology . as the inclination decreases from @xmath80 ( edge - on ) to @xmath81 ( pole - on ) , the polarization decreases , in general following a theoretical law ( developed for a single star at the center of an axisymmetric envelope but nonetheless interesting for our case ) that can be found for example in brown & mclean ( 1977 ) : the residual polarization ( sum of polarization of the scattered light and the unpolarized light from the star ) scales as @xmath82 . deviations from this theoretical law are found when the optical depth is high ( @xmath83 , too high to consider single scattering only ) or the grains are big ( @xmath84 , which probably introduces higher frequency variations in the polarization as a function of the scattering angle ) . with a @xmath80 inclination , the @xmath35 parameter is null , and only the @xmath34 parameter varies . as the inclination decreases , both the @xmath34 and @xmath35 parameters vary , and more so as the inclination decreases . in general , the behavior is the same as for thomson scattering . the simulations for a binary star in a circumbinary envelope give polarimetric variations that are double sine waves ( see figure [ fig - binary_as_t01_a01 ] ) , which translate to an ellipse that is traced out twice per orbit in the @xmath42 plane . the average stokes parameter @xmath35 is zero , as is expected from a geometry oriented in the plane of the sky in a east west direction , and the position angle is usually @xmath85 , perpendicular to the projection of the envelope s main axis and scattering angle . in order to decrease the number of parameters influencing the polarization and isolate causes and their effects , we can take a look at the contribution of only one star instead of considering the combined effects of the 2 stars . in figure [ fig - primary_as_t01_a01 ] we show the polarimetric variations for the primary only , as if it were in orbit around an `` invisible '' , or much fainter , secondary star , both stars being in orbit at the center of the same circumbinary shell . even though the orbit is circular , single - periodic variations are seen in addition to the usual double - periodic ones . we have shown in paper i that variable absorption effects introduce such single - periodic variations ; but if we neglect those variable absorption effects and impose a constant optical depth in our mie simulations , those single - periodic variations persist . this is due to the asymmetric phase function of dust grains . it is also seen that the peaks of polarization are not located exactly at phases 0.25 and 0.75 , as expected for an asymmetric phase function . therefore , in addition to variable absorption effects , non equal mass stars , orbital eccentricity , and asymmetric envelopes , mie scattering can also cause single - periodic variations . like the thomson scattering results presented in paper i , the amplitude of the variations increases significantly for larger size orbits . for an orbit that comes close to the inner edge of the circumbinary disk , the polarimetric variations are more apparent . in order to see the effects of geometries different than a circumbinary ellipsoidal envelope , we made simulations for flared circumbinary disks , prolate circumbinary envelopes , and circumstellar disks . we also investigated the effects of a non - coplanar geometry where the plane of symmetry of the envelope and the orbital plane differ from one another . first , we tried a geometry similar to a flared disk , with a conic opening ( with half - angle between 10 and @xmath86 ) in an ellipsoidal envelope . as the angle of the cone increases , so does the polarization and amplitude of the variations . we also studied prolate geometries instead of oblate ones , orienting the long axis of the envelope along 3 axes ( toward the observer , in a n s and e w orientation in the plane of the sky ) . finally , we inclined the envelope by @xmath87 with respect to the orbital plane . for all those cases , polarization and amplitude of variations are affected to various degrees , but the polarimetric variations have the same morphology . for circumbinary envelopes , the polarimetric curves we have produced so far seldom have an amplitude greater than 0.10% , whereas simmons ( 1983 ) could produce amplitude of many tenths of a percent , or even of a few percents , with a geometry more favorable for high amplitude polarimetric variations : a single circumstellar envelope externally illuminated by only one star . indeed , the most interesting simulations were done with a circumstellar disk around the primary , a `` naked '' secondary which does not have any circumstellar material in its environment , and no circumbinary disk . see figure [ fig - cs_45_1_5 ] for an example of such a geometry . as expected from this more asymmetric geometry , the polarimetric variations produced have a greater amplitude , ranging from 0.1 to 0.3% . the primary star that sits in the middle of its circumstellar disk does not produce any variation , as expected , but numerical noise is present at a level of @xmath88 % in polarization . variations for the secondary star , which revolves around the primary and its disk , are a sum of double- and single - periodic variations . average polarization , amplitude of the polarimetric variations , and morphology of the variability are all interesting , but periodic polarimetric observations can potentially give an additional and very interesting parameter for a binary star : its orbital inclination , which is found by using the formalism developed by bme ( equations [ eq - io1-p2 ] and [ eq - io2-p2 ] ) . we now investigate if their formalism , which was developed for thomson scattering , circular orbits , and constant optical depth , can be extended successfully to our simulations to find the orbital inclination . for an optical depth of 0.1 , we have found that the bme formalism will work ( i.e. , true inclinations can be found from the bme analysis of the polarimetric variations if @xmath89 ) as long as there are polarimetric variations of sufficient amplitudes for the mathematical analysis ( by sufficient , we mean for example , 0.002% for simulations without stochastic noise ) . the composition or the size of the grains is not important as long as some variations are introduced . big grains ( @xmath90 ) do not produce enough variations for the bme analysis to work properly . in general , the uncertainty on the inclination found by the bme formalism ( calculated with the method of propagation of errors ) is higher for dust grains than for electrons . this is understandable since , as stated before , grains produce smaller amplitude polarization variations , which will increase the uncertainty on the inclination found . the effects of optical depths were studied for some combinations of grain composition and size , with optical depths of 0.02 , 0.05 , 0.2 , and 0.5 . as expected , the polarization increases with the optical depth , as does the amplitude of the polarimetric variations . all the simulations studied here produced polarimetric variations with sufficiently large amplitudes , so again , the bme formalism works for @xmath7 . berger & mnard ( 1997 ) have studied multiple mie scattering in circumbinary shells with optical depths up to @xmath91 , and have found that even with such a high optical depth , the estimation of inclination from the bme formalism is still very good . although the inclination found by the bme equations is not significantly affected by the geometry in the case of the flared disk geometry , it is more seriously affected for prolate geometries and non - coplanar orbital and geometric planes . for the prolate envelope inclined at @xmath92 , the two cases where the long axis of the prolate envelope is in the plane of the sky give inclinations close to @xmath92 , but for the long axis coming toward the observer , the inclination found is @xmath93 too low . for the same inclination , the envelope inclined by @xmath87 with respect to the orbital plane gives orbital inclinations of @xmath94 or @xmath95 , significantly different than the @xmath96 found for the coplanar case . for the circumstellar disks , bme can find the inclination if it is @xmath97 . the lower threshold , compared to the circumbinary case , might be due to the amplitude of the variations , which are greater for the circumstellar envelope geometry than for the circumbinary envelope geometry . to investigate the possible effects of non - circular orbits on polarimetric observations , canonical simulations were performed with orbital eccentricities of 0.1 , 0.3 , and 0.5 . it should be noted that in these simulations , the orbital semi - major axes have been adjusted according to the eccentricity ( i.e. , decreased with increasing @xmath98 ) so the stars came at about the same distance from the inner edge of the envelope , irrespective of the eccentricity of their orbit ( the maximum distance between the center of the envelope and any of the two stars was always @xmath99 ) . this was done so we could study the influence of the eccentricity without having to deal with the influence of how close the stars come to the inner edge of the circumbinary envelope . the results found for mie scattering are identical to those we have found for thomson scattering ( see paper i ) . the eccentricity has no significant influence on the average level of polarization , but does change the amplitudes of the polarimetric variations : the amplitudes decrease as the eccentricity increases . as the eccentricity increases , the @xmath43 variations that are dominant for low eccentricities give way to @xmath45 variations . moreover , higher harmonics appear ; a fit with 1@xmath28 and 2@xmath28 is not sufficient to reproduce the polarimetric curves , so 3@xmath28 and 4@xmath28 harmonics are needed . when looking at the inclinations found by the bme formalism with the second - order coefficients of the fit , we find that the true inclinations are found for the lower eccentricities ( @xmath100 ) and the highest inclinations ( @xmath89 ) . for the highest eccentricity studied here , @xmath101 , the second - order coefficients can not be used to find the orbital inclination , as the bme formalism is unable to find the true inclination . for all eccentricities , even the highest one studied here @xmath101 , the first - order coefficients can be used to find almost all inclinations . there is a numerical problem with the true inclination of 80 , for which the bme formalism can not assign the right inclination . the lowest inclinations ( @xmath102 ) are harder to find . when the periastron is changed to values other than 0 , the results of the bme analysis are slightly modified . with an eccentricity of 0.5 , the second - order and first - order coefficients now both give reasonable results for @xmath89 . in future papers , we will present polarimetric observations of binary young stars , which are objects surrounded by circumstellar material ( dust grains ) . the binaries we have selected have short periods , so we believe the geometry adopted here ( circumbinary ellipsoidal envelope with a central cavity ) is suitable for these kinds of objects . some of the observed pre - main sequence binaries show periodic polarimetric variations , although they are in general of lower amplitude and less clear than those of binary hot stars ( surrounded by electrons ) , such as wolf rayet stars ( see , for example , st - louis et al . 1988 ; drissen et al . 1989b ; robert et al . this is in agreement with our previous comparison of the polarimetric properties of electrons and dust grains . non - periodic polarimetric variations that are known to exist even for single young stars ( bastien 1982 ; drissen , bastien , & st - louis 1989a ; mnard & bastien 1992 ) produce stochastic noise in the polarimetric curves of binary young stars , sometimes hiding the low amplitude periodic variations . the small amplitude variations ( @xmath103% ) often seen in binned data can be caused by non favorable inclinations or geometry , or not enough scatterers , but we believe it is in general agreement with one of our conclusions , that dust grains produce smaller amplitude variations than electrons . some stars do have high amplitude polarimetric variations of up to 0.7% . such large amplitudes were not quite produced by our simulations , although circumstellar disks , a more favorable geometry than circumbinary disks , do produce more substantial variations , up to @xmath104% . for the observed high amplitude polarimetric variations , a circumbinary envelope might thus not be an adequate geometry and circumstellar envelope(s)/disk(s ) could be present . we have presented numerical simulations of the periodic polarimetric variations produced by a binary star placed at the center of an empty cavity in a circumbinary ellipsoidal and optically thin envelope . mie single - scattering in envelopes made of single - size and single - composition grains was considered . the orbits were circular or eccentric . the mass ratio ( and luminosity ratio ) was equal to 1.0 . these parameters are to represent short - period spectroscopic binary young stars that are embedded in a circumbinary envelope and have evacuated the central regions of this envelope due to gravitational or radiative interactions . we have shown that the absolute amplitude of the variations is smaller for mie scattering than for thomson scattering , which will render more difficult the detection of polarimetric variations in binary stars surrounded by dust grains . in fact , polarimetric observations of binary young stars that will be presented in future papers show periodic variations that are sometimes much less obvious than for hot stars , for example . this may be due in part to the scattering properties of dust grains , although other factors , such as orbital inclination or scattering geometry , can also decrease ( or increase ) the amplitude of the polarimetric variations . the average polarization produced depends on the grains composition ( through the real and imaginary parts of the refractive index ) and size , and on the wavelength of observation ; a highly polarizing grain at a given grain size will not necessarily polarize as efficiently if it is smaller or bigger . for the four grain compositions that we have studied ( astronomical silicates , graphite , amorphous carbon , and dirty ice ) , the highest polarizations are produced by grains with sizes in the range @xmath105@xmath1 ( @xmath106@xmath3 for @xmath107 ) . polarization reversal was seen for astronomical silicates ( at @xmath108 , or @xmath109 ) , but not for dirty ices , graphite , and amorphous carbon grains , which is in agreement with previous works ( simmons 1982 ; daniels 1980 ) . the periodic variations produced are in general double - periodic ( seen twice per orbit ) , although we have shown that variable absorption effects ( pre- and post - scattering factors ) , non equal mass stars , orbital eccentricity , asymmetric envelopes , and mie scattering ( on spherical dust grains , which have an asymmetric scattering function ) introduce single - periodic variations ( seen once per orbit ) . circumstellar disks are more interesting than circumbinary disks , in the sense that they produce polarimetric variations that are more readily detectable ( @xmath5% versus @xmath6% ) . other geometries ( flared disks , prolate envelopes , non - coplanar disks ) did not produce any difference in the morphology or amplitude of the variations . we have shown that the bme formalism is still valid beyond its original limits ( circular orbits , no extinction effects , thomson scattering ) , under certain circumstances . in general , the bme formalism will be able to find the orbital inclination if this inclination is @xmath110 and if the amplitude of the polarimetric variations is sufficient . the composition or the size of the grains is not important as long as some variations are introduced , but the inclinations found by the bme formalism seem to be affected by prolate and non - coplanar geometries . the second - order fourier coefficients of the fit can be used for the lower orbital eccentricities ( @xmath100 ) and the highest inclinations ( @xmath89 ) . for all eccentricities , even the highest one studied here @xmath101 , the first - order coefficients can be used to find almost all inclinations . comparisons between polarimetric observations of binary stars and these simulations will help understand these systems ( presence of circumbinary and/or circumstellar disks , geometry of the envelope , orbital eccentricity , etc . ) although the interpretations might not be unique , as many parameters can have similar effects on the polarimetric variations ( mainly the amplitude and the presence of single - periodic variations ) . in that case , other types of observations ( imaging and spectroscopy ) would nicely complement polarimetric ones . n. m. would like to thank the conseil de recherche en sciences naturelles et gnie of canada , the fonds pour la formation de chercheurs et laide la recherche of the province of qubec , the facult des etudes suprieures and the dpartement de physique of universit de montral for scholarships , and p. b. for financial support . n. m. would also like to thank f. mnard for numerous discussions . we would like to thank the conseil de recherche en sciences naturelles et gnie of canada for supporting this research . bastien , p. 1982 , , 48 , 153 berger , j .- p . , & mnard , f. 1997 , in iau symposium 182 , low mass star formation from infall to outflow , ed . f. malbet & a. castets , 201 brown , j. c. , & mclean , i. s. 1977 , , 57 , 141 brown , j. c. , mclean , i. s. , & emslie , a. g. 1978 , , 68 , 415 brown , j. c. , aspin , c. , simmons , j. f. l. , & mclean , i. s. 1982 , , 198 , 787 daniel , j .- y . 1978 , , 67 , 345 daniel , j .- y . 1980 , , 87 , 204 drissen , l. , bastien , p. , & st - louis , n. 1989a , , 97 , 814 drissen , l. , robert , c. , lamontagne , r. , moffat , a. f. j. , st - louis , n. , van weeren , n. , & van genderen , a. m. 1989b , , 343 , 426 draine , b. t. 1985 , , 57 , 587 greenberg , j. m. 1968 , in nebulae and interstellar matter , ed . b. m. middlehurst & l. h. aller ( chicago : university of chicago press ) , 221 manset , n. , & bastien , p. 2000 , , 120 , 413 ( paper i ) mathis , j. s. , rumpl , w. , & nordsiek , k. h. 1977 , , 217 , 425 mnard , f. , & bastien , p. 1992 , , 103 , 564 robert , c. , moffat , a. f. j. , bastien , p. , st - louis , n. , & drissen , l. 1990 , , 359 , 211 rouleau , f. , & martin , p. g. 1991 , , 377 , 526 simmons , j. f. l. 1982 , , 200 , 91 simmons , j. f. l. 1983 , , 205 , 153 stassun , k , & wood , k. 1999 , , 510 , 892 st - louis , n. , moffat , a. f. j. , drissen , l. , bastien , p. , & robert , c. 1988 , , 330 , 286 van de hulst , h. c. 1981 , light scattering by small particles ( new york : dover publications , inc . ) wickramasinghe , n. c. 1967 , interstellar grains , ( london : chapman and hall ltd . ) wickramasinghe , n. c. , & guillaume , c. 1965 , , 207 , 366 wood , k. , bjorkman , j. e. , whitney , b. a. , & code , a. d. 1996 , , 461 , 828 llll grain composition & @xmath111 & @xmath112 & @xmath113 + be ( amorphous carbon ) & 2.240 & 0.781 & 2.37 + graphite & 2.524 & 1.532 & 2.95 + dirty ice & 1.33 & 0.09 & 1.34 + astronomical silicate & 1.715 & 0.030 & 1.72 + lllllll @xmath114 & & electrons & astron . & graphite & amorphous & dirty ice + & & & silicates & & carbon & + & @xmath115 ( % ) & 0.794 & 0.619 & 0.299 & 0.270 & 0.189 + 90@xmath116 & @xmath117 ( % ) & 0.0063&0.0048&0.0023&0.0021&0.0015 + & @xmath118&0.0079&0.0075&0.0077&0.0078&0.0079 + & @xmath115 ( % ) & 0.593 & 0.461 & 0.222 & 0.201 & 0.140 + 60@xmath116 & @xmath117 ( % ) & 0.0079&0.0061&0.0029&0.0027&0.0018 + & @xmath118&0.0133&0.0132&0.0131&0.0134&0.0129 + & @xmath115 ( % ) & 0.395 & 0.306 & 0.147 & 0.133 & 0.093 + 45@xmath116 & @xmath117 ( % ) & 0.0095&0.0073&0.0035&0.0032&0.0022 + & @xmath118&0.0241&0.0239&0.0238&0.0241&0.0237 +	following a previous paper on thomson scattering , we present numerical simulations of the periodic polarimetric variations produced by a binary star placed at the center of an empty spherical cavity inside a circumbinary ellipsoidal and optically thin envelope made of dust grains . mie single - scattering ( on spherical dust grains ) is considered along with pre- and post - scattering extinction factors which produce a time - varying optical depth and affect the morphology of the periodic variations . the orbits are circular or eccentric . the mass ratio ( and luminosity ratio ) is equal to 1.0 . we are interested in the effects that various parameters ( grain characteristics , geometry of the envelope , orbital eccentricity , etc . ) will have on the average polarization , the amplitude of the polarimetric variations , and the morphology of the variability . we show that the absolute amplitudes of the variations are smaller for mie scattering than for thomson scattering , which makes harder the detection of polarimetric variations for binary stars surrounded by dust grains . the average polarization produced depends on the grains composition and size , and on the wavelength of observation . among the four grain types that we have studied ( astronomical silicates , graphite , amorphous carbon , and dirty ice ) , the highest polarizations are produced by grains with sizes in the range @xmath0@xmath1 ( @xmath2@xmath3 for @xmath4 ) . composition and size also determine if the polarization will be positive or negative . in general , the variations are double - periodic ( seen twice per orbit ) . in some cases , because spherical dust grains have an asymmetric scattering function , the polarimetric curves produced show single - periodic variations ( seen once per orbit ) in addition to the double - periodic ones . a mixture of grains of different sizes does not affect those conclusions . circumstellar disks produce polarimetric variations of greater amplitude ( up to @xmath5% in our simulations ) than circumbinary envelopes ( usually @xmath6% ) . other geometries ( circumbinary flared disks or prolate envelopes , and non - coplanar envelopes ) do not present particularly interesting polarimetric characteristics . another goal of these simulations is to see if the 1978 bme ( brown , mclean , & emslie ) formalism , which uses a fourier analysis of the polarimetric variations to find the orbital inclination for thomson - scattering envelopes , can still be used for mie scattering . we find that this is the case , if the amplitude of the variations is sufficient and the true inclinations is @xmath7 . for eccentric orbits , the first - order coefficients of the fourier fit , instead of second - order ones , can be used to find almost all inclinations .
the multiple polylogarithm functions were defined in goncharov s paper @xcite as the following power series : @xmath7 where the @xmath8 are in the unit disk of the complex plane for @xmath9 and @xmath10 are positive integers . for @xmath11 , we get the classical @xmath3-th polylogarithm which was first introduced by leibniz @xcite in 1696 : @xmath12 and for @xmath13 , we obtain the multiple @xmath6-values which were first studied by euler @xcite : @xmath14 moveover , multiple polylogarithms can be represented as iterated integrals . recall that iterated integrals are defined as follows . let @xmath15 be smooth one - forms on a manifold @xmath16 and @xmath17 \rightarrow m$ ] be a piecewise smooth path . then we define inductively as follows : @xmath18 where @xmath19 is the restriction of @xmath20 on @xmath21 $ ] and @xmath22 is a function of @xmath23 on @xmath24 $ ] . more explicitly , it can be computed in the following way : @xmath25 where @xmath26 are the pullback one - forms on @xmath24 $ ] , @xmath27 . for example , @xmath28 in ( * ? ? ? * chap2 ) , the following formula was proved : @xmath29 @xmath30 this formula also provides the analytic continuation of multiple polylogarithms . we denote by @xmath31 the moduli space of @xmath32-labeled pointed stable curves of genus @xmath4 , where @xmath32 is a finite set . it s been studied by grothendieck @xcite , deligne , mumford , knudsen @xcite and many others . it is defined over @xmath33 . roughly speaking , a complex point of @xmath31 is a tree of complex projective lines with @xmath34 distinct smooth points marked by the set @xmath32 . here @xmath34 denotes the cardinality of the set @xmath32 . we know that it is a smooth irreducible projective variety of complex dimension @xmath35 . moreover , @xmath36 provides a natural compactification of the space @xmath37 of @xmath34 distinct points on @xmath38 modulo automorphisms of @xmath38 . by cross - ratio , @xmath37 is isomorphic to @xmath39 the boundary @xmath40 is a normal crossing divisor . it can be described by the combinatorial data of the set @xmath32 . for more detail , see section @xmath41 . now given any subset @xmath42 , with @xmath43 , there is a contraction morphism : @xmath44 which contracts stably all sections but those marked by @xmath45 . in particular , for any @xmath46 , let @xmath47 , the contraction morphism @xmath48 is the universal @xmath49-labeled curve with universal sections @xmath50 , for each @xmath51 . for more information and proofs of @xmath31 , we refer to @xcite . if the set @xmath52 , we ll denote this space by @xmath2 . and @xmath31 is non - canonically isomorphic to @xmath53 . from now on , we ll take @xmath54 + @xmath55 and fix the cyclic order @xmath56 on @xmath32 unless otherwise stated . in @xcite , for each multiple @xmath6-value ( [ mz ] ) , the authors construct two divisors @xmath0 and @xmath1 of @xmath31 and then show that @xmath57 is a framed mixed tate motive whose period is this value . in the end of that paper , they suggest to generalize their results to the following convergent iterated integral : @xmath58 where @xmath17 \rightarrow \mathbb{c}$ ] is a piecewise smooth simple path from @xmath4 to @xmath59 and @xmath60 , @xmath61 . in particular , by the formula ( [ mul - li ] ) , multiple polylogarithms are of this type . in this paper , we show that the analogous results hold for the iterated integral ( [ int1 ] ) . in section @xmath41 , we review the basic combinatorial facts about the boundary divisors of @xmath31 and the stable @xmath41-partitions of the set @xmath32 . next , we briefly recall the divisor @xmath62 in @xmath31 which was introduced in @xcite , and then prove some interesting combinatorial properties of @xmath62 . in the end , we proceed to study in detail some non - boundary divisors of @xmath31 which we ll use later on . in section @xmath63 , for the integral ( [ int1 ] ) , we define a meromorphic differential form @xmath64 of @xmath36 . let @xmath65 be its divisor of singularities in @xmath36 . we explicitly determine the divisor @xmath65 . then we use it to prove the key proposition that _ the divisor @xmath65 does not contain any @xmath66-dimensional face of the divisor @xmath62 , @xmath67_. in section @xmath68 , we review the definitions of framed hodge - tate structure and its period , and discuss their basic properties . finally combining with all the information of @xmath65 and @xmath62 in section @xmath63 and @xmath68 , we can prove : [ main ] let @xmath69 . for the iterated integral @xmath70 of ( [ int1 ] ) , @xmath71 , there exists two divisors @xmath65 and @xmath62 in @xmath31 , @xmath72 , such that @xmath73 carries an @xmath3-framed hodge - tatestructure with two canonical frames @xmath74}\in gr_{2n}^{w}h^{n}(\mathcal{\overline{m}}_{0,s}-a_s(\vec{a } ) ) ; \ ; { [ \delta_{b}(\gamma)]}\in ( gr_{0}^{w}h^{n}(\mathcal{\overline{m}}_{0,s},b_n))^{\vee}.\ ] ] and the period with respect to these frames is exactly the iterated integral @xmath70 , where @xmath75 $ ] is a meromorphic @xmath3-form on @xmath76 , and @xmath77 is a relative @xmath3-cycle . furthermore , if the @xmath78 are elements of a number field @xmath79 , @xmath61 , then @xmath80 is a framed mixed tate motive over @xmath79 . in section @xmath81 , we apply our construction to the dilogarithm . namely , we consider the following integral : @xmath82 by the theorem [ main ] above , @xmath83 carries a @xmath41-framed hodge - tate structure . we calculate it and prove that : [ thm2 ] the mixed hodge structure given in theorem [ main ] for the dilogarithm , that is , @xmath83 , is isomorphic to the one given by p. deligne . and for our case the period matrix is the following : + if @xmath84 , it equals @xmath85\ ] ] if @xmath86 , it is @xmath87\ ] ] with above theorem [ thm2 ] , the interesting question is that for the classical @xmath3-th polylogarithm ( @xmath88 ) , whether or not our construction is isomorphic to the canonical one given by deligne . the general situation is more delicate than the dilogarithm case . we have some results and it seems that they are not isomorphic if @xmath88 . more detail will appear in @xcite . another construction of the multiple polylogarithm motives has been given by a.goncharov in @xcite where he uses another sequence of blowups . as * framed * mixed tate motives , the two constructions should be equivalent . but our construction is canonical and more natural . first recall that there is a one - to - one correspondence between the boundary divisors of @xmath31 and the stable unordered @xmath41-partitions of the set @xmath32 . let @xmath89 and @xmath90 . we ll denote by @xmath91 the corresponding boundary divisor . [ def1 ] let @xmath92 with the given cyclic order @xmath93 . a subset @xmath0 of @xmath94 is called strictly ordered if there exists some @xmath95 and @xmath96 a positive integer , such that @xmath97 ( the subscripts are counted mod @xmath66 ) . that is , its elements are in consecutive order with respect to @xmath93 . given a @xmath41-partition @xmath98 of @xmath94 , @xmath99 , we say that @xmath98 is strictly ordered with respect to @xmath93 if one of the @xmath100 is a strictly ordered subset of @xmath94 . for example , take @xmath101 , then @xmath102 is strictly ordered , but @xmath103 is not . now consider the open @xmath3-simplex @xmath104 . as mentioned in the introduction , via cross - ratio , @xmath37 is identified with the subset @xmath105 of @xmath106 . under this identification , @xmath107 is a subset of @xmath37 . thus we have a natural map @xmath108 which embeds @xmath107 into @xmath36 . let @xmath109 be the zariski closure of the boundary of the closure of @xmath110 in @xmath36 , which is called algebraic stasheff polytope ( see @xcite ) . then we have the following : [ prop1 ] @xmath109 is a union of boundary divisors indexed by the stable @xmath41-partitions of @xmath32 which are strictly ordered with respect to the cyclic order @xmath93 . that is , they correspond to breaking a circle into two connected arcs . furthermore , @xmath109 is an `` algebraic stasheff polytope '' . i.e. , there is a bijection between the irreducible components @xmath111 of @xmath109 and the codimension one faces @xmath112 of the stasheff polytope @xmath113 such that a subset of @xmath111 s has a non - empty intersection of expected codimension if and only if the respective subset of @xmath112 s has this property . see @xcite , proposition 2.1 . for two unordered stable @xmath41-partitions @xmath114 clearly , @xmath115 or @xmath68 , and @xmath116 if and only if @xmath117 , which implies that the boundary divisors @xmath118 . [ lemma1 ] ( 1 ) @xmath119 if and only if @xmath120 and @xmath121 . + ( 2 ) @xmath122 if and only if @xmath123 . \(1 ) by definition , @xmath119 if and only if one of the following holds : @xmath124 by the fact 4 in @xcite , this is exactly the sufficient and necessary condition for that @xmath120 and @xmath125 . \(2 ) by definition , @xmath122 means that there are distinct elements @xmath126 such that @xmath127 , @xmath128 , @xmath129 , @xmath130 . then it follows from the proof of the fact 4 in @xcite . a family of stable @xmath41-partitions @xmath131 of @xmath32 is called good if @xmath132 , for @xmath133 . a family of boundary divisors @xmath134 is called compatible , if the corresponding partitions @xmath135 are good . [ lemma2 ] given @xmath136 pairwise distinct boundary divisors @xmath137 , then @xmath138 if and only if @xmath131 is good . first , if @xmath139 is not good , then @xmath140 for some @xmath133 . since @xmath141 and @xmath142 are distinct , by lemma [ lemma1 ] , @xmath143 hence , @xmath144 now suppose that @xmath139 is good . the by ( * ? ? ? * proposition 3.5.2 ) , there exists a stable @xmath32-tree @xmath145 with @xmath136 ( internal ) edges @xmath146 such that @xmath147 , where @xmath148 is the stable @xmath41-partition of @xmath32 corresponding to the one edge @xmath32-tree obtained by contracting all ( internal ) edges but @xmath149 . denote by @xmath150 the closure of the stratum parametrizing the stable curves of the combinatorial type @xmath145 . then for each @xmath151 , we have @xmath152 . hence , @xmath153 . and since the boundary divisors meet transversely , we can conclude that in this case @xmath154 . now fix an @xmath155 , and let @xmath156 with the induced cyclic order @xmath157 . let @xmath158 be the contraction morphism which contracts the section marked by @xmath159 . similarly , we have the natural embedding of @xmath160 into @xmath161 . and we also have the corresponding algebraic stasheff polytope @xmath162 . now we have the following key property : [ prop3 ] under the contraction morphism @xmath163 , the vertices of @xmath109 are mapped to the vertices of @xmath162 . let @xmath164 be a vertex of @xmath62 . then by proposition [ prop1 ] , there exists exactly @xmath3 compatible boundary divisors of @xmath62 , say , @xmath165 such that @xmath166 . consider the images @xmath167 in @xmath161 , for @xmath168 . there are two possible cases for the @xmath169 s : case 1 : @xmath170 ; or @xmath171 ; + case 2 : @xmath172 , with @xmath43 and @xmath173 . in case 1 , we get @xmath174 and in case 2 , we have @xmath175 where @xmath176 is a stable partition of @xmath49 , and @xmath177 is an irreducible component of @xmath162 . it s clear that if @xmath178 belong to case 2 , then @xmath179 . now we claim that among the @xmath180 , @xmath181 , @xmath182 , there exists exactly one of them belonging to case 1 . indeed , first suppose all of them belong to case 2 . then we have @xmath3 pairwise distinct boundary divisors of @xmath183 and their intersection is empty due to the dimension consideration . but @xmath184 is contained in this intersection , contradiction . secondly , suppose there are two of them in case 1 . say they are @xmath185 and @xmath186 . but then @xmath187 , thus @xmath188 . it contradicts that @xmath166 . hence the claim is proved . by the claim , we see that @xmath184 is contained in the intersection of @xmath189 distinct compatible divisors of @xmath162 . by proposition [ prop1 ] , @xmath184 is a vertex of @xmath162 . using proposition [ prop3 ] , we can prove the following : [ prop4 ] let @xmath190 be the contraction morphism which forgets all sections but the ones marked by @xmath191 . then the images of the vertices of @xmath109 are contained in the set @xmath192 . under the identification of @xmath193 with @xmath194 , @xmath195 is the interval @xmath24 $ ] . thus the vertices are @xmath4 , @xmath59 . the map @xmath196 is a composition of @xmath189 contraction morphisms , each of which forgets only one section of those labeled by the subset @xmath197 . now we can apply proposition [ prop3 ] repeatedly to those @xmath189 contraction morphisms . therefore , the images of the vertices of @xmath109 are either @xmath4 or @xmath59 . next , we ll introduce some non - boundary divisors in @xmath31 and prove some properties which we need later on . for each @xmath198 , consider the morphism @xmath199 which contracts all sections but those labeled by @xmath200 . now let @xmath201 , that is , it lies in the open stratum @xmath202 . then the inverse image @xmath203 is a divisor of @xmath31 . the following proposition describes how @xmath203 intersects the boundary divisors of @xmath31 . [ nonb1 ] let @xmath204 be a stable @xmath41-partition of @xmath32 and @xmath91 be the corresponding boundary divisor . then : + ( 1 ) if @xmath205 and @xmath206 , @xmath207 and @xmath208 , or @xmath209 and @xmath210 belong to different parts of @xmath98 , then @xmath211 ( 2 ) otherwise , we can assume that @xmath212 . then:@xmath213 where @xmath214 is some fixed element of @xmath180 and @xmath215 is the map contracting all sections but those marked by @xmath216 ; and if @xmath217 , @xmath23 is some fixed element of @xmath218 ; otherwise , @xmath219 , then @xmath23 is the only element of @xmath220 not in @xmath180 . for the first case , it s easy to check that the image of @xmath91 under @xmath221 is contained in the set @xmath222 . but @xmath223 , thus @xmath224 for the second case , consider the morphism:@xmath225 which is the product of contracting all sections labeled by @xmath218 but @xmath23 and contraction of all sections labeled by @xmath180 but @xmath214 . by the fact 2 of @xcite , the restriction of @xmath226 on @xmath227 is an isomorphism which is independent of the choices of @xmath214 and @xmath23 . then we have the following commutative diagram : @xmath228 where @xmath229 and @xmath230 are the restrictions of the maps @xmath226 and @xmath231 on the divisor @xmath91 respectively , and @xmath232 is the projection on the first factor . thus we obtain that : @xmath233 since @xmath229 is an isomorphism , we have @xmath234 via the map @xmath226 . next we have the following description of @xmath203 : [ nonb2 ] ( 1 ) . @xmath203 is an irreducible , reduced , smooth divisor of @xmath31 . it is a tate variety , that s , its motive is a direct sum of pure tate motives . we ll use the induction on @xmath3 . when @xmath235 , @xmath236 and @xmath237 is the identity map . and the proposition is obvious . let s suppose that it s true for @xmath238 , @xmath239 , and consider @xmath240 . now @xmath241 . let @xmath242 . then we have the following commutative diagram : @xmath243 where @xmath244 is the map forgetting the section @xmath245 , @xmath20 is the product of @xmath244 and the morphism from @xmath31 to @xmath246 contracting all sections but those marked by @xmath247 ; and @xmath248 is the projection on the first factor . moreover , we have:@xmath249 hence,@xmath250 it s known that @xmath251 is the universal curve on @xmath183 , so @xmath244 has geometrically reduced fibers , and by induction,@xmath252 is reduced , thus @xmath240 is reduced . in @xcite , keel proved that the map @xmath20 in the above diagram is isomorphic to a sequence of blowups of @xmath253 along all the boundary divisors of @xmath183 . from the above diagram , we see that : @xmath254 by the explicit blowups given in @xcite , @xmath240 is the composition of blowups of @xmath255 along the intersections of @xmath203 with the boundary divisors . by the proposition [ nonb1 ] , these intersections are either empty or of the form : @xmath256 where @xmath257 and @xmath258 . by the induction assumption , @xmath259 is irreducible and smooth , and @xmath255 is an irreducible smooth divisor of @xmath253 . altogether we see that @xmath240 is the composition of blowups of an irreducible smooth variety along an irreducible smooth subvariety . therefore , it is irreducible and smooth . and this proves the first part of the proposition . for the second part , we need the following formula of the motive of the blowup of a variety along a subvariety . this was proved by manin in ( * ? ? ? * corollary , page 463 ) . let @xmath260 be the blowup of a variety @xmath261 along a subvariety @xmath262 of codimension @xmath263 . then we have : @xmath264 where @xmath265 , @xmath266 and @xmath267 denote the motives of @xmath260 , @xmath261 and @xmath262 respectively ; @xmath268 is the tate motive , @xmath269 . therefore , if both @xmath261 and @xmath262 are tate varieties , by the formula above , the blowup @xmath260 is also a tate variety . because the product of tate varieties is also a tate variety , and by induction , @xmath240 is the composition of blowups of a tate variety along a tate subvariety . thus it s a tate variety . this proves the second part . now we consider the intersections of the divisors @xmath203 . let @xmath270 , and @xmath271 . then we have : [ nonb3 ] the intersection @xmath272 is a smooth irreducible subvariety ; it s also a tate variety . first let s look at the case @xmath273 . notice for any irreducible boundary divisor @xmath91 , where @xmath98 is some stable @xmath41-partition of @xmath32 , we have @xmath274 by the proposition [ nonb1 ] , @xmath275 is either empty or of the form : @xmath276 here we use the notations in proposition [ nonb1 ] . now by induction on @xmath3 , we see that @xmath277 is obtained by the blowups of smooth irreducible tate varieties along the tate subvarieties . hence it s also smooth , irreducible and tate . when @xmath278 , similarly for any irreducible boundary divisor @xmath91 , we have @xmath279 then similarly , use proposition [ nonb1 ] and induction on @xmath3 , it s true for any @xmath136 . [ nonb4 ] let @xmath280 be stable @xmath41-partitions of @xmath32 . for each @xmath281 , @xmath141 is the corresponding boundary divisor . then the intersection @xmath282 is either empty or a smooth irreducible tate variety . by the combinatorial description of the intersection @xmath283 , ( for @xmath284 , see ( * ? ? ? * fact 4 ) , and for general @xmath66 , it can be done inductively ) we see that the non - empty intersection in the corollary can be written as the product of irreducible smooth tate varieties . thus , it s also an irreducible smooth tate variety . let @xmath286 and consider the following @xmath3 contraction morphisms : @xmath287 which forget all sections but those labeled by @xmath288 . we define the meromorphic @xmath3-form @xmath285 of @xmath31 as follows : @xmath289 furthermore , we have another useful description of @xmath285 . let @xmath290 be the birational proper morphism of the products of @xmath291 , @xmath61 . and let @xmath292 be the affine coordinates of @xmath293 , then by definition , we have @xmath294 we define @xmath295 to be the divisor of singularities of @xmath285 . following @xcite , for @xmath296 , we define @xmath297 by @xmath298 and @xmath299 we say that a @xmath41-partition of @xmath32 has _ type @xmath300 _ if one part of it is of form @xmath301 where @xmath94 is a non - empty subset of @xmath297 . now we have the explicit description of the divisor @xmath295 : [ divisor - a ] the divisor @xmath295 of singularities of @xmath64 on @xmath31 consists of the following irreducible components : + ( a ) the boundary divisors @xmath91 corresponding to those stable @xmath41-partitions of @xmath32 which have some type @xmath300 , @xmath302 ; + ( b ) the non - boundary divisors @xmath303 , for each @xmath304 ; where @xmath305 is the morphism contracting all sections but those labeled by @xmath288 . moreover , @xmath295 is a normal - crossing divisor . we ll prove it by induction on @xmath3 . for the case @xmath235 , @xmath306 and @xmath307 . then @xmath308 , where @xmath23 is the affine coordinate of @xmath309 . thus @xmath310 , and the proposition is clear in this case . assume that it s true for @xmath3 . now @xmath311 and @xmath312 . let @xmath313 , @xmath314 , and @xmath315 the divisor of singularities of the meromorphic form @xmath316 in @xmath183 . consider the morphism @xmath226 of the product of two contraction maps : @xmath317 then we obtain that @xmath318 let @xmath319 be the divisor of singularities of the form @xmath320 in @xmath321 . clearly , the divisor of singularities of the meromorphic form @xmath322 in @xmath323 is @xmath324 , where @xmath248 and @xmath325 are the projections on the first and second factors respectively . and it s a normal crossing divisor . by @xcite , @xmath226 is isomorphic to a sequence of blowups of @xmath253 along all the boundary divisors of @xmath183 . by the following lemma [ blowup ] on blowups , and proposition [ nonb1 ] and [ nonb2 ] , we can conclude that @xmath326 is a normal crossing divisor in @xmath31 . @xmath295 is contained in it . thus it s also a normal crossing divisor . on the other hand , the meromorphic form @xmath285 does not necessarily have as singularities all the exceptional divisors of @xmath226 in @xmath326 . in fact , @xmath295 equals @xmath326 minus those spurious divisors . by ( * ? ? ? * lemma 3.8 ) , the spurious divisors are those irreducible boundary divisors in @xmath31 which get blown down by @xmath226 to a subvariety of the product which is not a stratum of the divisor @xmath324 . by ( * ? ? ? * lemma 1 , page 554 ) , the exceptional divisors of @xmath226 are exactly those boundary divisors corresponding to the following @xmath41-stable partitions : @xmath327 , with @xmath328 , @xmath329 , and @xmath330 . by induction , the divisor @xmath315 consists of the following four types of irreducible divisors ( we identify the boundary divisors with the corresponding @xmath41-stable partitions here ) : @xmath331 where @xmath332 is a non - empty subset of @xmath297 , @xmath333 . first , notice the pullback of the divisor @xmath303 equals @xmath334 which is not an exceptional divisor of @xmath226 , thus it s a component of the divisor @xmath295 . for each @xmath335 , by ( * fact 3 , page 552 ) , the pullback of the boundary divisor @xmath336 in the list above is equal to the sum of the following two components : @xmath337 by the condition @xmath338 , @xmath339 is not an exceptional divisor of @xmath226 , hence it s a component of the divisor @xmath295 ; but @xmath340 is an exceptional divisor . we need to check if it s a spurious divisor . notice that @xmath341 now clearly @xmath342 is a stratum of the divisor @xmath324 if and only if @xmath300 is in the divisor @xmath319 , therefore , @xmath340 is a component of the divisor @xmath295 if and only if @xmath300 is in the divisor @xmath319 . for @xmath343 , it depends upon whether or not @xmath344 . and @xmath345 is always in the divisor @xmath319 , hence @xmath346 is a component of @xmath295 . now there are only two more components of @xmath295 . one is @xmath347 . it s not an exceptional divisor and is contained in the pullback of @xmath348 . the other one is @xmath349 , when @xmath350 . and it is contained in the pullback of @xmath319 . thus we have all the components of @xmath295 listed in the proposition . this concludes the proof . now let s prove a lemma on blowups which is used in the proof of the preceding proposition . [ blowup ] let @xmath16 be a smooth complex variety of dimension @xmath3 , @xmath1 an irreducible smooth subvariety of @xmath16 with codimension @xmath351 , and @xmath352 a normal - crossing divisor of @xmath16 with smooth irreducible components @xmath353 , @xmath354 . suppose that for each @xmath353 , only one of the following cases happens : @xmath355 , @xmath356 , or : @xmath357 is an irreducible smooth subvariety with codimension @xmath358 . let @xmath359 be the blowup of @xmath16 along @xmath1 . let @xmath360 be the strict transform of @xmath353 , @xmath361 , and @xmath362 the exceptional divisor . then the divisor @xmath363 is a normal - crossing divisor in @xmath364 . let @xmath365 . if @xmath366 , the @xmath244 is locally at @xmath367 an isomorphism . so we only need to consider the case @xmath368 . then @xmath369 . by rearranging the indices of the @xmath353 , if necessary , we may assume that @xmath355 for @xmath370 , @xmath371 for @xmath372 , and @xmath373 for @xmath374 . here @xmath375 , @xmath376 and @xmath377 . since @xmath378 is a normal - crossing divisor , there exists an open neighborhood @xmath379 of @xmath380 with local coordinates @xmath381 such that @xmath382 and @xmath383 , for @xmath370 , @xmath384 ; for @xmath372 , @xmath385 . and @xmath386 . here $ ] is the homogenous coordinate of @xmath388 . then a point @xmath389 in the inverse image of @xmath390 has the property that @xmath391 . now in @xmath392 , suppose that @xmath393 for @xmath394 , and @xmath395 for @xmath396 . then @xmath397 . now since @xmath398 , it can not happen that all @xmath391 . then there exists some @xmath399 such that @xmath400 . consider the open neighborhood @xmath401 of @xmath402 . and the restriction of the blowup @xmath244 on the @xmath403 has the following formula:@xmath404 where @xmath405 , @xmath406,@xmath407 . therefore for @xmath408 , @xmath409 and @xmath410 . thus @xmath392 is a normal - crossing divisor in @xmath364 . next we ll prove the main theorem in this section : [ thm1 ] the divisor @xmath65 of singularities @xmath64 does not contain any @xmath66-dimensional face of the algebraic stasheff polytope @xmath109 , @xmath411 . here @xmath412 with @xmath413 and @xmath414 . first observe that if @xmath65 contains some @xmath66-dimensional face of @xmath109 , then it also contains the vertices of this face . therefore it suffices to show that @xmath65 does not contain any vertex of @xmath109 . we ll prove that each irreducible component of @xmath65 listed in the preceding proposition [ divisor - a ] ca nt contain any vertex of @xmath109 . first consider the non - boundary divisor @xmath303 , for @xmath415 . clearly , its image under the contraction morphism : @xmath416 is @xmath417 . on the other hand , by the proposition [ prop4 ] in section 2 , the image of a vertex of @xmath62 under @xmath231 is @xmath4 , or @xmath59 . since @xmath415 , @xmath303 does nt contain any vertex of @xmath62 . let s consider the boundary divisor in @xmath65 . it has one of the following types @xmath418 by the proposition [ prop1 ] in section 2 , we know that a vertex of @xmath62 is contained in the intersection of exactly @xmath3 irreducible components of @xmath62 . notice that the intersection of any @xmath419 irreducible boundary divisors of @xmath31 is empty , therefore , it s enough to show that none of the boundary divisors in @xmath65 appears as an irreducible component of @xmath62 . by the proposition [ prop1 ] in section 2 , the irreducible components of @xmath62 correspond to the stable @xmath41-partitions of @xmath32 which are strictly ordered with respect to the cyclic order @xmath56 . the partition @xmath420 is not strict with respect to @xmath93 because @xmath413 then @xmath421 and @xmath422 separate any element of @xmath423 from @xmath4 ; similarly , @xmath424 is not strict with respect to @xmath93 because @xmath414 and @xmath425 and @xmath422 block any element of @xmath426 to @xmath59 ; finally @xmath427 is not strict with respect to @xmath93 because @xmath59 and @xmath4 separate @xmath422 from any element of @xmath428 . hence , none of the boundary divisors in @xmath65 contains a vertex of @xmath62 . this concludes the proof . we ll continue to use the notations in the previous sections . now consider the convergent iterated integral : @xmath429 where @xmath17 \rightarrow \mathbb{c}$ ] is a piecewise smooth simple path from @xmath4 to @xmath59 and @xmath430 , @xmath61 , @xmath431 is an open @xmath3-simplex in @xmath432 , and @xmath433 . as mentioned in the introduction , we can identify the open stratum @xmath37 with the set @xmath434 . then @xmath435 is a subset of @xmath37 , thus we have a natural map @xmath436 which embeds @xmath435 into @xmath36 . clearly we have the following equality : @xmath437 where @xmath285 is the meromorphic @xmath3-form in @xmath31 defined in section 2 . let @xmath438 be the closure of @xmath435 in @xmath31 . then the zariski closure of the boundary of @xmath438 is @xmath62 . this follows from the fact that @xmath62 is the zariski closure of the boundary of the closure of @xmath439 in @xmath31 . hence we have a relative homology class:@xmath440\in h_n(\mathcal{\overline{m}}_{0,s}(\mathbb{c}),b_n(\mathbb{c});\mathbb{q})\ ] ] the meromorphic @xmath3-form @xmath285 gives rise to a cohomology class : @xmath441\in h^{n}_{dr}(\mathcal{\overline{m}}_{0,s}(\mathbb{c})-a_s(\vec{a})(\mathbb{c } ) ; \mathbb{c}).\ ] ] now given the iterated integral ( [ ite1 ] ) , we ll define : @xmath442 then we have the following main theorem : [ mt ] @xmath443 carries an @xmath3-framed hodge - tate structure with the frames coming from @xmath444 and @xmath438 . the period is just the iterated integral @xmath445 . moreover , if all the @xmath417 are elements of a number field @xmath79 , then @xmath446 is a framed mixed tate motive over @xmath79 . before we prove the theorem , first let s briefly recall the definitions of framed hodge - tate structure and its period . a mixed hodge @xmath447-structure consists of the following data : + 1 ) a finite - dimensional @xmath447-vector space @xmath448 with a finite increasing filtration @xmath449 called the weight filtration ; + 2 ) a finite decreasing filtration @xmath450 of @xmath451 called the hodge filtration ; + these data satisfy the following conditions : + for each associated graded piece @xmath452 , one has the decomposition:@xmath453 with @xmath454 here `` '' means the complex conjugation of @xmath455 with respect to @xmath456 . the hodge numbers are the integers @xmath457 . by definition , hodge - tate structure _ @xmath458 is a mixed hodge @xmath447-structure @xmath458 with the hodge numbers @xmath459 unless @xmath460 . this means that for the weight filtration , @xmath461 and @xmath462 is a finite direct sum of @xmath463 . it also implies that , for each @xmath464 , the natural map @xmath465 is an isomorphism . an _ @xmath3-framed hodge - tate structure _ @xmath458 is a hodge - tate structure @xmath458 equipped with a nonzero vector in @xmath466 and a nonzero functional on @xmath467 , that is , we have two nonzero morphisms : @xmath468 to define the period of an @xmath3-framed hodge - tate structure , we need to choose a map of @xmath447-vector spaces @xmath469 which lifts @xmath470 , that is , @xmath471 . now let @xmath472 . consider the composition : @xmath473 where the first one is @xmath164 and the second is provided by ( [ ht ] ) . it gives rise to a vector @xmath474 . the period is the number @xmath475 . a different choice of the lifting @xmath79 will change this period by @xmath476 `` weight @xmath189 period '' . in this sense , periods are multi - valued . for more information about periods of framed mixed tate motives or periods of framed hodge - tate structures , we refer to ( * ? ? ? * chapter 5 ) and ( * ? ? ? * section 3.2 ) . now let s prove some lemmas about hodge - tate structures . [ h - t1 ] the category of hodge - tate structures is abelian . it s closed under subquotients and extensions . it s straightforward ( c.f . * thorm 2.3.5 ) ) . [ h - t3 ] let @xmath0 be a complex algebraic variety . suppose that @xmath477 where @xmath478 are closed subvarieties of @xmath0 , @xmath479 . let @xmath480 be a fixed positive integer . for any non - empty subset @xmath481 , denote by @xmath482 . suppose that all the cohomology groups of @xmath483 carry the hodge - tate structures for all @xmath484 . then @xmath485 also carries a hodge - tate structure . we ll use the mayer - vietoris exact sequence and the induction on @xmath66 . if @xmath486 , then it s obvious . assume that it s true for @xmath487 , and consider the case @xmath488 . we can write @xmath0 as : @xmath489 where @xmath490 . by the mayer - vietoris exact sequence for cohomology , we have : @xmath491 it s known that the mayer - vietoris sequence is also an exact sequence of mixed hodge structures . therefore , we get the short exact sequence of mixed hodge structures : @xmath492 notice that @xmath493 therefore , by the induction assumption , both @xmath494 and @xmath495 are hodge - tate structures . now by the lemma [ h - t1 ] , @xmath485 is a hodge - tate structure . [ h2 ] let @xmath261 be a complex algebraic variety , and @xmath0 be a subvariety of codimension @xmath59 . then if both @xmath496 and @xmath497 are hodge - tate , so are the relative cohomology @xmath498 and @xmath499 . consider the long exact sequence of cohomologies for the pair @xmath500 and we know that it s also a long exact sequence of mixed hodge structures . similar to lemma [ h - t3 ] , by the lemma [ h - t1 ] , we can conclude that the relative cohomology @xmath498 carries a hodge - tate structure , then by duality , so does @xmath499 . now let s prove the theorem : proof of theorem [ mt ] . by deligne @xcite , there is a canonical mixed hodge structure on the relative cohomology @xmath501 . first we ll show that @xmath502 is a hodge - tate structure . by the proposition [ divisor - a ] and corollary [ nonb4 ] , we know that the intersections of the components of @xmath65 are either empty or tate varieties . by lemma [ h - t3 ] and [ h2 ] , @xmath503 carries a hodge - tate structure . for the same reason , @xmath504 also has a hodge - tate structure . by the lemma [ h2 ] again , @xmath501 carries a hodge - tate structure . now we need to show that @xmath444 and @xmath438 give the frames . first , there are natural non - zero morphisms of pure tate structures by @xmath444 and @xmath438 : @xmath505 : \mathbb{q}(-n)\rightarrow gr^w_{2n}h^{n } ( \mathcal{\overline{m}}_{0,s}-a_s({\vec{a}})),\ ] ] @xmath440:\mathbb{q}(0)\rightarrow gr_0^{w}h_n ( \mathcal{\overline{m}}_{0,s},b_n)\ ] ] then composing with the canonical isomorphisms : @xmath506 and @xmath507 we get the frame morphisms : @xmath508^{\prime } : \mathbb{q}(-n)\rightarrow gr^w_{2n } h^{n}(\mathcal{\overline{m}}_{0,s}-a_s({\vec{a } } ) , b_n - b_n\cap a_s({\vec{a}}))\ ] ] @xmath440^{\prime}:\mathbb{q}(0)\rightarrow gr_0^{w}h_{n } ( \mathcal{\overline{m}}_{0,s}-a_s({\vec{a } } ) , b_n - b_n\cap a_s({\vec{a}})).\ ] ] since @xmath430 , for @xmath61 , by the description of @xmath509 and theorem [ thm1 ] in section 3 , we have @xmath510 therefore , @xmath438 provides an element of the relative betti homology @xmath440\in h_{n}(\mathcal{\overline{m}}_{0,s}-a_s({\vec{a } } ) , b_n - b_n\cap a_s({\vec{a}}))\ ] ] which is a lift of the frame morphism @xmath511^{\prime}$ ] . the form @xmath444 obviously gives an element of the relative de rham cohomology class @xmath508\in h^{n}(\mathcal{\overline{m}}_{0,s}-a_s({\vec{a } } ) , b_n - b_n\cap a_s({\vec{a}})).\ ] ] and their pairing is exactly the iterated integral in the theorem . now suppose that @xmath417 , @xmath512 are elements of a number field @xmath79 . we ll show that @xmath513 is a framed mixed tate motive over @xmath79 . for this , we ll follow the proof of ( * ? ? ? * theorem 4.1 ) very closely . in @xcite , goncharov constructed the abelian category of mixed tate motive over any number field @xmath79 , where he used the theory of triangulated category of mixed motives by voevodsky in @xcite . let s apply it to our case . first , consider the standard cosimplicial variety : @xmath514 here @xmath515 , and @xmath516 is the disjoint union of the codimension @xmath66 strata of the divisor @xmath517 . according to the standard procedure , we can get a complex @xmath518 of varieties from the above cosimplicial variety with @xmath519 at the degree @xmath4 . then it gives an object in voevodsky s triangulated category of mixed motives over @xmath79 . in fact it also belongs to the triangulated subcategory @xmath520 of mixed tate motives over @xmath79 . and there is a canonical @xmath23-structure @xmath23 on @xmath520 . then @xmath521 is our mixed tate motive . similarly , by using the construction above for the hodge - tate structures and the fact that the hodge realization is a fully faithful functor on the category of pure tate motives , we obtain the frame morphisms for our motive coming from @xmath522^{\prime}$ ] and @xmath511^{\prime}$ ] defined above . the theorem is proved . in this section , we ll apply our preceding construction to the dilogarithm and show that the corresponding period matrix is the same as the one given by p. deligne . hence they are isomorphic . recall that @xmath82 now let s consider the moduli space @xmath523 . it s known that @xmath523 is the blow - up of @xmath524 at three points @xmath525 on the diagonal . let @xmath526 denote this blow - up . and let @xmath527 be the affine coordinates on @xmath524 . we list the ten boundary divisors of @xmath523 as follows ( see figure 1 ) . moreover we can describe the divisors @xmath538 and @xmath539 explicitly . by definition , @xmath538 is the divisor of singularities of the pullback meromorphic @xmath41-form @xmath540 of @xmath523 . we can check the following lemma directly and leave it to the reader . first we state the results . in the following statements , we only list the ( relative ) betti homology groups because by the comparison theorem the derham cohomology groups are isomorphic to the corresponding betti cohomology groups tensor with @xmath548 . [ m5.2 ] ( a ) if @xmath549 then : @xmath554 furthermore , there are natural bases such that the corresponding period matrix is @xmath85.\ ] ] hence , it coincides with the one given by deligne and they are isomorphic as hodge - tate structures . now we ll prove these two theorems . before we do it , let s give a more explicit description of the affine variety @xmath556 . we know @xmath557 is the blow - up of @xmath558 at the points @xmath559 , @xmath560 , @xmath561 , hence @xmath562 in which @xmath563 denotes the blow - up of @xmath564 at the points @xmath565 ( see figure 1 ) . we ll use the same letter @xmath566 for this blow - up . proof of theorem [ m5.1 ] : the idea is to decompose @xmath556 into the union of two subspaces , then use the mayer - vietoris exact sequences . + * case ( a ) : @xmath549 . * in this case we have : @xmath567 where @xmath568 denotes the unique intersection point of @xmath569 and the proper transform of @xmath570 . let @xmath571 , then it s open and isomorphic to @xmath572 which denotes the blow - up of @xmath573 at the point @xmath574 . let @xmath575 in @xmath564 . define @xmath576 , then it s clear that @xmath577 is the deformation retract of @xmath578 by the line contraction : @xmath579 . so we have the decomposition : @xmath580 . we claim that @xmath581 deformation retracts to @xmath582 , where @xmath583 is the @xmath66-sphere of radius @xmath584 . indeed , notice that @xmath585 , hence it is isomorphic to @xmath586 which deformation retracts to @xmath582 under the map @xmath587 , where @xmath588 and @xmath589 . hence , @xmath590 . by the alexander duality , we have @xmath591 the natural generator of @xmath592 is a simple loop around the divisor @xmath593 . since @xmath594 deformation retracts to @xmath595 , we have @xmath596 and @xmath597 . now @xmath598 , so it follows from ( * ? ? ? * page 473 - 474 ) that : @xmath599 now apply mayer - vietoris exact sequence , we obtain : + @xmath600 ; + @xmath601 @xmath602 + we also see that the generator for @xmath603 is a small simple loop around @xmath543 . * case ( b):@xmath86. we see in this case : @xmath604 , where @xmath568 is the intersection point of @xmath569 and @xmath593 ( the proper transform of @xmath570 ) ; @xmath605 is the intersection point of @xmath606 and @xmath607 ( the proper transform of @xmath608 ) . let @xmath609 . it s open and isomorphic to @xmath610 . as we did in case ( a ) , we can construct a small open neighborhood @xmath611 of @xmath577 and @xmath612 of @xmath613 such that : ( 1 ) @xmath611 deformation retracts to @xmath577 , and @xmath612 deformation retracts to @xmath614 ; ( 2 ) @xmath615 ; ( 3 ) @xmath616 deformation retract to @xmath617 for some small @xmath584 . we define @xmath618 . then we have the decomposition : @xmath619 . and @xmath620 deformation retracts to the disjoint union of two copies of @xmath582 . now we can proceed as in case ( a ) and find @xmath621 , @xmath622 , other homology groups vanish . theorem [ m5.1 ] is proved . proof of theorem [ m5.2 ] : now we ll use the results of theorem [ m5.1 ] to compute the relative homology and cohomology groups . + case ( a):@xmath549 . * first , let s look at the divisor @xmath623 ( see figure 2 ) . it consists of five irreducible components . let s label them as follows ( see the figure 1 and figure 2 ) : + @xmath624 , @xmath625 , @xmath626 , @xmath627 , @xmath628 + we have five intersection points of these components : + @xmath629 let @xmath630 be the disjoint union of the irreducible components @xmath631 , and @xmath632 be the disjoint union of the points @xmath633 ; @xmath634 . then we can compute @xmath635 by the following bicomplex @xmath636 : @xmath637 where @xmath638 denote the vector space of singular @xmath151-chain on @xmath261 with coefficients in @xmath447 . the vertical differential @xmath639 is the differential of the singular chain complex of @xmath261 and the horizontal differential @xmath584 is defined as following : @xmath640 where each @xmath641 is a @xmath642-chain of @xmath643 , @xmath644 . and @xmath645 where for each pair @xmath646 , @xmath647 is a @xmath642-chain of @xmath648 and @xmath649 represents the image of @xmath647 in @xmath643 under the incusion @xmath650 . that is , we have the following diagram : @xmath651 consider the spectral sequence with @xmath652 terms given as : @xmath653 the differential @xmath654 is induced by the horizontal differential @xmath584 . use the explicit generators of the homology groups in the theorem [ m5.1 ] , we immediately get the @xmath655 terms and @xmath656 : @xmath657 hence the spectral sequence degenerates at the @xmath658 terms . so , @xmath659 and the filtration induced by this spectral sequence coincides with the weight filtration . now let s turn to the relative de rham cohomology group @xmath660 . it is defined as the @xmath151-th cohomology of the total complex of the following bicomplex @xmath661 : @xmath662 where @xmath663 denote the vector space of @xmath664 complex @xmath151-forms on @xmath261 , the vertical differential @xmath665 is the exterior differentiation of forms and @xmath584 is defined as follows:@xmath666 where @xmath667 is a @xmath642-form of @xmath556 and @xmath668 is the restriction of @xmath667 on @xmath643 . @xmath669 where @xmath670 is a @xmath642-form of @xmath195 and for each pair @xmath646 , @xmath671 is the restriction of @xmath672 on @xmath648 . we have the following diagram : @xmath673 similar to the case for homology , we see that the corresponding spectral sequence degenerates at the @xmath658 terms . so we obtain : @xmath674 , and other relative cohomology groups vanish . for the case @xmath86 , the proof is similar and we leave it to the reader . we have the following three cochains in the total complex associated to the derham bicomplex @xmath661 : @xmath675 where @xmath676 is a function on @xmath677 satisfying @xmath678 and @xmath679 is a @xmath59-form on the component @xmath680 of @xmath195 . since @xmath679 and @xmath681 are holomorphic on @xmath680 and @xmath550 respectively , they are closed forms . hence @xmath682 are cocycles and represent elements in @xmath683 . + next , let s consider the cycles in the betti bicomplex @xmath684 . + let @xmath685 , where @xmath686 denotes the boundary of @xmath687 in @xmath195 . it s just the @xmath81 sides of the dotted pentagon in figure @xmath41 . let @xmath688 , where @xmath689 , and @xmath690 is a small positive number , @xmath691 . it s clear that the boundary @xmath692 consists of @xmath41 small cycles @xmath693 and @xmath694 . since @xmath695 does not contain the blow - up points , @xmath696 is isomorphic to @xmath695 and the boundary @xmath697 . the key point is that @xmath698 is contained in @xmath630 , which means exactly that @xmath699 is a cycle in the total complex . finally , let @xmath700 , where @xmath701 is the inverse image of a torus and since the three blow - up points are not on the torus , @xmath94 is isomorphic to its image . clearly the boundary of @xmath94 is zero . hence @xmath702 is a cycle . now we can calculate the period matrix @xmath703 between @xmath704 and @xmath705 , here @xmath706 , @xmath707 . it s straightforward that @xmath708 , @xmath709 + @xmath710 , and @xmath711 , @xmath712 . since @xmath713 is equal to the integral of the @xmath59-form @xmath714 over a small circle with center @xmath543 , we have @xmath715 . and @xmath716 therefore , we get the period matrix : @xmath717.\ ] ] since it s nonsingular , it follows that @xmath718 and @xmath719 are bases . for part(b ) , it s clear that : for the cocycles , we have @xmath720 ; for the cycles , we have @xmath721 . therefore , its period matrix is + @xmath722.\ ] ] a. a. beilinson , a. b. goncharov , v. v. schechtman and a. n. varchenko , _ aomoto dilogarithms , mixed hodge structure and motivic cohomology of pairs of triangle on the plane _ , in the grothendieck festschrift , i , prog . in math . 86 , birkhuser , boston , 1990 . p. deligne , _ theorie de hodge , i _ , actes cong . 1970 , t. 1 . 425 - 430 . p. deligne , _ theorie de hodge , ii _ , publ . 40,(1971),5 - 58 . deligne , _ theorie de hodge , iii _ , inst . hautes . 44,(1974),5 - 77 . p. deligne , _ resum des premiers exposs de a.grothendieck_ , sga 7 , lnm 288 , 1 - 24 . p. deligne , ; d. mumford , _ the irreduciblity of the space of curves of given genus _ , publ . math . inst . hautes . 36,(1969),75 - 109 . p. deligne and a. b. goncharov , _ groupes fondamentaux motiviques de tate mixte _ , ann . , @xmath723 srie , t. 38 , 2005 , 1 - 56 . l. euler , _ opera omnia _ , ser.1 , vol xv , teubner , berlin 1917 , 217 - 267 . p. griffiths and j. harris , _ principles of algebraic geometry _ , wiley and sons , 1978 . a. b. goncharov and yu . i. manin , _ mutiple @xmath6-motives and moduli spaces @xmath2 _ , compositio math . 140 ( 2004 ) , 1 - 14 . a. b. goncharov , _ polylogarithms in arithmetic and geometry _ , proc . of icm . vol . 1 , 2 ( zurich , 1994 ) , 374 - 387 , birkhauser , basel , 1995 . a. b. goncharov , _ polylogarithms and motivic galois groups _ , sympos . pure math . motives , vol . 55 , part 2 , ( 1994 ) , 43 - 97 . a. b. goncharov , _ volume of hyperbolic manifolds and mixed tate motives _ , j. amer . 12 ( 1999 ) , n2 , 569 - 618 . a. b. goncharov , _ multiple polylogarithms and mixed tate motives _ , math.ag/0103059 . a. b. goncharov , _ periods and mixed motives_. math.ag/0202154 . r. hain , _ classical polylogarithms _ , motives , sympos . pure math . vol . 55 , part 2 , ( 1994 ) , 1 - 42 . r. hain and r. macpherson , _ higher logarithms _ , ill . 34 , ( 1990 ) . c. gerhardt ( ed ) and g. leibniz , _ mathematische schriften iii/1 _ , georg olms verlag , hildescheim and new york , 1971 , pp . 336 - 339 . l. lewin , _ dilogarithms and associated functions _ , north - holland , 1981 . s. keel , _ intersection theory of moduli space of stable n - pointed curves of genus zero _ , trans.ams , 330 , n2 ( 1992 ) , 545 - 574 . m. kontsevich and yu . i. manin , _ quantum cohomology of a product _ , invent . math.,124 , 313 - 339 ( 1996 ) . f. f. knudsen , _ the projectivity of the moduli space of stable curves ii the stacks @xmath724 _ , math . scand . 52 ( 1983 ) , 163 - 199 i. manin , _ correspondencs , motifs and monoidal transformations _ , math.ussr-sbornik 6 ( 1968 ) , 439 - 470 . i. manin , _ frobenius manifolds , quantum cohomology , and modulis spaces _ , vol . 47 , colloquim publications , ams , 1999 . v. voevodsky , _ triangulated category of motives over a field _ , in : cycles , transfers , and motivic homology theories , 188 - 238 , ann . of math . , 143 , princeton univ . press , nj , 2000 . q. wang , _ multiple polylogarithm motives and moduli spaces @xmath2 _ , ph.d . dissertation , brown university , may 2004 . q. wang , _ a geometric construction of motivic classical polylogarithm _ , in preparation .	in this paper , we give a natural construction of mixed tate motives whose periods are a class of iterated integrals which include the multiple polylogarithm functions . given such an iterated integral , we construct two divisors @xmath0 and @xmath1 in the moduli spaces @xmath2 of @xmath3-pointed stable curves of genus @xmath4 , and prove that the cohomology of the pair @xmath5 is a framed mixed tate motive whose period is that integral . it generalizes the results of a. goncharov and yu . manin for multiple @xmath6-values . then we apply our construction to the dilogarithm and calculate the period matrix which turns out to be same with the canonical one of deligne .
predictive coding is a brain plausible principle to account for how diversity of perceptual input sequences can be predicted or generated for different intentions in the top - down pathway as well as how the corresponding intention can be inferred or recognized for a particular observation of perceptual sequence in the bottom - up pathway by using the prediction error minimization principle . within this predictive coding framework , it has been largely assumed that such prediction / generation and inference / recognition can be conducted through multiple levels across different cortical areas whereas necessary functional hierarchy can be developed via accumulated learning [ 1 , 2 , 3 , 4 ] . the current study examines how a spatio - temporal hierarchy that is adequate for robust generation and recognition of compositional dynamic visual patterns in the pixel level can be developed by proposing a novel deterministic predictive coding type deep recurrent neural network model . the proposed model , referred to as the predictive multiple spatio - temporal scales rnn ( p - mstrnn ) is a combination of prior - proposed models of the multiple timescales rnn ( mtrnn ) [ 4 ] and the deconvolutional neural network model [ 15 ] where both spatial and temporal multiple scales properties are used as macroscopic constraints to develop an effective functional hierarchy . there have been prior studies on predicting and generating dynamic visual datasets [ 6 , 7 , 12 , 13 ] . in particular , similar to the current paper , lotter et al . in 2016 [ 13 ] also adopted predictive coding in their work . however , p - pmstrnn differs from the previously proposed generative models for dynamic vision processing , because in the current model not only prediction in the top - down pathway but also inference in the bottom - up pathway can be performed . also , p - mstrnn is advantageous because it can deal with temporal hierarchy in addition to the spatial hierarchy by using the multiple temporal scales properties in the model . the current study is assumed to be the first attempt to achieve both generation / prediction and recognition / inference of dynamic visual images within one neural network model . the model was tested with using a video dataset consisting of multiple types of human movement patterns generated by multiple subjects . after the model was trained for multiple movement patterns , we examined how those multiple patterns are regenerated by using a functional hierarchy self - organized during the learning . we also scrutinized how concatenation of the learned movement patterns is developed further via additional learning . furthermore , we examined how test movement patterns are recognized by inferring the corresponding intention . this recognition test was conducted by following a test framework , referred to as imitative synchronization [ 11 ] , where p - mstrnn attempts to imitate perceived test movement patterns . we closely analyzed how the model network recognizes possible shifts in test movement patterns by inferring the corresponding shifts of intentions using the on - line error regression scheme . the video dataset used in training and testing consists of multiple cyclic human movement patterns performed by multiple subjects . a set of simulation experiments examined the spatio - temporal hierarchy as it develops during learning , and revealed that robust recognition of test patterns depends on learning spatial - temporal interdependencies with exemplar patterns exhibiting high degrees of variance . the newly proposed model , the predictive multiple spatio - temporal scales rnn ( p - mstrnn ) is a hierarchical neural network model that utilizes spatio - temporal features for generating and recognizing dynamic visual images . the p - mstrnn model is based on the multiple spatio - temoral scales rnn ( mstrnn ) [ 8 ] . the original mstrnn is a classification model for dynamic visual images without prediction mechanism . p - mstrnn differs from mstrnn because it learns , generates , and recognizes patterns using the principle of prediction error minimization in the predictive coding framework . p - mstrnn consists of a series of context layers and one input / output layer in the bottom , as shown in figure 1 . each context layer predicts its own neural state at each next time step by receiving the top - down and bottom - up signals from the neighboring layers . the neural state of each context layer represents its top - down intention . the first context layer predicts the next step visual inputs by receiving both the current visual input and the top - down signal from the second layer . in the upward direction , the prediction error signal back - propagates inversely in the same pathway for the purpose of updating the connectivity weights as well as the intention at each context layer in the learning and recognition process . the model utilizes a convolutional neural network ( cnn ) and a deconvolutional neural network [ 9 , 15 ] . however , unlike the cnn , every context layer in the p - mstrnn model is composed of two different types of units that function differently from the units composing cnn layers . the first type is the feature unit and the other is the context unit . both feature unit and context unit are based on the leaky integrator neural units which decay to the value of the previous time step , thereby enabling both types of units to reflect a temporal hierarchy that is not possible in a cnn [ 10 ] . the set of feature units in the same layer forms a feature map ( fm ) and context units form a context map ( cm ) . fms mainly contribute to spatial processing by receiving synaptic inputs from neighboring feature units . the cms , on the other hand , contribute mainly to development of adequate dynamic information processing by taking advantage of their adjustable recurrent connectivity . the timescale that determines decay rate is differently set in all layers . the layers in the lower level have smaller time constants and the higher level layers have larger time constants . therefore , the neural activity in the lower context layer is constrained to be faster while the one in the higher layers is constrained to be slower . the forward dynamics of fms for the generalized layer are shown in equation ( 1 ) and ( 2 ) . @xmath0 @xmath1 where @xmath2 is the @xmath3 fm s internal state in the @xmath4 layer at time step t , @xmath5 is the internal state of @xmath3 fm in the @xmath4 layer at t-1 time steps . @xmath6 is the timescale in the @xmath4 layer and @xmath7 is the number of fms in the @xmath8 layer . the operation @xmath9 is the convolution operator . @xmath10 is the kernel connecting the @xmath11 fm in the @xmath8 layer at t-1 time steps and @xmath3 fm in the @xmath4 layer at time step t. @xmath12 is the total number of cm in the @xmath4 layer and @xmath13 is the activation of the @xmath14 cm in the @xmath4 layer at time step t-1 . @xmath15 is the kernel connecting the @xmath14 cm in @xmath4 layer at the previous time step t-1 and the @xmath3 fm in the @xmath4 layer at the current time step t. @xmath16 is the input data frame from outside of the network and @xmath17 is the kernel between input and the current fm . @xmath18 is the bias of @xmath3 fms in the @xmath4 layer . the first term of equation ( 1 ) represents the decayed internal states of fms from the previous time step t-1 with the decay rate @xmath19 . the second term is the input to the current fm from the activation value of fms in the @xmath8 layer . the third term represents the input from the cms in the same layer at the previous time step t-1 . the fourth term indicates the data frame fed to the current fm . after the internal state of the current fm is computed , the activation value is obtained through a hyperbolic tangent activation function ( equation ( 2 ) ) . when the layer which the current @xmath3 fm belongs to is other than the first layer , the term from the input data frame @xmath16 no longer exists . also , highest layer internal dynamics do not include the second term in equation 1 due to the absence of any higher layer to provide values . equations ( 3 ) and ( 4 ) detail the forward dynamics of cms . @xmath20 @xmath21 where @xmath22 is the internal state of the @xmath23 cm in the @xmath4 layer at time step t and @xmath13 is the activation value of the @xmath14 cm in the @xmath4 layer at time step t-1 . @xmath24 is the recurrent weight connecting the @xmath23 cm in time step t with the previous @xmath14 cm in time step t-1 . @xmath25 is the weight connecting the current @xmath23 cm with the @xmath11 fm in the @xmath8 layer at time step t-1 . @xmath26 is kernel connecting the @xmath27 fm in @xmath28 layer in t-1 step to current cm . @xmath29 is the element - wise multiplication operator . the first term of equation ( 3 ) is the leaky integrator input from the internal state of the cm from the previous time step t-1 . the second term is the recurrent input to the current cm from the previous time step s cm activation value . @xmath30 is same size as cms in the same layer . accordingly , the activation values of the previous cm are multiplied with @xmath31 in an element - wise manner . in the cms , this second term represents the recurrency and reinforces the temporal processes of the network . the third term represents the input from the activation value of fms in the @xmath8 layer . the fourth term is from the lower layer fms . this term utilizes the bottom - up pathway allowing input data frames from outside of the network to be processed through all layers of the network . for the internal state calculation of the cms , if the layer is the highest layer , then there is no input from the upper l+1 layer and the third term in equation ( 3 ) is unused . likewise , if the layer is the lowest layer , then an input from the lower layer does not exist and the fourth term in equation ( 3 ) is unused . after the internal state is calculated , an activation value is obtained using the hyperbolic tangent function in equation ( 4 ) . when calculating the convolution , there are some cases in which the input map size is smaller than the size of the output map . in these cases , zero - padding is used for the input maps . as for the element - wise multiplication , the map and weight sizes are always the same . in the output layer , the output is calculated as , @xmath32 @xmath33 where @xmath34 is the internal state of the output layer , @xmath35 is the number of fms in the @xmath36 layer . @xmath37 is the kernel connecting the @xmath11 fm in the @xmath36 layer with the output map . @xmath38 is the bias for the output map . the internal state of the output is calculated from the fms in the first layer by convolution , and the activation value of the output layer is obtained by applying the scaled hyperbolic tangent function as suggested by lecun et al . [ 16 ] . the open loop generation method was used for training the network . when the network utilizes the open loop generation , the network receives the current input frame from the dataset and generates a ( single or multiple ) step prediction as an output frame for each step . we adopted mean square error ( mse ) for the cost function . a conventional back - propagation through time ( bptt ) method was used for training the network . the weights , kernels , biases and initial states of the context layers were optimized using gradient descent . along with open loop generation , closed loop generation was also used for training the network . the closed loop generation method is a scheme in which next step output prediction is computed by using a copy of the output prediction from the previous step as the current input . in the closed loop method , error is also calculated in comparison to the target signals . during training , error from open loop generation was used to optimize network parameters . as the training proceeded , both open loop error and closed loop error decreased . however , closed loop error was always higher than that of open loop because closed loop prediction generates accumulated error over time steps without correction from factors outside of the network . therefore , network training was terminated when the closed loop error reached a predefined lower bound threshold . at this point , the network was guaranteed to successfully perform both open loop generation and closed loop generation . in order for the network to learn to generate multiple data sequences , it must infer optimal initial states of context units in all layers for each sequence , as well as optimal connectivity weights . inferred initial states for each training sequence represent intentions to generate corresponding sequence . is the prediction of the next step input and @xmath39 is the target . ] the novelty of this model is that the trained network can be used for both generation and recognition of sequence patterns . once the network is trained , intended patterns are generated through the top - down pathway by setting the intention states of context units . the intention then propagates to the lower layers , eventually producing the intended or predicted image in a pixel level . on the other hand , current input patterns can be recognized by inferring optimal intention states of context units . the process , imitative synchronization via error regression which will be explained later , is performed by first going through the bottom - up pathway for recognition to find the optimal internal states . optimal prediction and generation are then achieved through the top - down pathway . in the framework of neuroscience , the brain receives sensory input by a bottom - up pathway and predicts the next input through a top - down path at every moment . when the actual input is different from what the brain predicted , the brain is surprised and modifies its prediction by reflecting the discrepancies . in this way , the brain always accurately predicts the future and actively interacts with the outer world [ 1 , 14 ] . the idea of imitative synchronization via error regression is the same as the way the brain recognizes and predicts sensory inputs . suppose the network is already trained to generate several types of human movement patterns in a pixel level video sequence given corresponding initial states . the network now attempts to actively imitate incoming streams of human movement patterns from an online camera at every moment by predicting future frames . this task is called imitative synchronization . to achieve imitative synchronization , the network must predict the next input pattern based on the recognized intention of the current input . the recognition process is performed by finding the network s optimal internal states that minimize error between predictions and actual inputs . this error minimizing method is called error regression . during the test , if the human suddenly changes current movement pattern to another type of pattern , the network must quickly read the intention of the changed pattern and modify its generation in a pixel level . this active recognition and generation task of imitative synchronization can be achieved by error regression . through the error regression , for recognizing the input at the current time step @xmath40 , thereby producing optimal prediction for future steps , the network utilizes past information . the network adapts its internal states by trying to imitate input history that is already known to the network . when the network produces optimal reconstruction outputs for past inputs , future prediction is also made based on the same intention the network has . this process , error regression , occurs only within the temporal window from time step @xmath41 to @xmath40 . the network first produces closed loop prediction starting from the time step @xmath41 to @xmath40 and error is calculated by comparing the actual input history and the output of the network in the same range of time . the error is then back propagated through time ( bptt ) until it reaches the first step of window at @xmath41 , which is called intention states . this scheme is shown in figure 2 in detail . in the figure , the prediction error is back propagated through the path presented as red arrows and finally reaches initial states inside the temporal window at time step @xmath41 . the initial states are then modified to produce minimum reconstruction error between closed loop generation and the actual input history within the window while maintaining weights and biases unchanged . until the prediction error decreases to produce an output pattern close to the input history , forward and backward propagation and modification of intention states are iterated multiple times inside the temporal window . when the error is low enough that the network successfully imitates input history , prediction for future frame @xmath42 is computed based on modified internal states . after that , the temporal window shifts one step forward in order to cover next time steps . in this manner , the test input pattern can be recognized in an on - line manner step by step and therefore allowing active prediction as close as possible to the input pattern . an experimental dataset was designed to reveal the spatio - temporal structure of the proposed p - msrtnn . the dataset consists of whole body human movement patterns generated by following a hierarchically defined movement syntax . sub - primitives ( arms and legs ) were defined first . each whole body movement primitive was composed of these sub - primitives , with sub - primitives being shared by all movements . experiment 1 analyzed the self - organization of the spatio - temporal hierarchy as the model learned one subject s six primitive movements . after training these primitives , the network was trained with additional ( previously unlearned ) concatenations of prior - learned primitive patterns . experiment 2 examined the recognition capability of the trained model . specifically , how the robustness of recognition is dependent on variance in training exemplar patterns was tested by changing the number of subjects preparing training patterns . the dataset for this experiment consists of six whole body movement patterns . each whole body movement pattern was hierarchically generated by combining predefined sub - primitives using legs and arms . figure 3 describes the sub - primitives . there are three types of arm sub - primitives . sub - primitive 1 ( a1 ) is laterally extending arms . sub - primitive 2 ( a2 ) is vertically extending arms . sub - primitive 3 ( a3 ) is drawing a large circle with arms . in action space , these arm sub - primitives are represented as a1r ( arm sub primitive 1 , right ) , a1l(arm sub primitive 1 , left ) , a2r , a2l , a3r and a3l . leg sub - primitives appear in the second row of figure 3 . there are three types of leg sub - primitives . the first leg sub - primitive ( l1 ) is raising the right and left leg alternatively . the second leg sub - primitive ( l2 ) is standing still , moving neither leg . the third leg sub - primitive ( l3 ) is bending both legs . all are shown in figure 3 . there are a total of six whole body action primitives . their syntax is presented in table 1 , [ cols="^,^,^,^,^,^,^,^,^,^,^,^,^ " , ] & 0.0391 & 0.0522 + table 3 shows the network trained under the multi - subject training condition yielded better results in imitative synchronization than the network trained under the single - subject training condition . this suggests that increased variance in training results in more robust recognition ability . this result is analogous to that of [ 11 ] which used lower dimensional data compared to the present study . under the multi - subject condition , identical movement primitives differed slightly from one another , and this variance was well preserved in the closed - loop output generation of dynamic images . figure 9 shows the corresponding analysis of neural activity in different layers in the network for five subjects . ( a ) shows neural activations of cms in the first layer and ( b ) shows those of cms in the fourth layer . in the lower layer ( figure 9.a ) , the same primitives formed clusters . at the same time , variations of neural activation patterns corresponding to the different subjects in the same cluster are well preserved ( note the different shapes and the positions of the trajectories ) . as in the analysis of experiment 1 , higher layer intentions represented in terms of different fixed points in ( b ) determine the corresponding cyclic activation patterns in the lower level shown in ( a ) by means of parameter bifurcation . in training with multiple subjects as shown in figure 9.b , three clusters emerge ( one for each primitive ) , and fixed points vary within these clusters ( reflecting subject variations ) . this implies that the higher level influences the lower level through the differentiation of fixed points within clusters of more or less robustly established primitives . the current paper introduced a novel dynamic neural network model for both generating and recognizing dynamic visual image patterns at the pixel level within a predictive coding framework . p - msrtnn is distinguished from other generative models in that generation and recognition are possible in one model . our simulation experiments showed that the proposed model characterized by its multiple spatio - temporal scales property can learn to generate dynamic visual images representing human movement patterns through the development of an internal spatio - temporal hierarchy . the model can also robustly recognize test movement patterns containing multiple transitions performed by unfamiliar subjects through imitative synchronization by inferring intention states with the error regression scheme applied to context units . future studies should focus on scaling the model in terms of size of frames , as well as in terms of number and complexity of movement patterns . also , applying the proposed model to the robotics domain is considered . since the model utilizes temporal hierarchy , top - down and bottom - up pathway which are also used in the human brain , this model would be beneficial to robots to act more natural and human - like way . moreover , robots with error regression scheme , which can be regarded as memory retrieving process , would be interesting . this work was supported by a national research foundation of korea ( nrf ) grant funded by the korea government ( msip ) ( no.2014r1a2a2a01005491 ) . we thank ahmadreza ahmadi for his help with the cuda program . [ 1 ] rao , r.p . and ballard , d.h . , 1999 . predictive coding in the visual cortex : a functional interpretation of some extra - classical receptive - field effects . _ nature neuroscience _ , * 2*(1 ) , pp.79 - 87 . [ 8 ] lee , h. , jung , m. and tani , j. , 2016 . characteristics of visual categorization of long - concatenated and object - directed human actions by a multiple spatio - temporal scales recurrent neural network model . _ arxiv preprint _ arxiv:1602.01921 . [ 11 ] ahamdi , a. and tani , j. , 2016 , october . towards robustness to fluctuated perceptual patterns by a deterministic predictive coding model in a task of imitative synchronization with human movement patterns . in international conference on neural information processing ( pp . 393 - 402 ) . springer international publishing . [ 12 ] xingjian , s.h.i . , chen , z. , wang , h. , yeung , d.y . , wong , w.k . and woo , w.c . convolutional lstm network : a machine learning approach for precipitation nowcasting . in advances in neural information processing systems ( pp . 802 - 810 ) . [ 15 ] zeiler , m.d . , taylor , g.w . and fergus , r. , 2011 , november . adaptive deconvolutional networks for mid and high level feature learning . in 2011 international conference on computer vision ( pp . 2018 - 2025 ) . ieee .	the current paper presents a novel recurrent neural network model , predictive multiple spatio - temporal scales rnn ( p - mstrnn ) , which can generate as well as recognize dynamic visual patterns in a predictive coding framework . the model is characterized by multiple spatio - temporal scales imposed on neural unit dynamics through which an adequate spatio - temporal hierarchy develops via learning from exemplars . the model was evaluated by conducting an experiment of learning a set of whole body human movement patterns , which was generated by following a hierarchically defined movement syntax . the analysis of the trained model clarifies what types of spatio - temporal hierarchy develops in dynamic neural activity as well as how robust generation and recognition of movement patterns can be achieved by using the error minimization principle .
complex networks@xcite have been recognized to be powerful representations of complex systems and can advance our understanding of their dynamics . they are studied in diverse disciplines ranging from earth@xcite and climate science@xcite to the neurosciences@xcite . in this perspective , systems are considered to be composed of subsystems ( i.e. , nodes ) which can or can not interact with each other according to some underlying physical coupling topology ( represented by links between nodes ) . while such structural networks have been studied extensively and are considered to serve as the physical substrate on which dynamical patterns can emerge , it is only recently that research into network dynamics gained strong momentum@xcite . the system dynamics may be represented by an interaction ( or functional ) network in which nodes represent subsystems and links reflect interactions between them . in field studies , such networks are usually derived via time series analysis techniques where nodes are associated with sensors and links are derived from the strength and/or direction of interactions as assessed by estimators of signal interdependence@xcite . using concepts from network theory , a plethora of methods have been developed to probe and assess diverse global and local properties of networks such as clustering coefficient , average shortest path length , synchronizability , or the degree distribution . moreover , numerous studies demonstrated that network topologies can decisively depend on the tendency of links to connect nodes with certain local properties@xcite . for instance , if people of a group tend to make friends with people speaking the same language , then this group is likely to separate into distinct circles of friends according to languages@xcite . this tendency of links to connect nodes with similar or equal properties is called _ assortativity _ and was frequently studied with respect to node degrees ( i.e. , the number of links connected to nodes ) . assortativity according to node degrees can be quantified by the assortativity coefficient@xcite and is considered to reflect the extent of degree - degree correlations present in the network topology . if links preferentially connect nodes of similar ( dissimilar ) degree , such networks are called assortative ( disassortative ) . the presence of degree - degree correlations was found to have far - reaching consequences for network resilience ( disassortative networks are more vulnerable to attacks than assortative networks)@xcite , for the ability of a network to globally synchronize ( disassortative networks appear to be easier to synchronize than assortative ones)@xcite , and the tendency of a network to separate into distinct groups ( assortative networks show a stronger tendency to disintegrate into different groups than disassortative networks)@xcite . moreover , in assortative networks , hubs ( nodes with relatively high degree ) tend to be closely interconnected thereby forming a resilient core which may facilitate the spread of information over the network . many technological and biological networks appear to be disassortative@xcite , while social@xcite , seismic@xcite , and functional brain networks derived from measurements of neural activity ( e.g. , electroencephalogram , magnetoencephalogram , or functional magnetic resonance imaging data)@xcite were reported to be assortative . structural brain networks may exhibit assortativity@xcite or disassortativity@xcite depending on the investigated scale@xcite , and other factors had recently been identified that might affect assortativity@xcite . studies that investigated the impact of neurological or neurodegenerative diseases on structural or functional brain networks reported on increased assortativity in patients suffering from alzheimer s disease@xcite , dementia@xcite or schizophrenia@xcite , in patients with brain tumors@xcite , with autism spectrum disorder@xcite or with psychogenic non - epileptic seizures@xcite as compared to healthy controls . epilepsy represents one of the most common neurological disorders , affecting approximately 1% of the world s population@xcite . in about 25% of individuals with epilepsy , seizures can not be controlled by any available therapy . early conceptualizations of epileptic seizures as either focal or generalized@xcite are being challenged by an increasing evidence of seizure dynamics ( generation , spread , and termination ) within a network of brain regions ( so called epileptic network)@xcite which generate and sustain normal , physiological brain dynamics during the seizure - free interval . using concepts from network theory , previous studies reported on a specific temporal evolution of global properties ( clustering coefficient , average shortest path length , synchronizability ) of epileptic networks@xcite during seizures . this finding is of importance for improving our understanding of seizure dynamics in humans , given the similarity of topological evolution across different types of epilepsies , seizures , medication , age , gender , and other clinical features . in order to provide a more complete view , here we report on the temporal evolution of assortativity of functional brain networks underlying epileptic seizures . we derive these networks from multichannel electroencephalograms which were recorded prior to , during , and after @xmath0 epileptic seizures from 60 patients ( cf . section [ sec : methods ] ) . we assess the robustness of our findings by comparing results obtained from networks derived using two different estimators of signal interdependence . moreover , we demonstrate in a simulation study ( cf . appendix a ) that degree - degree correlations can be spuriously induced by the finite size and frequency content of time series and may not necessarily reflect properties of the underlying dynamics . thus , we compared our results with those obtained from random network ensembles specifically designed to account for the aforementioned influences . after reporting results in section [ sec : results ] , we conclude with interpreting our findings and how they may relate to results obtained in earlier studies ( section [ sec : conclusion ] ) . we retrospectively analyzed electroencephalographic recordings of focal onset seizures from 60 patients with drug - resistant epilepsy . seizure onsets were localized in different anatomical regions . all patients had signed informed consent that their clinical data might be used and published for research purposes . the study protocol had been approved by the ethics committee of the university of bonn . electroencephalograms ( eegs ) were recorded prior to , during , and after 100 epileptic seizures as reported in previous studies @xcite . the multichannel ( @xmath1 channels ) eegs were recorded via chronically implanted strip , grid or depth electrodes from the cortex and/or from within relevant brain structures . data were sampled at 200hz , digitized using a 16-bit analog - to - digital converter , and filtered within a frequency range of 0.570hz . a bipolar referencing was applied prior to subsequent steps of analysis@xcite . electroencephalographic seizure onset and seizure end was automatically determined@xcite . we pursued a sliding window approach which allowed for a time - resolved analysis of the evolution of assortativity of seizure - related functional brain networks . we divided each multichannel recording into non - overlapping consecutive windows of length @xmath2 sampling points ( @xmath3s duration ) , and for each channel we normalized data from each window to zero mean and unit variance . in order to derive a functional network for each window , we associated each channel @xmath4 with a node @xmath4 and derived links by estimating signal interdependencies between all pairs of time series . we employed two commonly utilized methods to assess signal interdependence . the first method makes use of the absolute value of the correlation coefficient between time series at nodes @xmath4 and @xmath5 : @xmath6 where @xmath7 denotes the time series at node @xmath4 with mean @xmath8 and estimated standard deviation @xmath9 . the second method can take into account possible time lags ( e.g. , due to propagation of electrical activity along anatomical pathways during the seizure ) when assessing signal interdependencies and is defined as the maximum value of the absolute cross correlation function , @xmath10 with @xmath11 where @xmath12 denotes the time lag . both estimators are symmetric ( @xmath13 and @xmath14 ) , are confined to the interval @xmath15 $ ] , and assess linear dependencies of time series . here we refrain from assessing nonlinear dependencies since previous studies@xcite reported results obtained from estimators of nonlinear and of linear signal dependence to be qualitatively similar when pursuing network analyses of neuroscientific data . links can now be defined via thresholding the values of signal interdependence such that the adjacency matrix @xmath16of the functional network has entries @xmath17 where @xmath18 $ ] denotes the threshold and @xmath19 either denotes @xmath20 or @xmath21 . we choose @xmath22 such that the resulting network possesses a predefined link density@xcite @xmath23 with the number of nodes @xmath24 and the mean degree @xmath25 ( @xmath26 denotes the degree of node @xmath4 ) . in the following , we set @xmath27 to define links . to assess assortativity of the functional networks we employ the assortativity coefficient @xmath28 , which is defined as the pearson correlation coefficient between the degrees of nodes at both ends of a link@xcite . reformulating the pearson correlation coefficient in terms of the degrees of nodes@xcite ( cf . appendix b ) , we obtain @xmath29 where @xmath30 , and @xmath28 is confined to the interval @xmath31 $ ] by definition . positive ( negative ) values of @xmath28 indicate an assortative ( disassortative ) network . note that the assortativity coefficient is not well defined for regular graphs , i.e. , networks whose nodes share all the same degree . in the following , we denote the assortativity coefficient of the functional networks with @xmath32 or @xmath33 depending on whether signal interdependencies are estimated by the correlation coefficient ( eq . [ eq : cc ] ) or by the maximum value of the absolute cross correlation function ( eq . [ eq : mc ] ) . it was demonstrated in previous studies@xcite that the finite length of empirical time series and their frequency content together with the applied analysis methodology to infer networks can introduce spurious properties in interaction networks . these properties are not related to the analyzed dynamics but reflect the way how networks are derived from finite multichannel data . since the dynamics of epileptic seizures is well known for its complex temporal changes in frequency content@xcite , it is important to investigate whether and to which extent the assortativity coefficient reflects spurious properties induced by the analysis methodology . we detail the findings of our investigation in the appendix ( cf . appendix a ) and briefly note here that the assortativity coefficient takes on larger positive values the larger the relative amount of low - frequency contributions and the shorter the length of time series . in order to account for these influences and to distinguish them from those reflecting the dynamics , we compare the assortativity coefficient of a functional network with those obtained from an ensemble of random networks@xcite . to generate a random network , we proceed as above but estimate signal interdependencies between pairs of surrogate time series@xcite , which preserve the length of the eeg time series from which they are derived , their frequency contents , and their amplitude distributions but are in all other aspects random . the surrogate time series comply with the null hypothesis of independent linear stochastic processes . the resulting random network possesses the same link density as the functional network . we denote the assortativity coefficient of a random network with @xmath34 or @xmath35 depending on the applied estimator for signal interdependencies . in the top panels of fig . [ fig:1 ] , we show time courses of the assortativity coefficients @xmath36(top left ) and @xmath37(top right ) for an exemplary recording of a seizure . both coefficients indicate that the majority of functional brain networks obtained for this recording is assortative . we observe @xmath37to increase at the beginning of the seizure and to decrease at the end of the seizure . a similar course in time is also noticeable for @xmath36(which exhibits a larger variability during the recording ) . for this recording , @xmath36and @xmath37indicate functional brain networks to possess a more assortative structure before than after the seizure . in order to investigate whether these time courses could be explained by the finite size of time series and their frequency content , which varies over time , we generate for each functional brain network an ensemble of 20 random networks and determine the assortativity coefficients @xmath38and @xmath39 . with @xmath40and @xmath41 , we denote the respective mean values calculated from the ensembles . after the seizure and to a lesser extent also before the seizure , @xmath41approximates the values of @xmath37quite well ( cf . top panels of fig . [ fig:1 ] ) . during the seizure , however , pronounced differences between @xmath37and @xmath41as well as between @xmath36and @xmath40indicate that the increased assortativity can not be related to simple alterations in frequency content nor be related to the finite size of time series but reflect features of the seizure dynamics . to summarize our findings for all recordings of 100 focal onset seizures , we need to account for the different durations of seizures ( mean seizure duration : @xmath42 ) . to this end , we partition each seizure into ten equidistant time bins@xcite , assign the estimated assortativity coefficients to their respective time bins , and determine , for each time bin , the mean values ( @xmath43 , @xmath44 , @xmath45 , @xmath46 ) . in addition , we also determined mean values from an equal number of pre - seizure and post - seizure windows . despite the fact that anatomical locations of seizure onset varied across patients , a common time course of the assortativity coefficients becomes apparent ( bottom panels of fig . [ fig:1 ] ) . @xmath43increases at the beginning of seizures , keeps increasing until it reaches a plateau at the end of the seizures , and decreases again after the seizures . a similar behavior can be observed for @xmath44 , which increases at the beginning of the seizures , reaches a plateau already in the middle of the seizures from where it decreases again already prior to the end of seizures . time courses of @xmath45and @xmath46(bottom panels of fig . [ fig:1 ] ) obtained from random networks indicate that values of @xmath43and @xmath44cannot be solely explained by the varying frequency content and the finite length of time series . we compare values of the assortativity coefficients with those obtained from random networks by determining the differences @xmath47 and @xmath48 for each window . we assign these values to their respective time bins , and determine , for each time bin , their mean values ( @xmath49 , @xmath50 ) . the time courses of @xmath51and @xmath52(shown in fig . [ fig:2 ] ) are remarkably similar : both indicate an ongoing increase of assortativity during the seizures , peak in the last fifth part of the seizures , and decrease again after the seizures . and @xmath52 for pre - seizure , discretized seizure , and post - seizure time periods of 100 epileptic seizures . all error bars indicate standard error of the mean . lines are for eye - guidance only.,scaledwidth=40.0% ] we pursued a time - resolved analysis of functional brain networks derived from multichannel eeg data from epilepsy patients recorded before , during , and after seizures . we assessed the assortativity coefficient which quantifies the tendency of nodes to preferentially connect to nodes with a similar degree ( degree - degree correlations ) . our results suggests that seizure dynamics are characterized by assortative functional networks , i.e. , by positive degree - degree correlations . we demonstrated in a simulation study that assortative network structures can also spuriously arise due to the finite size and frequency content of time series and thus like the clustering coefficient and the average shortest path length@xcite do not necessarily reflect properties of the underlying dynamics . by comparing with suitable random networks , however , we corrected for the aforementioned influences and still observed functional brain networks to be assortative during seizures . this result was obtained for two different estimators of signal interdependence that we employed to derive functional brain networks . thus we conclude that positive degree - degree correlations are inherent to seizure dynamics as assessed by our analysis methodology . we observed the assortativity coefficient to change in a characteristic way during seizures which indicates a reorganization of the underlying functional brain networks . while functional brain networks before and after the seizures were less assortative , assortativity slowly increased during seizures and reached a maximum prior to the end of seizures . since this change was observable irrespective of the anatomical onset location , this functional reorganization might be a generic feature of focal seizure dynamics . our observations agree with and complement earlier findings@xcite for the clustering coefficient and the average shortest path length of functional brain networks which were found to increase during seizures while synchronizability decreased . a re - increased synchronizability observed at the end of seizures might indeed promote neuronal synchronization which in turn was hypothesized to be a potential mechanism for seizure termination@xcite . our study sheds light on how the topology of functional brain networks changes during seizures . these changes either promote or weaken the ability of networks to show synchronization : with increasing assortativity during seizures , the hubs of the functional brain network become mutually interconnected while the low - degree nodes become connected to each other . this segregation of the network is likely to result in different groups of brain regions which are if at all only sparsely interconnected . the lower interconnectivity between groups is reflected in larger values of the average shortest path length , while larger values of the clustering coefficient indicate strong connectivity within groups . at the same time , the segregation coming along with increasing assortativity appears to weaken the synchronizability and possibly to hinder the synchronization of the network as a whole . this is in agreement with various numerical studies which reported synchronizability to decrease with increasing assortativity or with increasing clustering coefficient@xcite . in this perspective , seizure termination might be characterized by a reintegration of groups of brain regions within the network thereby diminishing segregation which in turn may lower the threshold for the network to show global synchronization . our findings can inform the development of model studies which may be able to further improve our understanding of mechanisms leading to or aborting global network synchronization . in addition , our study underlines the importance of hubs and their connectivity within the epileptic network . identifying hubs and their role within epileptic networks could help to further our understanding of the generation , propagation , and termination of seizures and might guide the improvement of seizure prevention strategies . we thank g. ansmann and f. ghanbarnejad for helpful comments . this work was supported by the deutsche forschungsgemeinschaft ( grant no . le660/4 - 2 ) . in a previous study@xcite , it was demonstrated that the finite length and the frequency content of time series can introduce spurious properties in interaction networks and can affect their clustering coefficient and average shortest path length . here we investigate whether the assortativity coefficient @xmath28 ( eq . [ eq : assort ] ) is also affected by these influences . we make use of a model which implements the null hypothesis that time series are obtained from independent stochastic processes@xcite . let @xmath53 be time series whose entries @xmath54,@xmath55 , are independently drawn from a uniform probability distribution @xmath56 on the interval @xmath57 . by choosing different length @xmath58 of time series and deriving interaction networks from @xmath53 as described in the methods section , we can study the influence of the length of time series on the assortativity coefficient . we additionally introduce the possibility to obtain time series @xmath7 with varying frequency content ( i.e. , varying serial correlations ) by applying a moving average , @xmath59 where @xmath60 is the size of the moving average and @xmath58 is the length of time series . for @xmath61 , no serial correlations are introduced and we obtain @xmath62 . for @xmath63 , the moving average acts as a low - pass filter . by choosing different values of @xmath60 ( @xmath64 ) and keeping @xmath65 constant , we can study the influence of the frequency content of time series on the assortativity coefficient . we note that @xmath66 and @xmath67 are independent for @xmath68 by construction . we generate an ensemble of @xmath69 networks with @xmath70 nodes each from the time series as defined in equation for given values of @xmath60 and @xmath58 . each realization @xmath71 of a network is derived by thresholding the values of signal interdependence @xmath20 or @xmath21 ( @xmath72 ) obtained from the @xmath73-th realization of pairs of time series ( @xmath74 , @xmath75 ) . the threshold was determined such that the resulting network possesses a link density @xmath27 . with @xmath76 and @xmath77 we denote the assortativity coefficient determined from the @xmath73-th realization of a network which was derived based on @xmath78 or @xmath79 , respectively . mean values determined from all @xmath80 realizations are denoted as @xmath81 and @xmath82 , respectively . in fig . [ fig:3 ] we show the dependence of @xmath83 ( left panel ) and @xmath84 ( right panel ) on the size @xmath60 of the moving average for time series of length @xmath85 . remarkably , for increasing values of @xmath60 and thus for an increasing relative amount of low - frequency contributions in time series , we observe the assortativity coefficient of interaction networks to increase . likewise , if we keep @xmath61 constant and decrease the length of time series @xmath86 , we observe a similar dependency for @xmath87 and @xmath88 on @xmath58 : the assortativity coefficient takes on larger values the shorter the time series . we note that the similarity between @xmath83 and @xmath87 may be traced back to variances of the correlation coefficients between time series @xmath89 and @xmath90 as well as between @xmath91 and @xmath92 which become approximately equal for @xmath86 , the effective length of time series , @xmath93 as argued in a previous study@xcite . to summarize , the finite length of time series and low - frequency contributions likely lead to interaction networks which would be classified as assortative networks according to the positiveness of the assortativity coefficient . this finding is particularly interesting in the light of network null models used to assess the significance of findings in field studies . the expectation value of the assortativity coefficient for frequently used null models like erds - rnyinetworks@xcite or random networks derived from the configuration model @xcite or by random link - switching @xcite is zero ( apart from effects due to a finite number of nodes ) . these models do not take into account the way how interaction networks are usually derived from multivariate field data . although the question of which null model to choose is still a matter of an ongoing scientific debate@xcite , we here use a null model@xcite which accounts for the above demonstrated effects of finite length and low - frequency content of time series : networks are derived from surrogate time series . these surrogate time series possess the same amplitude distribution , the same number of sampling points and approximately the same frequency content as the empirical time series but are random in all other aspects@xcite . in the following we derive equation from the definition of the assortativity coefficient . let us consider an undirected unweighted network consisting of @xmath24 nodes which is defined by its adjacency matrix @xmath94 . to simplify the following steps without changing any of the results , we consider each undirected link of the network to be represented by two directed links . all directed links are contained in the set @xmath80 , and the number of links is denoted as @xmath95 . each link @xmath96 connects two nodes which possess degrees @xmath97 and @xmath98 , respectively . the assortativity coefficient @xmath28 is defined as the pearson correlation coefficient between the degrees of nodes at both ends of a link @xcite , i.e. @xmath99 where cov and @xmath100 denote the covariance and the standard deviation , respectively . note that @xmath101 and @xmath102 since our network is undirected . the mean values @xmath103 and @xmath104 are determined from sums over the links @xmath96 which we can translate into sums over the nodes @xmath105 of the network by making use of the following observations : the number @xmath95 of directed links is equal to the sum of all degrees of nodes , @xmath106 the degree @xmath26 of a node @xmath4 is defined as @xmath107 . thus we obtain for the mean value @xmath103 of the degrees of nodes at one end of links , @xmath108 following the same line of reasoning , we obtain @xmath109 the mean value of the products of the degrees of nodes at both ends of links can be reformulated as @xmath110 where the factor @xmath111 accounts for the fact that we considered directed links . inserting equations , , and into equation , we finally obtain @xmath112 91ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1137/s003614450342480 [ * * , ( ) ] link:\doibase 10.1016/j.physrep.2005.10.009 [ * * , ( ) ] link:\doibase 10.1016/j.physrep.2008.09.002 [ * * , ( ) ] link:\doibase 10.1017/cbo9780511791383 [ _ _ ] ( , , ) link:\doibase 10.5194/npg-13 - 145 - 2006 [ * * , ( ) ] link:\doibase 10.1103/physreve.74.026113 [ * * , ( ) ] link:\doibase 10.5194/npg-15 - 389 - 2008 [ * * , ( ) ] link:\doibase 10.1016/j.physa.2003.10.045 [ * * , ( ) ] link:\doibase 10.1140/epjst / e2009 - 01098 - 2 [ * * , ( ) ] link:\doibase 10.1002/sam.10100 [ * * , ( ) ] link:\doibase 10.1016/j.clinph.2007.08.010 [ * * , ( ) ] link:\doibase 10.1038/nrn2575 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1016/j.clinph.2012.01.011 [ * * , ( ) ] link:\doibase 10.1017/cbo9780511755743 [ _ _ ] ( , , ) link:\doibase 10.1016/j.pneurobio.2005.10.003 [ * * , ( ) ] link:\doibase 10.1016/j.physrep.2006.12.004 [ * * , ( ) ] link:\doibase 10.1016/j.jneumeth.2009.05.015 [ * * , ( ) ] in @noop _ _ , ( , , ) pp . link:\doibase 10.1103/physreve.67.026126 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.89.208701 [ * * , ( ) ] in link:\doibase 10.1109/cdc.2005.1582890 [ _ _ ] ( ) pp . @noop * * , ( ) in link:\doibase 10.1007/11589990_197 [ _ _ ] , , vol . , ( , ) pp . link:\doibase 10.1103/physreve.68.036122 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.94.018102 [ * * , ( ) ] link:\doibase 10.1016/j.physa.2008.06.048 [ * * , ( ) ] link:\doibase 10.1093/brain / awq043 [ * * , ( ) ] link:\doibase 10.1523/jneurosci.2287 - 11.2011 [ * * , ( ) ] link:\doibase 10.1142/s0219635211002725 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1016/j.neuroimage.2011.08.044 [ * * , ( ) ] link:\doibase 10.1371/journal.pbio.0060159 [ * * , ( ) ] link:\doibase 10.1523/jneurosci.1929 - 08.2008 [ * * , ( ) ] link:\doibase 10.1016/j.neuroimage.2010.09.006 [ * * , ( ) ] link:\doibase 10.1103/physreve.75.021915 [ * * , ( ) ] link:\doibase 10.1523/jneurosci.3539 - 11.2011 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1016/j.neuroimage.2010.12.047 [ * * , ( ) ] link:\doibase 10.1371/journal.pone.0035673 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1186/1471 - 2202 - 10 - 101 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , , ) link:\doibase 10.1046/j.1528 - 1157.2002.26901.x [ * * , ( ) ] link:\doibase 10.3389/fnsys.2011.00012 [ * * , ( ) ] link:\doibase 10.1111/j.1528 - 1167.2011.03101.x [ * * , ( ) ] link:\doibase 10.1016/j.physd.2013.06.009 [ * * , ( ) ] link:\doibase 10.1016/j.clinph.2006.12.002 [ * * , ( ) ] link:\doibase 10.1016/j.eplepsyres.2008.02.002 [ * * , ( ) ] link:\doibase 10.1063/1.2966112 [ * * , ( ) ] link:\doibase 10.1016/j.expneurol.2009.02.001 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1063/1.3504998 [ * * , ( ) ] link:\doibase 10.1371/journal.pone.0022826 [ * * , ( ) ] link:\doibase 10.1093/brain / awl304 [ * * , ( ) ] link:\doibase 10.1016/j.clinph.2009.10.013 [ * * , ( ) ] link:\doibase 10.1063/1.3553181 [ * * , ( ) ] link:\doibase 10.1016/s0378 - 8733(99)00011 - 8 [ * * , ( ) ] link:\doibase 10.5194/npg-18 - 751 - 2011 [ * * , ( ) ] link:\doibase 10.1016/s0013 - 4694(98)00024 - 8 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/s1388-2457(00)00259-5 [ * * , ( ) ] link:\doibase 10.1016/s1388 - 2457(02)00344 - 9 [ * * , ( ) ] link:\doibase 10.1111/j.1528 - 1167.2010.02690.x [ * * , ( ) ] link:\doibase 10.1016/s0167 - 2789(00)00043 - 9 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1214/aoms/1177706098 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1007/bf02066689 [ * * , ( ) ] link:\doibase 10.1016/0097 - 3165(78)90059 - 6 [ * * , ( ) ] link:\doibase 10.1002/rsa.3240060204 [ * * , ( ) ] in @noop _ _ , ( , , ) @noop * * , ( ) link:\doibase 10.1016/s0378 - 8733(00)00026 - 5 [ * * , ( ) ] link:\doibase 10.1126/science.1065103 [ * * , ( ) ] link:\doibase 10.1016/j.physa.2003.06.002 [ * * , ( ) ] link:\doibase 10.1103/physreve.72.056708 [ * * , ( ) ] link:\doibase 10.1080/15427951.2010.557277 [ * * , ( ) ] link:\doibase 10.1371/journal.pone.0010012 [ * * , ( ) ] link:\doibase 10.1103/physreve.79.061916 [ * * , ( ) ] link:\doibase 10.1063/1.4732541 [ * * , ( ) ] link:\doibase 10.1016/j.neuroimage.2012.02.001 [ * * , ( ) ] link:\doibase 10.1016/j.neuroimage.2012.01.071 [ * * , ( ) ] link:\doibase 10.1016/j.neuroimage.2013.04.087 [ * * , ( ) ]	we investigate assortativity of functional brain networks before , during , and after one - hundred epileptic seizures with different anatomical onset locations . we construct binary functional networks from multi - channel electroencephalographic data recorded from 60 epilepsy patients ; and from time - resolved estimates of the assortativity coefficient we conclude that positive degree - degree correlations are inherent to seizure dynamics . while seizures evolve , an increasing assortativity indicates a segregation of the underlying functional network into groups of brain regions that are only sparsely interconnected , if at all . interestingly , assortativity decreases already prior to seizure end . together with previous observations of characteristic temporal evolutions of global statistical properties and synchronizability of epileptic brain networks , our findings may help to gain deeper insights into the complicated dynamics underlying generation , propagation , and termination of seizures . 15cm(3cm,26 cm ) copyright ( 2013 ) american institute of physics . this article may be downloaded for personal use only . any other use requires prior permission of the author and the american institute of physics . the article appeared in chaos 23 , 033139 ( 2013 ) and may be found at http://link.aip.org/link/doi/10.1063/1.4821915 doi : 10.1063/1.4821915 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ epilepsy is a disorder of the brain characterized by an enduring predisposition to generate epileptic seizures . it affects more than 50 million individuals worldwide , and for 25% of epilepsy patients , seizures remain poorly controlled despite maximal medical management . there is increasing evidence that focal - onset seizures that appear to originate from a circumscribed region of the brain result from complex interactions in a large - scale epileptic network , which comprises cortical and subcortical brain structures and regions . specific temporal changes of characteristics of epileptic brain networks were hypothesized to reflect an emergent self - regulatory mechanism for seizure termination . we aimed at shedding more light onto this mechanism as it would provide important clues as to how to efficiently control networks underlying seizure dynamics . to do so , we investigate in a time - resolved manner the assortativity of epileptic networks , i.e. , the preference for brain regions ( nodes ) to interact with other brain regions with similar properties . epileptic networks show assortative mixing patterns . these patterns change in a characteristic way and indicate that assortativity increases during seizures , reaches a maximum prior to the end of seizures , and decreases again at the end of seizures . we speculate that these changes may reflect a reorganization of the underlying functional brain networks , which may result in seizure termination . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
ytterbium , both atomic and singly ionised , is an element widely used in experiments involving trapped atoms and ions , such as laser cooling and trapping of neutral atoms and ions @xcite , atomic clocks @xcite , frequency standards @xcite , quantum computing experiments @xcite , quantum optics @xcite and atomic parity non - conservation experiments @xcite . knowledge of the @xmath1 transition line in atomic yb and corresponding frequency shifts for the stable isotopes is very important in these experiments as they allow for laser cooling and isotope selective photoionisation @xcite . various methods have been used to investigate these transitions , and corresponding isotope shifts @xcite . + we present our measurements of the ytterbium @xmath0 transition frequencies which differ from previously published results @xcite . using a simple method based upon observing and aligning fluorescence spots of atomic yb generated from an atomic oven in an evacuated glass beljar we have measured resonant transition frequencies and isotope shifts to an accuracy of 60 mhz . our results are compared to the results obtained using saturation absorption spectroscopy and were further verified by ionising and trapping yb ions . + the advantage of our technique is that it can be used to determine resonance frequencies when the atomic beam is not necessarily perpendicular to the propagation of the laser , which can be the case in many experiments . these doppler shifts also depend on the mean velocity of the atomic beam , so the measurement is quite useful . using our technique we were able to measure the doppler shift of the transition frequency under a range of angles and then derive average velocity of atoms along the direction of atomic beam . to drive the @xmath0 399 nm dipole transition an inexpensive in - house built ultra violet external cavity diode laser ( ecdl ) was used . the ecdl was formed with a laser diode ( frequency selected to 399 nm , sanyo diode : dl-4146 - 301s ) and a diffraction grating ( newport : 53--240r ) positioned in the littrow configuration . a piezo electric actuator was used to adjust the angle of the grating and thus tuning of the emission wavelength . to stabilise the emission wavelength a high finesse wavelength meter ( ws-07 , with specified wavelength accuracy of 60 mhz ) and ni - daq cards operating under labview real time were used . the wavemeter is calibrated using a 780 nm laser which is locked to within 1 mhz of the @xmath2rb d@xmath3 line . it measures the wavelength of the ecdl and sends the information to a labview program , via a ni - daq card ( na-6143 ) . the program determines a drift in wavelength and generates the corresponding error signal to keep the laser stable . + to perform saturation spectroscopy on the @xmath0 399 nm isotope transitions the ecdl frequency was required to scan over a range of 3 ghz . however , adjustment of the grating angle alone in the ecdl resulted in a mode hop free tuning range of only 800 mhz . to overcome this the piezo voltage and laser diode current were adjusted simultaneously to produce a mode hop free tuning range of over 5.5 ghz . for our technique an atomic flux propagating in a known direction was required as this enabled localised fluorescence spots that could be be spatially resolved . to generate the yb atomic beam a resistivly heated stainless steel oven tube ( 2 cm long and 1.5 mm diameter ) was used . a small piece of natural yb was placed inside the oven and yb atomic beam could be obtained by running an electric current of 5 to 6 amperes over the oven . the atomic oven was placed inside a vacuum chamber maintained at a pressure on the order of @xmath4 torr . we have devised a simple method that works by observing fluorescence spots produced by laser - atom interaction at resonant wavelengths . the experimental setup for our spot method can be seen in figure [ fig : scheme ] , here two pairs of non - overlapping counter propagating laser beams ( each with an intensity of approximately 4.5 mw@xmath5 ) are passed through an atomic beam . for orientation the cylindrical axis of the oven tube is considered as the reference axis and the lasers are aligned perpendicular to this reference axis unless otherwise stated . + ( a ) ( b ) the atoms from the oven interact with the four laser beams producing 4 fluorescence spots . in the frame of reference of laser beam propagation each laser interacts with atoms of the same velocity group . when viewed in the laboratory frame the four spots are seen to be misaligned when the frequency is detuned from resonance , but when on resonance the four spots are seen to align perpendicular to the propagation of the laser beams . in this situation each laser is interacting with atoms of zero velocity perpendicular to the atomic beam . spot alignment can be seen in figure [ fig : illustration ] where [ fig : illustration](b ) corresponds to when the laser frequency is on resonance and [ fig : illustration](a)/(c ) corresponds to when the laser frequency is detuned -20 mhz/+20 mhz from resonance respectively . this phenomenon occurs irrespective of the angle between the cylindrical axis of the ovens and the laser beams , which is discussed later . + ( a ) ( b ) ( c ) in order to obtain the @xmath6 transition frequencies , the frequency of the 399 nm laser diode was scanned slowly over 3 ghz . during the scan the florescence spots were imaged using an inexpensive ccd camera setup . after narrowing down the scanning range , the frequencies of the transitions of the different isotopes were individually measured by aligning the four spots in a straight line with the cylindrical axis of the atomic oven . this is shown in figure [ fig : illustration ] . we have measured the frequency of the @xmath7yb@xmath6 transition to be 751.52665 thz @xmath860 mhz which disagrees with a previously published value 751.525987761 thz @xmath860 khz @xcite by approximately 670 mhz . _ @xcite relied in their measurement on the stated accuracy of 20 mhz of a home built wavemeter and we speculate that the uncertainty of that wavemeter had been larger than anticipated by the authors . our measurement differs from the nist atomic spectra database @xcite by 260 mhz , which lists the energy of the yb @xmath9p@xmath10 level at 25068.222 @xmath11 ( 751.52639 thz ) . closer examining this discrepancy , the database lists meggers _ @xcite as the most recent source of their data . however , meggers _ @xcite lists the level energy at 25068.227 @xmath11 ( 751.52654 thz ) which is only 110 mhz away from our measurement , pointing to a possible typographical error on the nist atomic spectra database . considering that meggers _ @xcite used a natural mixture of yb isotopes , their result is consistent with our measurement . we also note that the result by meggers _ @xcite shows an expected 550 mhz discrepancy with the measurement by das _ _ @xcite . using our measurements the isotope shifts relative to the @xmath7yb @xmath0 transition were then calculated . our results are shown in table : [ tab : result ] along with recently published results . the frequency shift between the @xmath12yb @xmath0 , @xmath13yb @xmath14 , and @xmath13yb @xmath15 transitions is 55 mhz , while the separation between the @xmath16yb @xmath0 and the @xmath17yb @xmath18 transitions is 38 mhz @xcite . these close transitions result in overlapping of the fluorescence spots and can not be resolved by spot method . the isotope shifts data presented in table : [ tab : result ] for the yb @xmath0 transitions are in very good agreement with previously published work @xcite . l5cc shift from @xmath7yb ( mhz ) + isotope & transition & this work & ref.@xcite & ref.@xcite & ref.@xcite & ref.@xcite + 168 & @xmath0 & 1883 @xmath830 & 1887.400 @xmath80.05 & & & 1870.2 @xmath85.2 + 170@xmath19 & @xmath0 & 1149 @xmath860 & 1192.393 @xmath80.066 & 1175.7 @xmath88.1 & & 1172.5 @xmath85.7 + 171@xmath19 & @xmath18 & `` & 1153.696 @xmath80.061&1151.4 @xmath85.6&1104@xmath869&1136.2 @xmath85.8 + 171 & @xmath14 & 829 @xmath830 & 832.436 @xmath80.05 & 832.5 @xmath85.6 & 822 @xmath851 & 834.4 @xmath84.0 + 172@xmath19 & @xmath0 & 546 @xmath860 & 533.309 @xmath80.053 & 527.8 @xmath82.8 & 534@xmath833 & + 173@xmath19 & @xmath14 & '' & 515.972 @xmath80.2 & & + 173@xmath19 & @xmath15 & " & 587.986 @xmath80.056 & 578.1 @xmath85.8 & + 173 & @xmath20 & -264 @xmath830 & -253.418 @xmath80.05 & & -262 @xmath816 + 176 & @xmath0 & -509 @xmath830 & -509.310 @xmath80.05 & -507.2 @xmath82.5 & -554 @xmath835 & + + _ * can not be resolved with the spot method _ [ tab : result ] to determine the resonant frequencies and isotope shifts we used the ws7 wavemeter from high finesse ( specified to 60 mhz accuracy ) . the wavemeter is calibrated using a 780 nm laser which is locked to within 1 mhz of the @xmath2rb d@xmath3 line . a he - ne laser ( sios sl02/1 calibrated to 1 mhz ) was then used to provide confirmation of the calibration . a further two point check was performed to ensure calibration in the ultra violet frequency range . using our frequency doubling system , that converts 739 nm light into 369 nm , we measured the two wavelengths simultaneously and confirmed the wavemeter to operate within specifications . the wavemeter was calibrated before and after measurements and no change in calibration measurements were observed . the wavemeter is specified to have an accuracy of 60 mhz for absolute frequency measurements , but is more accurate for relative frequency measurements of closely spaced transitions . for the 399 nm transitions , the relative accuracy is translated to be 20 mhz . therefore the error on transition frequencies is 60 mhz and the error on isotope shifts is 30 mhz . however due to the actual line width of 20 mhz of these transitions and resulting overlap of some transition lines , isotope shift could not be resolved better than 60 mhz . + other sources of error include non - parallel alignment of the lasers where beam misalignment of 1 degree related to 15 mhz error on wavelengths . as the technique can resolve a 20 mhz detuning from resonance it was possible to align the lasers to better than 1 degree as misalignment would result in a visible deviation of the spots . power broadening of the transition line was insignificant as each beam was less than @xmath21 mw@xmath5 whereas the saturation intensity of the transition is 60 mw@xmath5 , which translates to a 3 mhz linewidth broadening . the potential resolution of the technique can be seen from figure [ fig : illustration ] where the frequency of the laser has been detuned from ( b ) resonance by ( a)-20 mhz and ( c)+20 mhz . the separation from the front of the oven to the furthest spot is approximately 22 mm , and by observing spot misalignment at a given detuning we extrapolate that the resolution of the technique can be 10 mhz in this particular setup . transition peaks for different isotopes of yb obtained with saturation absorption spectroscopy . the transitions corresponding to labels 1 to 4 are shown in table [ tab : saturationresults ] . the dashed line represents the doppler background.,width=302 ] to provide a comparison for the transition frequencies and isotope shifts obtained by our spot method , a typical saturation absorption spectroscopy experiment was set up . the saturation spectroscopy setup consisted of a pump and probe beam ( 127 mw@xmath5 and 2 mw@xmath5 respectively ) counter propagating and overlapping through ytterbium atomic beam . by scanning the wavelength of the external cavity diode laser the saturation plot shown in figure [ fig : saturation ] was produced . the light gray line represents the doppler background , which was obtained by scanning the wavelength but blocking the pump beam . with the pump beam included the dark trace was produced . saturation peaks for different isotopes can be seen at label 1 , 2 , 3 and 4 on the graph and corresponding frequencies are listed in table : [ tab : saturationresults ] and compared with the results obtained with the spot method . it can be seen that the results obtained from the saturation spectroscopy are in good agreement with the results from our fluorescence spot method . however the signal to noise ratio using the saturation spectroscopy setup is poor compare to spot method . further to this , saturation spectroscopy can only be used to determine the resonant transition frequencies but not doppler shifted transition frequencies in non - perpendicular laser - atomic beam configurations which can be useful in ion trap experiments . [ cols="^,^,^,^,^ " , ] + _ * not resolved _ [ tab : saturationresults ] in the previous section the atomic motion was perpendicular to the laser beam propagation when measuring resonant transition frequencies . however , in many experiments the atomic motion is not necessarily perpendicular to the laser beams . in this situation a component of the velocity of the atoms is parallel ( or anti parallel ) to the laser beam and the frequency of the resonance line becomes doppler shifted and this doppler shift depends on the temperature of the atomic oven . using our method these doppler shifted transition frequencies can be measured and average velocity of the atoms in the direction of the reference axis and oven temperature can be determined . + and pair ( 1 and 3 ) makes an obtuse angle @xmath22 with the reference axis . to align the pair of spots with the reference axis , laser frequency is blue detuned in ( a ) and red detuned in ( c ) . in ( b ) both spot pairs are aligned ( but not with the reference axis ) when the frequency is at resonance . ] and ( c ) 110@xmath23 . ( b ) also demonstrates the non - doppler shifted resonance where the spots align perpendicular to the laser beams , as seen in the fluorescence spot technique , title="fig : " ] ( a ) and ( c ) 110@xmath23 . ( b ) also demonstrates the non - doppler shifted resonance where the spots align perpendicular to the laser beams , as seen in the fluorescence spot technique , title="fig : " ] ( b ) and ( c ) 110@xmath23 . ( b ) also demonstrates the non - doppler shifted resonance where the spots align perpendicular to the laser beams , as seen in the fluorescence spot technique , title="fig : " ] ( c ) as described earlier when the frequency is detuned from the doppler free resonance two pairs of spots move in opposite directions , relating to the pairs of counter - propagating laser beams . to measure the doppler shifted wavelengths the earlier setup was adjusted by rotating the oven to form an angle with the lasers . this provided maximum atomic flux for the desired angle , while the cylindrical axis of the atomic oven offered a reference axis to which the spots could be aligned . this scheme is illustrated in figure : [ laserbeam ] . here the direction of propagation of the laser beams must be taken into account because one pair of the laser beams ( 2 and 4 ) makes an acute angle @xmath24 while the other pair ( beam 1 and 3 ) forms an obtuse angle @xmath22 with the reference axis . in both cases the magnitude of the angular difference from 90@xmath23 ( and hence the frequency shift ) is the same , but the sign is different . to align a pair of spots to the reference axis the laser frequency was detuned from resonance by a certain amount . blue detuning the laser frequency causes fluorescence spots ( corresponding to lasers 2 and 4 ) to align to the reference axis , while red detuning causes fluorescence spots ( from lasers 1 and 3 ) to align to the reference axis . figure : [ angleshift ] demonstrates these doppler shifted resonance spots when the angle between laser and the atomic beam is ( a ) @xmath25 and ( c ) @xmath26 . a particular advantage of the spot method can be realised in figure : [ angleshift](b ) , that when on resonance the four spots form a line perpendicular to the laser beams . this allows for the resonant transition frequency to be measured independent of the angle between the laser beam and atomic beam . the doppler shift on the resonance frequency depends on the velocity component , of the atoms , parallel ( or anti parallel ) to the propagation of the lasers , which is dependant upon the cosine of the angle @xmath27 between the atomic motion and the propagation of laser . therefore the frequency shift , @xmath28 , is given by @xcite @xmath29 transition of @xmath7yb as a function of angle between the laser beams and atomic beam , width=302 ] where @xmath30 is the doppler free transition frequency , @xmath31 is the speed of light , @xmath32 is the mean velocity of the atoms along the direction of reference axis and @xmath27 is the angle between the atom motion and laser beam . the @xmath33 term represents the mean velocity component parallel to the laser beam . + the reference axis was rotated to make angles of 63 , 70 , 75 and 80 degrees with one pair of lasers , and hence 117 , 110 , 105 and 100 degrees with the other pair of lasers . we measured the doppler shifted frequencies , when the spots aligned with the reference axis , and our measurements of these frequencies are shown in figure [ angle ] . equation [ dopplershift ] was fitted to the data and enables us to derive the average velocity of the atoms in the direction of atomic beam . using our data we have determined the average atomic velocity to be 260@xmath8 20 ms@xmath34 and therefore the temperature of the atomic beam can be estimated . + using frequencies measured from our technique we were able to photoionise and trap yb@xmath35 isotopes in an ion trapping experiment @xcite . to ionise a yb atom a 399 nm photon is required to drive the @xmath0 transition where a further 369 nm photon excites an electron past the continuum @xcite . in the ion trapping experiment the atomic oven and 399 nm laser made an angle of 63@xmath82@xmath23 . we were able to verify the resonance frequencies for the @xmath0 transitions by photoionisation of different yb isotopes . we have demonstrated a simple technique to determine the resonant transition frequencies and isotope shifts , of the 398.91 nm @xmath0 transition , for different stable yb isotopes . we present new values for the resonant frequencies that differ from previously published work @xcite . saturation absorption spectroscopy was also performed and the frequencies measured were in good agreement to those obtained via our florescence spot method . the fluorescence spot method has an advantage over absorption saturation spectroscopy as the former method is easier in setup and provides larger signal to noise ratio . by small modifications to the setup , the fluorescence spot method can also be used to measure the doppler shifted transition frequencies that occur when the laser and the atomic beam are not perpendicular and allowing for predictions of transition frequencies in realistic ion trap setups . furthermore , this technique allows us to derive the average thermal velocity of atoms in the direction of atomic beam . this work was supported by the uk engineering and physical sciences research council ( ep / e011136/1 , ep / g007276/1 ) , the european commission s sixth framework marie curie international reintegration programme ( mirg - ct-2007 - 046432 ) , the nuffield foundation and the university of sussex .	we demonstrate a simple technique to measure the resonant frequency of the 398.9 nm @xmath0 transition for the different yb isotopes . the technique , that works by observing and aligning fluorescence spots , has enabled us to measure transition frequencies and isotope shifts with an accuracy of 60 mhz . we provide wavelength measurements for the transition that differ from previously published work . our technique also allows for the determination of doppler shifted transition frequencies for photoionisation experiments when the atomic beam and laser beam are not perpendicular and furthermore allows us to determine the average velocity of the atoms along the direction of atomic beam .
one of the most exciting discoveries in cosmology after hubble s discovery of expansion of universe is the accelerated expansion of the universe@xcite . in order to explain the acceleration , it becomes necessary to modify the existing theory of gravitation proposed by einstein , ie . , the general theory of relativity . it can be done in two ways , the first one is to modifying the energy - momentum part of the field equation . this necessitates the existence of an exotic component called dark energy . the cosmological constant is the simplest dark energy model . since the effect of dark energy can be felt only at large scales there are no experimental supports to it till now . the second approach to explain the observed acceleration of the universe is to modify the curvature dependent part of field equation . this results in extended theories of gravity . there are a number of extended theories of gravity in literature . f(r ) gravity , gauss - bonnet gravity , teleparallel gravity and new massive gravity are some examples of extended theories of gravity . among them , f(r ) gravity is simple and successful@xcite . even though the f(r ) gravity is most simple and comparatively successful , we can not find out a unified f(r ) model which explains both galaxy rotation curves of different galaxies as well as the accelerated expansion of the universe . some f(r ) models like @xmath0 suffers the problem of ghosts . we must choose an f(r ) model that must be stable for different types of perturbations . the different studies towards the quantum gravity show that there must be terms of curvature invariants but there are no curvature invariants in general f(r ) theory@xcite . the action of f(r ) gravity is obtained by replacing the ricci scalar @xmath1 in the einstein - hilbert action by a general function of @xmath1 . varying the action thus obtained with respect to the metric we get the f(r ) field equation , it is a fourth order equation . comparing the field equation of general theory of relativity which is second order , the field equation of f(r ) gravity is more difficult to solve . static cylindrically symmetric interior solutions in f(r ) gravity is studied by sharif et al@xcite and godel solution is obtained by santos@xcite in f(r , t ) gravity . thermodynamics of evolving lorentzian wormholes is studied by saiedi@xcite . in this work we use a solution originally obtained by sebastiani et al@xcite . scattering of scalar field in this metric has been studied by the authors@xcite . the quasinormal modes are the characteristic sound of black hole , which was first pointed out by vishveshwara@xcite . chandrasekhar and detweller@xcite studied the quasinormal modes of schwarzschild black hole . since then there are a number of works appeared in this field . for a review on quasinormal modes please see the references@xcite and references therein . the study of quasinormal modes becomes relevant because it gives a direct way to detect black hole using the gravitational waves . there are a number of methods for evaluating quasinormal frequencies but the wkb method developed by schutz , will and iyer is the simplest@xcite . the paper is organized as follows . in section 2 we discuss the static spherically symmetric solution in f(r ) gravity in section 3 , we study the quasinormal modes of the extended black hole perturbed by electromagnetic field . in section 4 we study the quasinormal modes of extended black hole using scalar perturbation for massive and massless cases . wkb approximation is discussed in section 5 and finally in section 6 conclusion of the present work is given . starting with the f(r ) action as @xmath2 where @xmath3 is the determinant of the metric and @xmath4 is the general function of ricci scalar @xmath1 . we start with a static spherically symmetric solution.@xcite we take @xmath5 and this particular form of f(r ) is taken as it gives a valid black hole solution . we take this f(r ) as positive , where @xmath6 is an integration constant . this is chosen either as positive or zero and all terms in the radical are taken as positive . we write the static spherically symmetric solution as@xcite @xmath7 with constant @xmath8 , we get the solution as @xmath9 where @xmath10 is given by @xmath11 comparing with schwarzschild - de sitter solution , the @xmath6 term in the solution which is the coefficient of @xmath12 can represent the cosmological constant term . for in this case we choose @xmath13 and @xmath14 . we choose @xmath14 because we restrict our study to a space time which is asymptotically flat@xcite . we select @xmath15 as @xmath16 such that the metric should be positively related to the mass only in such cases we get the correct newtonian limit and @xmath17 is chosen as a length parameter . thus the extended metric in f(r ) gravity is given by@xcite , @xmath18 where @xmath19 is the black hole mass and @xmath17 is a length parameter of the metric . the black hole mass is related to the metric linearly and thus we assume a linear mass term with one length parameter . this length parameter can be adjusted to obtain an effective rotation curves of galaxies and gravitational lensing . f(r ) is chosen as above because only this form of f(r ) will give a static spherically symmetric black hole solution . f(r ) models with @xmath0 , @xmath20 etc.@xcite have been studied but they do not possess black hole solution . we are studying the evolution of maxwell field in this extended space time in f(r ) gravity . the maxwell s equation can be written as , @xmath21 where @xmath22 is the electromagnetic field tensor and @xmath23 is the vector potential and vector potential @xmath23 can be expressed as four dimensional vector spherical harmonics@xcite , @xmath24+\left [ \begin{array}{c } f^{lm}(t , r ) y_{lm}\\h^{lm}(t , r ) y_{lm}\\k^{lm}(t , r)\partial_{\theta } y_{lm}\\k^{lm}(t , r)\partial_{\phi } y_{lm } \end{array}\right ] \right),\ ] ] where @xmath25 and @xmath19 are angular momentum quantum number and azimuthal quantum number respectively . the first column has a parity of @xmath26 and the second column has @xmath27 . we define the tortoise coordinates as @xmath28 such that after some mathematical steps , we get the regge - wheeler@xcite equation as , @xmath29 with the potential @xmath30 given by , @xmath31 where @xmath32 for parity @xmath33 and @xmath34 for parity @xmath27 . using wkb approximation method given in section.5 we can evaluate quasinormal modes of electromagnetic perturbation . the electromagnetic quasinormal modes are given in table . 1 in this section , we study the quasinormal modes of the extended black hole in f(r ) gravity perturbed by scalar fields . the metric is given by eq.(5 ) . since we are considering the scalar perturbation , we can use the klein gordon equation . we consider here both the massless and massive scalar field . the klein - gordon equation for a massive scalar field is , @xmath35 where @xmath36 is the mass of the scalar field . in order to separate the wave function into radial , temporal , and angular parts , we write @xmath37 as @xmath38 substituting eq . 11 in eq . 10 and after straight forward calculation , the radial part is given by @xmath39\phi=0.\ ] ] introducing the tortoise coordinate as @xmath40 , eq.(8 ) can be written as , @xmath41 where the potential v(r ) is given by @xmath42 @xmath43 ranges from @xmath44 to@xmath45 the scalar field perturbation is studied and the corresponding quasinormal modes are obtained and are tabulated in table . ( 2 ) . using the third order wkb approximation method developed by schutz and will@xcite and finally modified by iyer and will@xcite , we can obtain the frequency of the quasinormal mode in general as , @xmath46-i(n+\frac{1}{2})(-2v_{0}'')^{\frac{1}{2}}(1+\omega),\ ] ] where @xmath47,\ ] ] and @xmath48,\end{aligned}\ ] ] where @xmath49 , and @xmath50 the potentials given in eq . 14 corresponding to electromagnetic and scalar perturbations are substituted in the wkb formula and obtained the corresponding complex frequencies . the quasinormal frequencies are tabulated in tables . in table.1 we give the electromagnetic perturbed quasinormal modes and in table . 2 the quasinormal modes when the extended black hole is perturbed by the scalar fields . we have plotted the data in fig.(1 ) , for electromagnetic field case and from this figure we can see that the real part of quasinormal frequencies decreases with increasing mode number @xmath51 for a given angular momentum quantum number @xmath25 . the fig.(2 ) and fig.(3 ) show scalar field perturbations for massless and massive field respectively . in fig.(1 ) fig.(2 ) and fig(3 ) , the top dots are for @xmath52 . the mode value of imaginary part of frequencies increases rapidly showing that the oscillation is damping and black hole is stable against electromagnetic and scalar perturbations . the absolute value of imaginary frequency increases with mode number showing damping . in this paper we have studied the quasinormal modes of extended black hole in f(r ) gravity . the black hole space - time is perturbed by electromagnetic and scalar waves and behavior of the resulting quasinormal modes are evaluated . the present study shows that the imaginary part of complex quasinormal modes for both cases increase showing damping of oscillations . thus the black hole is stable against both scalar and electromagnetic perturbations . the damping time increases with increasing mass of the field . ss wishes to thank csir , new delhi for financial support under csir - srf scheme . vck is thankful to ugc , new delhi for financial support through a major research project and wishes to acknowledge associateship of iucaa , pune , india 0 a. g. riess et al . , _ astronom . j _ * 116 * ( 1998 ) 1009 . s. perlmutter et al . _ j. _ * 517 * ( 1999 ) 565 s. capozziello and v. faraoni , _ beyond einstein gravity_. ( springer 2010 ) . t. p. sotiriou and v. faraoni , _ rev . phys . _ * 82 * 451 ( 2008 ) m. sharif and sadia arif _ mod . * 27 * 1250138 ( 2012 ) a. f. santos _ mod . * 28 * 1350141 ( 2013 ) h. saiedi _ mod . * 27 * 1250220 2012 l. sebastiani and s. zerbini._eur . j. c _ * 71 * ( 2011 ) 15915 s. sebastian and v. c. kuriakose _ mod . a. _ * 29 * 1450005 ( 2014 ) c. v. vishveshwara._nature ( london ) _ * 227 * ( 1970 ) 936 s. chandrasekhar and s. detweller . r. soc . a _ * 344 * ( 1975 ) 441 k. d. kokkotas and b. g. schmidt . _ * 2 * ( 1999 ) 2 h. p. nollert._class . _ * 16 * ( 1999 ) r159 b. f. schutz and c. m. will . _ astrophys . * 291 * ( 1985 ) l33 s. iyer and c. m. will . * 35 * ( 1987 ) 3621 s. iyer . * 35 * ( 1987 ) 3632 t. clifton and j. d. barrow _ phys * 72 * ( 2005 ) 103005 a. d. felice and s. tsujikawa _ living rev . relativity_13 3 ( 2010 ) starobinsky . a. a. _ phys . lett . b_91 99 - 102 ( 1980 ) r. ruffini , in black holes : les astres occlus ( gordon and breachm new york , 1973 ) regge . t and wheeler . j. a _ phys * 108 * ( 1957 ) 1063	in this paper we study the scalar and electromagnetic perturbations of an extended black hole in f(r ) gravity . the quasinormal modes in two cases are evaluated and studied their behavior by plotting graphs in each case . to study the quasinormal mode , we use the third order wkb method . the present study shows that the absolute value of imaginary part of complex quasinormal modes increases in both cases , thus the black hole is stable against these perturbations . as the mass of the scalar field increases the imaginary part of the frequency decreases . thus damping slows down with increasing mass of the scalar field .
the calculation of binding energies , masses of heavy quarkonia @xmath7 and another physical parameter using renormalon method has attracted the attention recently . the calculations , based on perturbative expansions , are primarily due to the knowledge of up to @xmath8 term ( @xmath9 ) of the static quark - antiquark potential @xmath10 @xcite and partial knowledge of the @xmath11 term there , and the ultrasoft gluon contributions to a corresponding effective theory @xmath11 hamiltonian @xcite ; and the knowledge of the pole mass @xmath2 up to order @xmath12 @xcite . another impetus in these calculations was given by the observation of the fact that the contributions of the leading infrared ( ir ) renormalon singularities ( at @xmath13 ) of the pole mass @xmath2 and of the static potential @xmath10 cancel in the sum @xmath14 @xcite . consequently , this cancellation effect must be present also in the total quarkonium mass @xmath15 @xcite , or more precisely , in @xmath3 where @xmath1 is the hard@xmath16soft part of the binding energy , i.e. , the part which includes the contributions of relative quark - antiquark momenta @xmath17 , @xmath18 , i.e. , soft / potential scales ( predominant ) and higher hard scales ( smaller contributions ) . in addition , the binding energy has contribution @xmath19 from the ultrasoft momenta regime @xmath17 , @xmath20 . the ultrasoft contribution is not related to the @xmath13 renormalon singularity , since this singularity has to do with the behavior of theory in the region which includes the hard ( @xmath21 ) and soft / potential ( @xmath22 ) scales . this contribution is based on ref . @xcite and we present numerical calculation of the binding ground energies @xmath23 ( separately the @xmath24 and the @xmath25 parts ) and the mass @xmath26 of the heavy @xmath0 system , by taking into account the leading ir renormalon structure of @xmath27 and @xmath28 , and combining some features as : ( a ) the mass that we use in the perturbation expansions is a renormalon - free mass @xcite which we choose to be the @xmath4 mass @xmath29 ; ( b ) borel integrations @xcite are used to perform resummations . before resummations we perform separation of the soft / potential ( @xmath24 ) and ultrasoft ( @xmath25 ) part of the binding energies , and apply the renormalon - based borel resummation only to the @xmath24 part . the renormalization scales used in the borel resummations are @xmath30 ( hard scale ) for @xmath31 , and @xmath32 for @xmath33 . the term corresponding to @xmath19 is evaluated at @xmath34 whenever perturbatively possible . further , the borel resummations are performed in three different ways : ( a ) using a slightly extended version of the full bilocal expansion refs . @xcite ; ( b ) using a new `` @xmath35-regularized '' full bilocal expansion introduced in the ref . @xcite ; ( c ) using the form of the borel transform where the leading ir renormalon structure is a common factor of the transform @xcite ( we call it @xmath36-method ) . the borel integrations for both @xmath2 and @xmath33 are performed by the same prescription ( generalized principal value pv @xcite ) so as to ensure the numerical cancellation of the renormalon contributions in the sum @xmath37 . furthermore , we demonstrate numerically that in the latter sum the residues at the renormalons are really consistent with the renormalon cancellation when a reasonable factorization scale parameter for the @xmath24-@xmath25 separation is used , while they become incosistent with the aforementioned cancellation when no such separation is used . the numerical results allow us to extract the mass @xmath38 from the known @xmath39 mass of the @xmath0 system . here we calculate the pole mass @xmath2 in terms of the @xmath4 renormalon - free mass @xmath29 and of @xmath40 , using elements of the renormalization group approach of ref . @xcite and the bilocal expansion method refs . the ratio @xmath41 has , divergent , perturbation expansion in @xmath4 scheme which is at present known to order @xmath9 ( ref . @xcite for @xmath42 ; @xcite for @xmath9 . let us define [ sm ] @xmath43 \ , \label{smexp } \\ r_1(\mu ) & = & \kappa_1 + \beta_0 l_m(\mu ) \ , \label{r1 } \\ r_2(\mu ) & = & \kappa_2 + ( 2 \kappa_1 \beta_0 + \beta_1 ) l_m(\mu ) + \beta_0 ^ 2 l_m^2(\mu ) \ , \label{r2 } \\ ( 4/3 ) \kappa_1 & = & 6.248 \beta_0 - 3.739 \ , \label{k1 } \\ ( 4/3 ) \kappa_2 & = & 23.497 \beta_0 ^ 2 + 6.248 \beta_1 + 1.019 \beta_0 - 29.94 \ , \label{k2}\end{aligned}\ ] ] where @xmath44 , while @xmath45 and @xmath46 are the renormalization scheme independent coefficients with @xmath47 being the number of light active flavors ( quarks with masses lighter than @xmath2 ) . the natural renormalization scale here is @xmath48 ( hard scale ) . let us consider @xmath49 as a function of the running coupling @xmath50 , and the perturbation expansion for @xmath49 reads : @xmath51 which has to be summed . then , one defines the borel transform as @xmath52 this series has a finite radius of convergence in the @xmath53-plane , and we introduce the borel function @xmath54 ( borel integral ) corresponding to @xmath55 as @xmath56 and in our case @xmath57 is known to order @xmath58 @xmath59 \ . \label{bsm1}\ ] ] it has renormalon singularities at @xmath60 @xcite . the behavior of @xmath61 near the leading ir renormalon singularity @xmath13 is determined by the resulting renormalon ambiguity of @xmath2 which has to have the dimensions of energy and should be renormalization scale and scheme indepenent the only such qcd scale being @xmath62 @xcite . the stevenson scale @xmath63 @xcite can be obtained in terms of the strong coupling parameter @xmath64 , where @xmath65 ( @xmath66 ) are the parameters characterizing the renormalization scheme , by solving the renormalization group equation ( rge ) @xcite @xmath67 \nonumber\\ & & - \frac{1}{\beta_0 a(\mu ) } + \frac{c_1}{\beta_0 } \ln \left ( \frac { 1 + c_1 a(\mu ) } { c_1 a(\mu ) } \right ) \ \rightarrow \label{solrge } \\ { \widetilde \lambda } & = & \mu \ \exp \left ( - \frac{1}{2 \beta_0 a(\mu ) } \right ) \left ( \frac{1 + c_1 a(\mu ) } { c_1 a(\mu ) } \right)^{\nu } \nonumber\\ & & \exp\left [ - \frac{1}{2 \beta_0 } \int_0^{a(\mu ) } dx \ \frac { ( c_2 + c_3 x + c_4 x^2 + \cdots ) } { ( 1 + c_1 x ) ( 1 + c_1 x + c_2 x^2 + \cdots ) } \right ] \ , \label{tl1}\end{aligned}\ ] ] where @xmath68 . expansion of expression ( [ tl1 ] ) in powers of @xmath69 then gives @xmath70 \ , \label{tl2}\ ] ] where @xmath71 \ , \nonumber \\ { \widetilde r}_3 & = & \frac{1}{48 \beta_0 ^ 3 } [ ( c_1 ^ 2 - c_2 ) ^3 - 6 \beta_0 ( c_1 ^ 2 - c_2 ) ( c_1 ^ 3 - 2 c_1 c_2 + c_3 ) \nonumber \\ & & + 8 \beta_0 ^ 2 ( c_1 ^ 4 - 3 c_1 ^ 2 c_2 + c_2 ^ 2 + 2 c_1 c_3 - c_4 ) ] \ .\end{aligned}\ ] ] the singular part of the borel transform @xmath57 around @xmath13 must have the form [ bstc ] @xmath72 + b_{s}^{\rm ( an.)}(b ; \mu ) \ , \label{bsrenan}\ ] ] @xmath73 and @xmath74 is analytic on the disk @xmath75 . the @xmath4 coefficients for @xmath76 : @xmath77 , @xmath78 and @xmath79 are already known @xcite , but for @xmath80 we have only estimates @xcite obtained by pad - related methods . we consider @xmath81 . thus , @xmath82 can be obtained : @xmath83 , @xmath84 and @xmath85 . the bilocal method @xcite consists of taking in the expansion ( [ bsrenan ] ) for the analytic part @xmath86 a polynomial in powers of @xmath53 , so that the expansion of @xmath61 around @xmath87 agrees with expansion ( [ bsm1 ] ) . for that , the residue parameter @xmath88 in eq . ( [ bsrenan ] ) has to be determined . using refs . @xcite : @xmath89 and according to ( [ bsrenan ] ) @xmath90 then , the @xmath88 is estimated ( see ref . @xcite ) , using @xmath91 tps and pad approximation @xmath92 $ ] : @xmath93 the bilocal expansion ( [ bsrenan ] ) has then for the analytic part the polynomial [ bsan ] @xmath94 where , by convention , @xmath95 . then , the bilocal formula , is @xmath96 + \sum_{k=0}^2 \ \ \beta_0^k } \ b^k \ . \label{bsbiloc}\ ] ] applying the ( generalized ) principal value ( pv ) prescription for the borel integration @xmath97 we obtain the pole mass @xmath2 in terms of the mass @xmath5 . the numerical integration is performed , using the cauchy theorem ( refs . @xcite ) . in figs . [ mqmu.fig ] ( a ) we present the resulting ( pv ) pole masses of the @xmath53 , as function of the renormalization scale @xmath98 . the spurious @xmath98-dependence is very weak . in addition , results of another method ( `` r''-method ) are presented in figs . [ mqmu.fig ] ( a ) , with the @xmath98-dependence stronger in the low-@xmath98 region ( @xmath99 ) . the r - method ( refs . @xcite ) consists in the borel integration of the function ( [ rsm ] ) @xmath100 where for @xmath91 the corresponding ( nnlo ) tps is used . when we take @xmath101 gev and we vary the values of the residue parameter @xmath88 , the bilocal method gives , at @xmath102 , variation @xmath103 mev . when the central values of @xmath88 are used , the variation of the obtained values of @xmath2 with @xmath98 , when @xmath104 grows from 1.0 to 1.5 , is about @xmath105 mev for @xmath27 ( for @xmath36-method : @xmath106 mev ) . when @xmath80 is varied , the variation is about @xmath107 and @xmath108 mev for @xmath27 . the uncertainty in @xmath109 can be taken as @xmath110 @xcite , corresponding to @xmath111 . this uncertainty is by far the major source in the variation of the pole masses : ( @xmath112 mev for bilocal method ( @xmath113 mev for @xmath36-method ) . the natural renormalization scale @xmath98 here is a hard scale @xmath114 , and will be denoted later in this work as @xmath115 in order to distinguish if from the `` soft '' renormalization scale @xmath98 used in the analogous renormalon - based resummations of the ( hard@xmath16)soft binding energy @xmath33 ( @xmath116 ) in sec . [ eqq ] . the fact that the two renormalization scales are different does not affect the mechanism of the ( @xmath117 ) renormalon cancellation in the bilocal calculations of the meson mass @xmath118 , because the renormalon ambiguity in each of the two terms is renormalization scale independent @xmath119 , as seen by eqs . ( [ tl2])([bstc ] ) . on the other hand , if @xmath36-type methods ( [ brsint ] ) [ cf . also eq . ( [ nmform ] ) ] are applied for the resummations of @xmath120 and @xmath33 , the renormalon ambiguities are renormalization scale independent in the approximation of the one - loop rge running , and the renormalon cancellation is true at this one loop level . the perturbation expansion of the ( hard @xmath16 soft @xmath16 ultrasoft ) binding energy @xmath121 of the @xmath7 heavy quarkonium vector ( @xmath122 ) or scalar ( @xmath123 ) ground state ( @xmath124 ) up to the @xmath125 @xmath126 was given in @xcite the reference mass scale used was the pole mass @xmath2 . the ground state energy expansion has the form @xmath127 \nonumber\\ & & + a^2(\mu ) \left [ k_{2,0 } + k_{2,1 } l_p(\mu ) + k_{2,2 } l_p^2(\mu ) \right ] \nonumber\\ & & + a^3(\mu ) \left [ k_{3,0 } + k_{3,1 } l_p(\mu ) + k_{3,2 } l_p^2(\mu ) + k_{3,3 } l_p^3(\mu ) \right ] + { \cal o}(a^4 ) { \big \ } } \ , \label{eqqexp}\end{aligned}\ ] ] where @xmath128 the coefficients @xmath129 of perturbation expansion ( [ eqqexp ] ) of the quarkonium ( @xmath130 ; @xmath131 ; @xmath122 or @xmath132 ) are given below . @xcite , @xcite for @xmath133 : [ kij ] @xmath134 } \ , \label{k30}\end{aligned}\ ] ] here , @xmath135 have been estimated in ref . @xcite , obtained from the condition of renormalon cancellation in the sum @xmath136 [ a3nf ] @xmath137 the coefficients @xmath129 in the expansion ( [ eqqexp ] ) originate from quantum effects from various scale regimes of the participating particles : ( a ) the hard scales ( @xmath138 ) ; ( b ) the soft and potential scales where the three momenta are @xmath139 ( @xmath140 in the soft and @xmath141 in the potential regime ) ; ( c ) ultrasoft scales where @xmath142 and @xmath143 are both @xmath144 . the coefficients are dominated by the soft scales ; the hard scales start contributing at the nnlo @xcite and are numerically smaller . for this reason , we will usually refer to the combined soft and hard regime contributions to the binding energy as simply soft ( @xmath24 ) contribution @xmath145 . strictly speaking , it is only the pure soft regime that contributes to the @xmath13 renormalon . however , for simplicity , we will resum the hard@xmath16soft contributions @xmath33 together . this will pose no problem , since the hard regime , being clearly perturbative , is not expected to deteriorate the convergence properties of the series for @xmath33 . the natural renormalization scale @xmath98 in the resummations of @xmath33 is expected to be closer to the soft scale ( @xmath146 ) . on the other hand , the @xmath125 coefficient @xmath147 obtains additional contributions from the from the ultrasoft ( @xmath25 ) regime . the leading ultrasoft contribution comes from the exchange of an ultrasoft gluon in the heavy quarkonium @xcite . it consists of two parts : 1 . the retarded part , which can not be interpreted in terms of an instantaneous interaction @xmath148 where @xmath149 is the qcd bethe logarithm - see refs . the non - retarded part can be calculated as expectation value of the @xmath25 effective hamiltonian @xmath150 in the coulomb ( i.e. , leading order ) ground state @xmath151 , where @xmath150 ( in momentum space ) was derived in refs . . direct calculation of the expectation value , here in coordinate space , then gives : [ k30usnr ] @xmath152 } \times { \big \ { - \frac{27 \pi^3}{8 } a^3(\mu ) \langle 1 \| \frac{1}{r } \| 1 \rangle \nonumber \\ & & - 17 \pi^2 \frac{a^2(\mu)}{m_q } \langle 1 \| \frac{1}{r^2 } \| 1 \rangle + \frac{4 \pi^2}{3 } \frac{a(\mu)}{m_q^2 } \langle 1 \| \delta({\bf r } ) \| 1 \rangle \nonumber \\ & & + 3 \pi \frac{a(\mu)}{m_q^2 } \langle 1 \| \ { \delta_{\bf r } , \frac{1}{r } \ } \| 1 \rangle { \big \ } } \ , \label{k30usnr1 } \\ & = & - 14.196 \left [ \ln \left ( \frac { \mu_f}{m_q \alpha_s^2(\mu ) } \right ) + 0.9511 \right ] \ . \label{k30usnr2}\end{aligned}\ ] ] here , @xmath153 is the coulomb energy of the state @xmath151 , and @xmath154 is the factorization energy between the soft ( @xmath155 ) and ultrasoft ( @xmath156 ) scale . the @xmath24@xmath25 factorization scale @xmath154 can be estimated as being roughly in the middle between the @xmath24 and @xmath25 energies on the logarithmic scale @xcite @xmath157 = \kappa \ m_q \alpha_s(\mu_s)^{3/2 } \ , \label{muf}\ ] ] where @xmath158 and @xmath159 ( @xmath160 ) . therefore , the ultrasoft part of the @xmath125 coefficient @xmath147 can be rewritten , by eqs . ( [ k30usr ] ) , ( [ k30usnr ] ) and ( [ muf ] ) , in terms of the @xmath24@xmath25 parameter @xmath161 as @xmath162 the soft scale @xmath163 appearing here will be fixed by the condition @xmath164 . the formal perturbation expansions for the separate soft and ultrasoft parts of the ground state binding energy ( [ eqqexp ] ) are then [ eqqsus ] @xmath165 + { \cal o}(a^4 ) { \big \ } } \ , \label{eqqs } \\ e_{q \bar q}(us ) & = & - \frac{4}{9 } m_q \pi^2 a^2(\mu ) { \big \ { } a^3(\mu ) k_{3,0}(us ) + { \cal o}(a^4 ) { \big \ } } \ . \label{eqqus}\end{aligned}\ ] ] the energy @xmath33 ( [ eqqs ] ) contains the leading ir renormalon effects , and @xmath19 ( [ eqqus ] ) does not . in these expressions , the common factor is the soft scale @xmath166 which is also present as the reference scale in the logarithms @xmath167 appearing with the coefficients @xmath129 ( when @xmath168 ) in eqs . ( [ eqqexp ] ) , ( [ lp ] ) . we will re - express @xmath2 everywhere in @xmath121 with the renormalon - free mass @xmath5 , and will consider the dimensionless soft - energy quantity @xmath169 . thus , we will divide the soft binding energy with the quantity @xmath170 , where @xmath171 can be any soft scale . we will fix this scale by the condition @xmath172 ( @xmath173 ) . further , in the logarithms @xmath174 we express the pole mass @xmath2 in terms of @xmath5 and powers of @xmath69 ( cf . sec . [ mass ] ) , and the powers of logarithms @xmath175 $ ] we re - express in terms of @xmath176 $ ] . this then results in the following soft binding energy quantity @xmath177 to be resummed @xmath178 \ , \label{fs}\end{aligned}\ ] ] where the coefficients @xmath179 depend on @xmath180 and on three scales : the renormalization scale @xmath98 ( @xmath181 ) , the ( fixed ) soft scale @xmath182 , and @xmath5 . the coefficient @xmath183 depends , in addition , on the parameters @xmath161 ( [ muf])-([k30us ] ) , @xmath163 , and @xmath135 ( [ a3nf ] ) . the coefficients @xmath179 are written explicitly in appendix [ app : sbe ] . the @xmath13 renormalon in the quantity @xmath177 is then of the type of the renormalon of the pole mass @xmath2 discussed in the previous sec . [ mass ] . however , if we divided in eq . ( [ fs ] ) by @xmath2 instead of @xmath184 and at the same time used in the resulting @xmath179-coefficients @xmath185 , the numerical resummations of @xmath177 by methods of sec . [ eqq ] would give us values for @xmath33 different usually by not more than @xmath186 ( we checked this numerically ) . we will briefly refer to these approaches later in this section as `` pole mass '' approaches . a version of such pole mass bilocal approach was applied in ref . @xcite for resummation of the unseparated @xmath187 . the ultrasoft part ( [ eqqus ] ) , on the other hand , has no @xmath188 renormalon . the mass scale used there should also be renormalon free ( @xmath5 ) . the renormalization scale @xmath98 there should be adjusted downward to the typical @xmath25 scale of the associated process @xmath189 ( @xmath190 ) in order to come closer to a realistic estimate @xmath191 the soft binding energy quantity to be resummed is @xmath177 of eq . ( [ fs ] ) . however , in the @xmath11 coefficient @xmath183 we have dependence on @xmath135 , and on the @xmath24-@xmath25 factorization scale parameter @xmath193 . then , the value of @xmath161 ( [ muf ] ) is obtained requiring that the residue parameter values be reproduced from the borel transform of the soft binding energy quantity @xmath177 of eq . ( [ fs ] ) . similarly as in eq . ( [ bsrenan ] ) , we have @xmath194 + b_{f(s)}^{\rm ( an.)}(b ; \mu)\ , \label{bfrenan}\ ] ] where the factor in front of the singular part was determined by the condition of renormalon cancellation of the sum @xmath37 . we now define in analogy with eq . ( [ rsm ] ) @xmath195 here we denoted , explicitly the dependence on the factorization scale @xmath154 and @xmath196 theoretically , @xmath197 should be a function with only a weak singularity ( cut ) at @xmath13 , and the nearest pole at @xmath198 ( @xcite ) . resummations such as pad approximations ( pa s ) are applied , then @xmath199(b)$ ] has physically acceptable pole structure @xmath200 for most of the values of @xmath201 and @xmath158 . and the residue parameter @xmath88 is reasonably stable under the variation of @xmath98 . in fig . [ nmbb.fig](a ) we show the dependence of @xmath88 on @xmath161 , at a typical @xmath98 value @xmath202 gev , for the @xmath203 system . the know value of @xmath88 is obtained by the @xmath199(b\!=\!1/2)$ ] expression at @xmath204 . in fig . [ nmbb.fig](b ) we present , for @xmath205 , the dependence of calculated @xmath88 on the renormalization scale @xmath98 . there , we include also the ( @xmath206$]-resummed ) curve for the case when no separation of the @xmath24 and @xmath25 parts of the energy is performed . in that case , the obtained values of @xmath88 are unacceptable . the other values of the input parameters are chosen to have the @xmath203 `` central '' values : @xmath207 ; @xmath101 gev ; @xmath208 gev ( @xmath209 ) and @xmath210 ) [ from : @xmath211 , i.e. , @xmath212 @xcite . for the rge running , we use four - loop @xmath4 @xmath213-function ( tps ) . variation @xmath214 [ @xmath203 ] implies @xmath216 . if , on the other hand , @xmath135 parameter is varied , then for @xmath203 @xmath217 . thus , the value of @xmath24-@xmath25 factorization scale parameter @xmath161 is influenced largely by the allowed values of the renormalon residue parameter , and significantly less by the allowed values @xmath135 of the @xmath11 coefficient of the static @xmath218 potential . therefore , we will consider the variations of @xmath88 and of @xmath161 to be related by a one - to - one relation , while the variations of @xmath135 will be considered as independent . in this way , we have the following value for the @xmath24-@xmath25 factorization scale parameter : @xmath219 and thus we obtain the @xmath11 tps ( [ fs ] ) for the soft part of the ground binding energy . now with the value of @xmath161 , we can perform the resummation of the soft part of the ground binding energy . the full bilocal method @xcite can be performed as in sec . [ mass ] , eqs . ( [ bsbiloc ] ) and ( [ bsint ] ) . therefore @xmath220 where the coefficients @xmath221 in the expansion of the analytic part are known up to order @xmath222 @xmath223 the result have spurious @xmath98-dependence , and for the of the borel transformation turns out to have a problematic behavior : the obtained pole of @xmath224(b)$ ] is unacceptably small in size : @xmath225 . theoretically , @xmath226 should have the nearest pole at @xmath198 @xcite . the reason for this problem appears to lie in the specific truncated form of the singular part taken in the bilocal method . while the latter part describes well the behavior of the transform near @xmath13 , it influences apparently strongly the coefficients @xmath221 and thus the analytic part , so that no reliable resummation of that part ( apart from tps ) can be done . this problem can be alleviated by introducing a `` form '' factor which suppresses that part away from @xmath227 , but keeps it unchanged at @xmath227 . we choose a gaussian type of function , and the `` @xmath35-regularized '' bilocal expressions for the borel transform is @xmath228 \nonumber \\ & & \times \exp \left [ - \frac{1}{8 \sigma^{2 } } ( 1\!-\!2 b)^2 \right ] + \sum_{k=0}^{3 } \ \frac{1}{k ! \ \beta_{0}^{k } } \ h^{(\sigma)}_{k } b^{k } \ . \label{fssig}\end{aligned}\ ] ] the corrective terms @xmath229 and @xmath230 in the coefficients of eq . ( [ fssig ] ) appear to ensure the correct known behavior of the renormalon part up to order @xmath231 . numerical analysis indicate that @xmath35 is between zero and one . namely , when @xmath35 decreases from @xmath232 to about @xmath233-@xmath234 , the value of the pole of the @xmath206 $ ] pad - resummed analytic part @xmath235 of eq . ( [ fssig ] ) gradually turns acceptable ( @xmath236 ) and rather stable when the renormalization scale @xmath98 varies in the interval @xmath237 $ ] ( except close to @xmath238 ) . when the value of @xmath35 falls below @xmath239 , the analytic part starts showing erratic behavior again and the borel resummation gives significantly differing results with the tps- and the pad - evaluated analytic parts . on these grounds , the obtained optimal @xmath35 turn out to be [ sig ] @xmath240 in fig . [ eqqsig.fig](a ) we present the borel - resummed soft part of ground state energy for the bottonium ( @xmath122 ) , as a function of the @xmath35 parameter of method ( [ fssig ] ) . finally , the results for the soft binding energy @xmath28 of the ground state of bottonium using the @xmath36-method @xcite , where we resum the function @xmath241 ( [ rfs ] ) and then employ the ( pv ) borel resummation as written in eq . ( [ brsint ] ) ( with @xmath242 instead of @xmath243 there ) . , as functions of the renormalization scale @xmath98 , are presented in fig . [ ebbmu.fig](a ) . we observe from the figure that the bilocal `` @xmath35-regularized '' method ( [ fssig ] ) ( @xmath244 ) gives the tps and pa results closer to each other . the methods @xmath35-tps , @xmath35-pa , and @xmath36-pa give similar results in the entire presented @xmath98-interval . @xmath36-tps appears to fail at low @xmath98 ( @xmath245-@xmath246 gev ) . in fig . [ ebbmu.fig](b ) we include , for comparison , the simple tps evaluation of @xmath28 , according to formula [ cf . ( [ fs ] ) ] @xmath247 \ , \label{fstps}\ ] ] where for @xmath8 tps case we take @xmath248 . in fig . [ ebbmu.fig](b ) the same input parameters are used as in fig . [ ebbmu.fig](a ) . we see that the perturbation series shows strongly divergent behavior already at @xmath11 . in this figure , we also included the `` perturbative '' ultrasoft part @xmath249 calculated according to [ see eqs . ( [ k30us ] ) and ( [ eqqus ] ) ] . @xmath250 the estimate of the perturbative part is given in eq . ( [ eqqus2 ] ) , where it was essential to take for the renormalization scale a @xmath25 scale @xmath251 . for the bottonium case , this scale is below @xmath252 gev , the energy at which we can not determine perturbatively @xmath6 . this indicates that in the bottonium the @xmath25 part of the binding energy has an appreciable nonperturbative part . the lowest energy at which we can still determine perturbatively @xmath109 is @xmath253-@xmath254 gev , giving @xmath255 . although this is a soft scale for @xmath203 , we will use this also as an ultrasoft scale . then by eq . ( [ eqqus2 ] ) @xmath256 the nonperturbative contribution coming from the gluonic condensate is given by @xcite @xmath257 where we used @xmath258 gev , and the value of the gluon condensate @xmath259 @xcite . ( [ ebbusp ] ) and ( [ ebbusnp ] ) give @xmath260 the finite charm mass contributions has been calculated in ref . @xcite ( refs . the contribution to the mass @xmath261 is @xmath262 the estimates ( [ ebbus ] ) , and ( [ mcnot0 ] ) then give a rough estimate of the @xmath25 and @xmath263 contributions to the bottonium mass @xmath264 mev . the mass of the @xmath39 vector bottonium ground state is well measured @xmath265 mev with virtually no uncertainty @xcite . therefore , the pure perturbative `` soft '' mass is @xmath266 where the uncertainty is dominated by the uncertainty of the @xmath25 regime contribution . our numerical results for @xmath28 and for @xmath27 allow us , by varying the input value of @xmath38 , to adjust the sum @xmath267 to the value given in eq . ( [ mus ] ) . for the soft binding energy we apply the @xmath35-regularized bilocal methods @xmath35-tps and @xmath35-pa , and @xmath36-tps and @xmath36-pa , with the input parameters : @xmath212 ; @xmath208 gev ( @xmath209 ) , thus @xmath268 [ @xmath269 ; @xmath270 ; @xmath205 ; @xmath271 ; @xmath207 ; @xmath272 . for @xmath273 we apply the bilocal - tps and @xmath36-tps method , with renormalization scale @xmath274 , both methods giving us very similar results [ cf . [ mqmu.fig ] ( a ) ] . the bilocal - tps method is applied for @xmath273 when @xmath35-tps and @xmath35-pa are applied for @xmath28 ; the @xmath36-tps is applied for @xmath273 when @xmath36-tps and @xmath36-pa are applied for @xmath28 . the extracted values of @xmath275 are then @xmath276 finally , the average of the central values gives us @xmath277 our result for the mass @xmath278 , is compared in table [ table7 ] with recently values of @xmath38 . the only input parameter common to all these methods is @xmath109 . our central value was @xmath279 [ @xmath280 since such @xcite , or similar @xcite , values follow from the ( nonstrange ) semihadronic @xmath281 decay data which are very precise @xcite . on the other hand , the world average as of september 2002 is @xmath282 @xcite . most of the authors during the last four years used central value @xmath283 . therefore , for comparisons , we convert our results ( [ mbb ] ) to this central value of @xmath109 more specifically , from @xmath111 to @xmath284 . this can be easily done by inspecting in table [ table2 ] the column under @xmath109 , an increase in the central values of @xmath38 by @xmath285 , @xmath286 , @xmath287 and @xmath288 mev , respectively . this gives the average @xmath288 mev higher than in eq . ( [ mbbav ] ) @xmath289 there are two important numerical effects in our result . the first is the separate evaluation of the `` perturbative '' ultrasoft energy part at the corresponding low renormalization energy ( @xmath290 gev ) , eqs . ( [ eqqus2 ] ) and ( [ ebbusp ] ) . if we had not separated the ( `` perturbative '' ) ultrasoft from the soft part of the binding energy , the use of the common renormalization energy scale @xmath98 ( @xmath291 gev ) in the resummation then would have given us the central value of @xmath292 by about @xmath293 mev higher . then the extracted value of @xmath38 would have gone down by about @xmath294 mev , giving the value @xmath295 , with the central value close to that of l03 in table [ table7 ] . on the other hand , that latter value is quite clearly lower than the value ps02 in table [ table7 ] , by about @xmath296 mev , principally because of the @xmath13 renormalon effect which were taken into account here and in ref . thus , the renormalon effect brings down the extracted central value of @xmath38 by about 150 mev , but the separate evaluation of the ultrasoft contribution brings it up by about @xmath297 mev . i would like to thank g. cvetic for helpful and discussions . this work was supported by project usm-110321 of the utfsm . we write down here the explicit coefficients @xmath179 of the expansion ( [ fs ] ) for the soft part of the ground state binding energy . the logarithms appearing in these expressions involve three scales [ @xmath298 and @xmath299 @xmath300 the coefficients @xmath179 are @xmath301 [ f2 ] @xmath302}\nonumber \ , \label{f20 } \\ f_{2}^{(1 ) } & = & { \big [ } -42.7469 - 11 l_1 ^ 2 - 3.6667 l_{\mu}^2 + l_{\mu } ( 26.6944 - 7.3333 l_2 ) + 6.8056 l_2 \nonumber\\ & & + l_1 ( -33.0556 + 14.6667 l_{\mu } + 7.3333 l_2 ) { \big ] } \nonumber \ , \label{f21 } \\ f_{2}^{(2 ) } & = & { \big [ } 1.16286 + ( 3/9 ) l_1 ^ 2 + ( 1/9 ) l_{\mu}^2 + l_1 ( 1 - ( 4/9 ) l_{\mu } - ( 2/9 ) l_2 ) \nonumber\\ & & + l_{\mu } ( -0.81482 + ( 2/9 ) l_2 ) - 0.18518 l_2 { \big ] } \nonumber \ . \label{f22}\end{aligned}\ ] ] [ f3 ] @xmath303}\nonumber \ , \label{f30 } \\ f_{3}^{(1 ) } & = & { \big [ } -1274.33 - 1277.92 l_1 - 471.125 l_1 ^ 2 - 121 l_1 ^ 3 + 1182.32 l_{\mu } \nonumber\\ & & + 843.667 l_1 l_{\mu } + 272.25 l_1 ^ 2 l_{\mu } - 335.813 l_{\mu}^2 - 181.5 l_1 l_{\mu}^2 \nonumber\\ & & + 30.25 l_{\mu}^3 + 124.501 l_2 + 108.361 l_1 l_2 + 90.75 l_1 ^ 2 l_2 \nonumber\\ & & -186.708 l_{\mu } l_2 - 181.5 l_1 l_{\mu } l_2 + 90.75 l_{\mu}^2 l_2 + 36.729 l_2 ^ 2 \nonumber\\ & & + ( 4.0686 + 15.5964 l_1 - 11.6973 l_{\mu } - 3.8991 l_2 ) s ( s+1 ) { \big ] } \nonumber \ , \label{f31 } \\ f_{3}^{(2 ) } & = & { \big [ } 70.8992 + 70.2453 l_1 + 28.9722 l_1 ^ 2 + 7.3333 l_1 ^ 3 - 65.9925 l_{\mu } \nonumber\\ & & -51.6667 l_1 l_{\mu } - 16.5 l_1 ^ 2 l_{\mu } + 20.5972 l_{\mu}^2 + 11 l_1 l_{\mu}^2 -1.8333 l_{\mu}^3 \nonumber\\ & & - 5.1939 l_2 - 6.5741 l_1 l_2 - 5.5 l_1 ^ 2 l_2 + 10.9167 l_{\mu } l_2 \nonumber\\ & & + 11 l_1 l_{\mu } l_2 - 5.5 l_{\mu}^2 l_2 - 2.0972 l_2 ^ 2 { \big ] } \nonumber \ , \label{f32 } \\ f_{3}^{(3 ) } & = & { \big [ } -1.21475 - 1.21714 l_1 - ( 5/9 ) l_1 ^ 2 - 0.14815 l_1 ^ 3 + 1.16286 l_{\mu } \nonumber\\ & & + l_1 l_{\mu } + ( 1/3 ) l_1 ^ 2 l_{\mu } - 0.40741 l_{\mu}^2 - ( 2/9 ) l_1 l_{\mu}^2 + 0.03704 l_{\mu}^3 \nonumber\\ & & + 0.05429 l_2 + ( 1/9 ) l_1 l_2 + ( 1/9 ) l_1 ^ 2 l_2 \nonumber\\ & & - 0.18518 l_{\mu } l_2 - ( 2/9 ) l_1 l_{\mu } l_2 + ( 1/9 ) l_{\mu}^2 l_2 + 0.03704 l_2 ^ 2 { \big ] } \nonumber \ . \label{f33}\end{aligned}\ ] ] m. peter , phys . rev . lett . * 78 * ( 1997 ) 602 [ arxiv : hep - ph/9610209 ] ; nucl . b * 501 * ( 1997 ) 471 [ arxiv : hep - ph/9702245 ] . y. schrder , phys . lett . b * 447 * ( 1999 ) 321 [ arxiv : hep - ph/9812205 ] ; nucl . * 86 * ( 2000 ) 525 [ arxiv : hep - ph/9909520 ] . b. a. kniehl , a. a. penin , v. a. smirnov , and m. steinhauser , nucl . b * 635 * ( 2002 ) 357 [ arxiv : hep - ph/0203166 ] . n. brambilla , a. pineda , j. soto and a. vairo , phys . d * 60 * ( 1999 ) 091502 [ arxiv : hep - ph/9903355 ] ; b. a. kniehl and a. a. penin , nucl . b * 563 * ( 1999 ) 200 [ arxiv : hep - ph/9907489 ] . n. gray , d. j. broadhurst , w. grafe and k. schilcher , z. phys . c * 48 * ( 1990 ) 673 ; l. v. avdeev and m. y. kalmykov , nucl . b * 502 * , 419 ( 1997 ) [ arxiv : hep - ph/9701308 ] . k. g. chetyrkin and m. steinhauser , nucl . b * 573 * ( 2000 ) 617 [ arxiv : hep - ph/9911434 ] ; k. melnikov and t. v. ritbergen , phys . b * 482 * ( 2000 ) 99 [ arxiv : hep - ph/9912391 ] . a. h. hoang , m. c. smith , t. stelzer and s. willenbrock , phys . d * 59 * ( 1999 ) 114014 [ arxiv : hep - ph/9804227 ] . n. brambilla , a. pineda , j. soto and a. vairo , nucl . b * 566 * ( 2000 ) 275 [ arxiv : hep - ph/9907240 ] . m. beneke , phys . b * 434 * ( 1998 ) 115 [ arxiv : hep - ph/9804241 ] . a. pineda , jhep * 0106 * ( 2001 ) 022 [ arxiv : hep - ph/0105008 ] . t. lee , jhep * 0310 * , 044 ( 2003 ) [ arxiv : hep - ph/0304185 ] . c. contreras , g. cvetic and p. gaete , phys . d * 70 * ( 2004 ) 034008 [ arxiv : hep - th/0311202 ] . i. i. bigi , m. a. shifman , n. g. uraltsev and a. i. vainshtein , phys . d * 50 * ( 1994 ) 2234 [ arxiv : hep - ph/9402360 ] . a. h. hoang , z. ligeti and a. v. manohar , phys . * 82 * , 277 ( 1999 ) [ arxiv : hep - ph/9809423 ] ; phys . d * 59 * , 074017 ( 1999 ) [ arxiv : hep - ph/9811239 ] . o. i. yakovlev and s. groote , phys . d * 63 * , 074012 ( 2001 ) [ arxiv : hep - ph/0008156 ] . t. lee , phys . d * 67 * , 014020 ( 2003 ) [ arxiv : hep - ph/0210032 ] . i. caprini and j. fischer , phys . d * 60 * , 054014 ( 1999 ) [ arxiv : hep - ph/9811367 ] . g. cveti and t. lee , phys . rev . d * 64 * , 014030 ( 2001 ) [ arxiv : hep - ph/0101297 ] ; g. cveti , t. lee and i. schmidt , phys . b * 520 * , 222 ( 2001 ) [ arxiv : hep - ph/0107069 ] ; n. n. khuri , phys . b * 82 * , 83 ( 1979 ) ; c. n. lovett - turner and c. j. maxwell , nucl . b * 452 * , 188 ( 1995 ) [ arxiv : hep - ph/9505224 ] ; p. ball , m. beneke and v. m. braun , nucl . b * 452 * , 563 ( 1995 ) [ arxiv : hep - ph/9502300 ] . g. cveti , phys . d * 67 * , 074022 ( 2003 ) [ arxiv : hep - ph/0211226 ] . m. beneke and v. m. braun , nucl . b * 426 * ( 1994 ) 301 [ arxiv : hep - ph/9402364 ] ; m. beneke , phys . rept . * 317 * ( 1999 ) 1 . m. beneke , phys . b * 344 * , 341 ( 1995 ) [ arxiv : hep - ph/9408380 ] . o. v. tarasov , a. a. vladimirov and a. y. zharkov , phys . b * 93 * , 429 ( 1980 ) ; s. a. larin and j. a. vermaseren , phys . b * 303 * , 334 ( 1993 ) [ arxiv : hep - ph/9302208 ] . t. van ritbergen , j. a. vermaseren and s. a. larin , phys . b * 400 * , 379 ( 1997 ) [ arxiv : hep - ph/9701390 ] . j. r. ellis , i. jack , d. r. jones , m. karliner and m. a. samuel , phys . d * 57 * , 2665 ( 1998 ) [ arxiv : hep - ph/9710302 ] . v. elias , t. steele , f. chrishtie , r. migneron and k. sprague , phys . d * 58 * , 116007 ( 1998 ) [ arxiv : hep - ph/9806324 ] . t. lee , phys . d * 56 * ( 1997 ) 1091 [ arxiv : hep - th/9611010 ] ; t. lee , phys . b * 462 * ( 1999 ) 1 [ arxiv : hep - ph/9908225 ] . g. cveti , j. phys . g * 30 * , 863 ( 2004 ) [ arxiv : hep - ph/0309262 ] . g. cveti , c. dib , t. lee and i. schmidt , phys . d * 64 * , 093016 ( 2001 ) [ arxiv : hep - ph/0106024 ] . s. titard and f. j. yndurain , phys . d * 49 * ( 1994 ) 6007 [ arxiv : hep - ph/9310236 ] ; a. pineda and f. j. yndurain , phys . d * 58 * ( 1998 ) 094022 [ arxiv : hep - ph/9711287 ] ; phys . d * 61 * ( 2000 ) 077505 [ arxiv : hep - ph/9812371 ] . a. czarnecki and k. melnikov , phys . lett . * 80 * , 2531 ( 1998 ) [ arxiv : hep - ph/9712222 ] ; phys . d * 65 * , 051501 ( 2002 ) [ arxiv : hep - ph/0108233 ] ; a. h. hoang and t. teubner , phys . d * 58 * , 114023 ( 1998 ) [ arxiv : hep - ph/9801397 ] ; j. h. khn , a. a. penin and a. a. pivovarov , nucl . b * 534 * , 356 ( 1998 ) [ arxiv : hep - ph/9801356 ] . k. melnikov and a. yelkhovsky , phys . d * 59 * , 114009 ( 1999 ) [ arxiv : hep - ph/9805270 ] . a. a. penin and a. a. pivovarov , phys . b * 435 * , 413 ( 1998 ) [ arxiv : hep - ph/9803363 ] ; nucl . b * 549 * , 217 ( 1999 ) [ arxiv : hep - ph/9807421 ] . m. b. voloshin , nucl . b * 154 * , 365 ( 1979 ) ; h. leutwyler , phys . lett . b * 98 * , 447 ( 1981 ) ; m. b. voloshin , sov . j. nucl . phys . * 36 * , 143 ( 1982 ) [ yad . * 36 * , 247 ( 1982 ) ] . b. l. ioffe and k. n. zyablyuk , eur . j. c * 27 * , 229 ( 2003 ) [ arxiv : hep - ph/0207183 ] . n. brambilla , y. sumino and a. vairo , phys . d * 65 * , 034001 ( 2002 ) [ arxiv : hep - ph/0108084 ] . d. eiras and j. soto , phys . b * 491 * , 101 ( 2000 ) [ arxiv : hep - ph/0005066 ] . a. h. hoang , arxiv : hep - ph/0008102 . b. v. geshkenbein , b. l. ioffe and k. n. zyablyuk , phys . d * 64 * , 093009 ( 2001 ) [ arxiv : hep - ph/0104048 ] . r. barate _ [ aleph collaboration ] , eur . j. c * 4 * , 409 ( 1998 ) . s. bethke , nucl . suppl . * 121 * , 74 ( 2003 ) [ arxiv : hep - ex/0211012 ] . m. beneke and a. signer , phys . b * 471 * , 233 ( 1999 ) [ arxiv : hep - ph/9906475 ] . j. h. khn and m. steinhauser , nucl . b * 619 * , 588 ( 2001 ) [ erratum - ibid . b * 640 * , 415 ( 2002 ) ] [ arxiv : hep - ph/0109084 ] . g. corcella and a. h. hoang , phys . b * 554 * , 133 ( 2003 ) [ arxiv : hep - ph/0212297 ] . m. eidemller , phys . d * 67 * , 113002 ( 2003 ) [ arxiv : hep - ph/0207237 ] . .[table2 ] the separate uncertainties @xmath304 ( in mev ) for the extracted value of @xmath38 from various sources : 1 . ) @xmath25 [ @xmath305 mev ] ; 2 . ) @xmath306 gev ; 3 . ) @xmath307 ; 4 . ) @xmath110 [ @xmath111 ] ; 5 . ) @xmath214 [ @xmath216 ] ; 6 . ) @xmath310 ; 7 . ) @xmath311 ; 8 . @xmath312 ; 9 . @xmath313 ( @xmath314 mev ) . [ cols="^,>,>,>,>,>,>,>,>,>",options="header " , ]	methods of borel integration to calculate the binding ground energies and mass of @xmath0 quarkonia are presented . the methods take into account the leading infrared renormalon structure of the `` soft '' binding energy @xmath1 and of the quark pole masses @xmath2 , and the property that the contributions of these singularities in @xmath3 cancel . the resummation formalisms are applied to quantities which do not involve renormalon ambiguity , such as @xmath4 mass @xmath5 and @xmath6 .
in the minimal supersymmetric extension of the standard model ( mssm ) @xcite , radiative corrections involving supersymmetric ( susy ) particles modify the tree level relation between fermion masses and their yukawa couplings @xcite . in this paper we review the sources and behavior of these susy yukawa corrections and describe their phenomenological effects . the behavior of the susy yukawa corrections is most easily derived in the context of an effective field theory ( eft ) , in which we take the low energy effective theory of the mssm below the susy scale to be a two higgs doublet model ( 2hdm ) and absorb the effects of susy radiative corrections into the parameters of the eft . at tree level , the fermion yukawa couplings and masses arise from the lagrangian , @xmath2 + { \rm h.c . } \ ] ] where we use third generation quark notation and the higgs doublets are @xmath3 the higgs vacuum expectation values ( vevs ) are constrained by @xmath4 and their ratio is parameterized by @xmath5 . the fermion masses arise from replacing the higgs fields with their vevs : @xmath6 so that @xmath7 and @xmath8 . at tree level , down - type fermions receive their masses from @xmath9 while up - type fermions receive their masses from @xmath10 . this is a consequence of the holomorphicity of the superpotential in unbroken susy ; if susy were unbroken , it would be true to all orders in the eft . in a model with more than one higgs doublet , tree - level flavor - changing neutral higgs couplings can be eliminated if the right - handed fermion singlets with each value of hypercharge are allowed to couple to only one higgs doublet @xcite . under this requirement , called `` natural flavor conservation '' , there are two different configurations possible for the couplings of the two higgs doublets @xmath9 and @xmath10 of a 2hdm to quarks : in the type i 2hdm all the quarks couple to @xmath9 and none to @xmath10 , while in the type ii 2hdm the down - type right - handed quark singlets couple to @xmath9 while the up - type right - handed quark singlets couple to @xmath10 . the most general 2hdm without natural flavor conservation , in which the right - handed fermion singlets with each value of hypercharge couple to both higgs doublets , is called the type iii 2hdm . in unbroken susy , the higgs couplings take the form of a type ii 2hdm . however , susy must be broken in nature . when susy is broken , non - holomorphic higgs couplings to fermions arise at one loop in the eft . we first consider down - type fermions . at one loop , diagrams such as those in fig . [ fig : mssmdiags2 ] give rise to a correction to the coupling of @xmath11 quarks to @xmath9 and a non - holomorphic coupling of @xmath11 quarks to @xmath10 : @xmath12 examining eq . [ eq : effectivel ] , we see that in the eft , the right - handed down - type quarks couple to both @xmath9 and @xmath10 . thus,_the low energy effective theory of the mssm is a type iii 2hdm_. the @xmath11 quark mass at one loop is obtained by inserting the higgs vevs into eq . [ eq : effectivel ] : @xmath13 solving for @xmath14 we find , @xmath15 here @xmath16 because of the loop suppression . however , the third term in eq . [ eq : mbcorr ] is enhanced by a factor of @xmath0 , and thus can be quite large at large @xmath0 @xcite . this @xmath0 enhanced correction to @xmath14 has significant phenomenological consequences at large @xmath0 because it modifies the relation between @xmath14 and @xmath17 . at tree level ( so that @xmath18 ) and generate the down - type quark masses at one loop through susy breaking terms . ] at tree level , @xmath19 . at one loop , @xmath17 can be significantly smaller or larger than the tree - level value , depending on the sign of @xmath20 . in the up - type fermion sector , an analogous calculation gives , @xmath21 in particular , there is no @xmath0 enhancement and the corrections to the relation between @xmath22 and @xmath23 are small . these corrections can still be important ; for example , they enter the mssm higgs mass calculation at the two loop level and lead to a few gev change in the upper bound on @xmath24 in some regions of parameter space @xcite . we now examine the behavior of @xmath25 as the susy mass scale is varied . it is easy to see from , e.g. , the first diagram in fig . [ fig : mssmdiags2 ] that @xmath25 are independent of @xmath26 . this diagram scales like @xmath27 , where the factor of @xmath28 comes from a helicity flip in the gluino line , the factor of @xmath29 comes from the dimensionful higgs - squark - squark coupling , and the factor of @xmath30 comes from the three - point loop integral ( here @xmath26 is the largest of the masses in the loop and we neglect external momenta ; see eq . [ eq : iabc ] ) . thus if all the soft susy breaking parameters and the @xmath29 parameter are scaled by a common factor , @xmath25 are unchanged ; i.e. , they do not decouple with @xmath26 when the ratios between susy parameters are held fixed . these susy yukawa corrections are also universal ; i.e. , they are independent of the external momenta and thus process - independent . in general , there are additional contributions to physical processes at one loop that depend on the external momenta ; however , these are suppressed by powers of @xmath31 . although @xmath25 do not decouple with @xmath26 , they do not represent a violation of the decoupling theorem @xcite : in the limit of large pseudoscalar higgs mass @xmath32 , the effects of @xmath25 appear only in the couplings of the heavy higgs bosons , as we will now illustrate . because in the type iii 2hdm the right - handed down - type fermions couple to both higgs doublets , there is no longer a special yukawa basis for the higgs fields . we are then free to write them in a new basis chosen such that one of the doublets has zero vev ( the `` vev basis '' ) : @xmath33 where @xmath34 is the standard model ( sm ) higgs vev and we have performed the rotation ( @xmath35 ) @xmath36 note that in the vev basis @xmath37 contains the goldstone bosons while @xmath38 contains @xmath39 and @xmath40 . the two cp - even neutral higgs mass eigenstates @xmath41 and @xmath42 are given in terms of @xmath43 and @xmath44 by @xmath45 in the limit of large @xmath32 , @xmath46 goes to zero : @xmath47 so @xmath48 . thus in the limit of large @xmath32 , @xmath37 contains the goldstone bosons , the sm higgs vev , and the light higgs boson @xmath41 , while @xmath38 contains all the heavy higgs states , @xmath49 , @xmath40 and @xmath42 . the effective lagrangian in the vev basis can be written as , @xmath50 where @xmath51 and @xmath52 are the yukawa couplings in the vev basis . inserting the higgs vevs we obtain the @xmath11 quark mass : @xmath53 in particular , @xmath54 is fixed by the @xmath11 quark mass , while @xmath52 is unconstrained . thus the coupling of @xmath37 to quarks must be sm - like while that of @xmath38 is unconstrained and can contain the effects of @xmath25 . in the decoupling limit @xcite of large @xmath32 , @xmath48 so the @xmath55 coupling approaches its sm value , while the effects of @xmath25 are confined to the heavy higgs bosons in @xmath38 . we now write down the susy radiative corrections to the down - type mass - yukawa relation ( keeping only the @xmath0 enhanced contributions ) @xcite : @xmath56 where the correction @xmath57 is given by is proportional to the product @xmath58 , it has been suggested in ref . @xcite to use the sign of @xmath57 and the sign of @xmath29 ( measured in some other process ) to determine the sign of @xmath28 and test the anomaly - mediated susy breaking scenario @xcite , which predicts a negative @xmath28 . ] @xmath59 here @xmath60 and @xmath61 is a loop integral , @xmath62/ \nonumber \\ & & [ ( a^2-b^2)(b^2-c^2)(a^2-c^2 ) ] , \label{eq : iabc}\end{aligned}\ ] ] which is positive for any @xmath63 and goes like @xmath64 . solving eq . [ eq : mbyb ] for @xmath17 @xcite , we find @xmath65 where we have not expanded the denominator to one loop order . it was proven in ref . @xcite that eq . [ eq : ybmb ] includes a resummation of terms of order @xmath66 to all orders of perturbation theory . in particular , @xmath57 in eq . [ eq : ybmb ] gets no @xmath0 enhanced corrections at higher orders of the form @xmath67 . the remaining higher order corrections to @xmath57 are not @xmath0 enhanced and are thus insignificant compared to the one loop piece ( eq . [ eq : deltab ] ) . the proof in ref . @xcite that eq . [ eq : ybmb ] receives no higher order @xmath0 enhanced contributions relies on the facts ( 1 ) that @xmath0 enters the calculation at higher orders only multiplied by @xmath14 , e.g. , from the @xmath68 term in the bottom squark mass matrix ; and ( 2 ) that no factors of @xmath69 can arise from the loop diagrams to cancel the @xmath14 factor multiplying @xmath0 , because the yukawa coupling operator is dimension four . expanding eq . [ eq : ybmb ] to one loop order , we have @xmath70 an expression of this form arises in on - shell diagrammatic one - loop calculations , in which @xmath57 enters through the @xmath11 quark mass counterterm . although eqs . [ eq : ybmb ] and [ eq : ybmb1loop ] are equivalent at one loop , they differ at higher orders : eq . [ eq : ybmb1loop ] does not contain a resummation of terms of order @xmath66 to all orders of perturbation theory . , in which the @xmath11 mass operator is renormalized , because the @xmath11 mass operator is not dimension four . see ref . @xcite for details . ] in our eft analyses in the remainder of this paper we will use eq . [ eq : ybmb ] in order to take advantage of the resummation of higher order terms . the remainder of this paper is organized as follows . we examine the effect of the susy yukawa corrections on the process @xmath71 in sec . [ sec : hbb ] and on the @xmath72 coupling in sec . [ sec : h+tb ] . in sec . [ sec : fcnceft ] we describe how flavor - changing neutral higgs interactions arise in the eft from the susy yukawa corrections , and in sec . [ sec : fcncproc ] we summarize recent results for specific flavor changing processes . in sec . [ sec : ckm ] we discuss the renormalization of the ckm matrix . in sec . [ sec : resum ] we briefly discuss resummation of the susy yukawa corrections to flavor changing processes . finally in sec . [ sec : conclusions ] we give a summary and outlook . accurate knowledge of the @xmath71 branching ratio is important for determining the reach of the upcoming higgs searches at the tevatron and lhc @xcite . also , once a light higgs boson is discovered , precision measurements of its branching ratios ( e.g. , at a future @xmath74 linear collider ) can be used to distinguish a sm higgs boson from a mssm higgs boson in some regions of parameter space @xcite . the susy yukawa corrections can have a significant effect on these branching ratios @xcite . for example , the ratio @xmath75 is sensitive to the susy yukawa corrections to @xmath71 . coupling also receives @xmath0 enhanced susy yukawa corrections , but they are proportional to electroweak gauge couplings and are expected to be much smaller than the corrections to the @xmath55 coupling . ] at tree level , the ratio @xmath75 is the same in the sm and the mssm ( @xmath76 ) . further , in the mssm the @xmath55 and @xmath77 couplings have the same dependence on the cp - even higgs mixing angle @xmath78 , so that @xmath75 is also insensitive to the radiative corrections to @xmath78 . thus in the context of the mssm , a deviation of @xmath75 from its sm value provides a direct window onto the susy yukawa corrections @xcite . we begin by computing the susy corrections to @xmath71 in the eft approach ; that is , we assume @xmath79 and neglect the external momenta . in the next section . ] using eqs . [ eq : effectivel ] and [ eq : ybmb ] , the @xmath55 coupling is given by @xcite @xmath80 \label{eq : ghbbeft}\end{aligned}\ ] ] where the tree level coupling is @xmath81 and @xmath82 . as shown in sec . [ sec : intro ] , the susy radiative corrections to @xmath83 should decouple in the limit of large @xmath32 . indeed , we have , @xmath84 so that @xmath83 approaches its sm value as expected in the decoupling limit @xcite . up to this point we have neglected the effects of the external momenta on @xmath71 . these give contributions of order @xmath85 , where @xmath86 . these contributions can be sizeable because of the @xmath0 enhancement , especially when the susy scale is relatively low . for example , for @xmath87 , @xmath88 is of order one even for @xmath89 tev . in order to treat the external momenta correctly , we must do an on - shell diagrammatic calculation . we focus here on the susy - qcd contributions to @xmath83 , for which diagrammatic calculations are available in refs . enhanced contributions . ] for compactness we give here an approximate formula expanded in powers of @xmath90 and @xmath91 , where we have taken @xmath92 and @xmath93 . defining @xmath94 $ ] , we have @xcite @xmath95 the first term in eq . [ eq : dg / g ] was found before in the eft calculation ; it remains constant as @xmath96 at fixed @xmath32 . the remaining terms are neglected in the eft approach ; they decouple as @xmath96 but remain constant as @xmath97 at fixed @xmath26 . because of their @xmath30 dependence they can be identified as coming from higher dimensional operators that were neglected in our eft calculation . the correction @xmath98 decouples in the limit that both @xmath32 and @xmath96 . the susy - electroweak corrections to @xmath71 are of order @xmath99 instead of order @xmath100 and arise from the second diagram of fig . [ fig : mssmdiags2 ] . the susy - electroweak yukawa corrections are easily calculated in the eft approach ( see eq . [ eq : deltab ] ) . there are also large corrections to @xmath41 decays at order @xmath99 from the radiative corrections to the cp - even higgs mixing angle @xmath78 @xcite . these corrections are usually taken into account by inserting the radiatively corrected value of @xmath78 into the tree level higgs couplings to fermions . at tree level , the off - diagonal elements of the cp - even higgs mass matrix are @xmath101 @xcite . when @xmath32 is small and @xmath0 is large , @xmath102 can be quite small , of the same order as its one loop radiative corrections . then it is possible to tune the radiatively corrected @xmath103 to zero by a careful choice of the mssm parameters , thereby driving the mixing angle @xmath78 to zero . if @xmath104 , the @xmath55 and @xmath77 couplings both vanish at tree level . however , the susy - qcd yukawa corrections to the @xmath55 coupling are nonzero , because they depend on the @xmath41 couplings to bottom squarks which remain finite at @xmath104 . thus at @xmath104 , the @xmath77 coupling vanishes while the @xmath55 coupling is nonzero due to the susy - qcd yukawa corrections . if we then vary the supersymmetric parameters , we can tune @xmath78 so that the tree level value of the @xmath55 coupling cancels its susy - qcd yukawa corrections . then the corrected @xmath55 coupling vanishes while the @xmath77 coupling is again nonzero . thus the @xmath55 and @xmath77 couplings vanish at different values of @xmath78 , or equivalently , at different points in susy parameter space . this behavior was first pointed out in ref . @xcite . of course , an on - shell diagrammatic calculation is needed to take into account terms of order @xmath85 from the external momenta . in ref . @xcite a partial on - shell diagrammatic calculation was performed , which includes an on - shell calculation of the corrections to @xmath78 and the on - shell susy - qcd corrections to @xmath71 , but does not include the @xmath105 susy - electroweak yukawa corrections to @xmath71 . a full one - loop diagrammatic calculation of the susy - electroweak corrections to @xmath71 was performed in ref . @xcite , although compact explicit expressions are not available . a similar calculation is in progress @xcite . the @xmath0 enhanced susy yukawa corrections modify the @xmath72 coupling through , e.g. , the first diagram of fig . [ fig : mssmdiags2 ] with @xmath107 ( @xmath108 ) replaced by @xmath109 ( @xmath110 ) and taking the charged higgs states in @xmath9 , @xmath111 . these corrections have important implications for the charged higgs boson search through @xmath112 at the tevatron @xcite because the reach in @xmath113 and @xmath0 depends on the effective @xmath114 and @xmath11 yukawa couplings . the one loop susy yukawa corrections to @xmath112 can be very large , of order the tree level amplitude , so that reliable calculations require a resummation or improvement . using the eft approach , we have at large @xmath0 @xcite @xmath115 with @xmath57 as in eq . [ eq : deltab ] . as described in sec . [ sec : intro ] , in this result @xmath57 gets no @xmath0 enhanced corrections at higher orders of the form @xmath116 @xcite . the remaining higher order corrections to @xmath57 are not @xmath0 enhanced . the susy yukawa corrections to the @xmath72 coupling have also been computed diagrammatically in refs . @xcite ( susy - qcd ) and @xcite ( susy - electroweak ) , giving results of the form @xmath117 this is equivalent to eq . [ eq : h+tbresum ] at one loop level , but in eq . [ eq : h+tbnoresum ] @xmath57 receives @xmath0 enhanced radiative corrections at higher orders . in ref . @xcite the tree - level input parameters were carefully chosen to keep the one - loop corrections in eq . [ eq : h+tbnoresum ] small . this is analogous to using running quark masses to absorb qcd corrections ; however it does not take into account the higher order @xmath0 enhanced susy corrections . the @xmath0 enhanced susy yukawa corrections lead to flavor changing neutral higgs interactions in the eft . this was first pointed out in ref . @xcite for @xmath1-@xmath118 mixing and expanded upon in ref . @xcite for leptonic @xmath1 meson decays . similar results have been found in diagrammatic calculations for a number of processes ; these will be discussed in sec . [ sec : fcncproc ] . in this section we describe how flavor changing neutral higgs interactions arise in the eft . as shown in sec . [ sec : intro ] , for @xmath119 the low energy effective theory of the mssm is the type iii 2hdm . in the type iii 2hdm , there is no natural flavor conservation and flavor changing neutral higgs couplings arise in general @xcite . in three generation notation , eq . [ eq : leffvevbasis ] becomes @xmath120 where we again use the vev basis , eq . [ eq : vevbasis ] . here @xmath121 are fermion generation indices and @xmath122 are general yukawa matrices , not necessarily diagonal . the down - type quark mass matrix is obtained by replacing the higgs fields with their vevs : @xmath123 note that @xmath124 does not appear in eq . [ eq:3genmass ] because @xmath38 has zero vev . diagonalizing the mass matrix in eq . [ eq:3genmass ] diagonalizes the yukawa matrix @xmath125 , which parameterizes the down - type fermion couplings to @xmath37 . the yukawa matrix @xmath126 , which parameterizes the down - type fermion couplings to @xmath38 , is not in general diagonal in the fermion mass basis . rewriting eq . [ eq:3genyuks ] in the fermion mass basis , we have @xmath127 clearly , the couplings of @xmath37 to down - type fermions are flavor diagonal , while those of @xmath38 are not . the off - diagonal elements in @xmath126 give flavor changing couplings of the states in @xmath128 ( @xmath40 and @xmath44 ) to down - type fermions . note that the flavor changing effects of the susy yukawa corrections in physical processes should decouple at large @xmath32 , for which @xmath129 and @xmath130 . there are two potential sources of flavor changing neutral higgs couplings in the mssm : minimal and non - minimal flavor violation . in the mssm with minimal flavor violation , the sfermion mass matrices are flavor - diagonal in the same basis as the quark and lepton mass matrices . then the only source of flavor changing is the ckm matrix , as in the sm . in this case the flavor changing neutral higgs couplings arise from one loop diagrams that contain a generation - changing @xmath131 , @xmath49 , or chargino coupling and an associated ckm matrix element . non - minimal flavor violation occurs when the sfermion mass matrices are not flavor - diagonal in the same basis as the quark and lepton mass matrices . when the sfermion mass matrices are diagonalized , flavor changing gluino - squark - quark and neutralino - sfermion - fermion couplings arise . non - minimal flavor violation is present in the most general mssm , and can lead to large flavor - changing effects in contradiction to experiment in , e.g. , @xmath132-@xmath133 mixing . this is the susy flavor problem , which must be solved in any realistic mssm scenario . attempts to solve the susy flavor problem include flavor - blind susy breaking scenarios ( e.g. , minimal supergravity ) , in which the sfermion mass matrices are flavor diagonal in the same basis as the quark and lepton mass matrices at the susy - breaking scale . however , a small amount of non - minimal flavor violation is typically regenerated as the sfermion and fermion mass matrices are run down to the electroweak scale . if non - minimal flavor violation is present in the mssm , then the flavor changing neutral higgs couplings receive contributions from one loop diagrams that contain a flavor - changing gluino or neutralino coupling , in addition to the contributions present in the minimal flavor violation scenario . the susy contributions to @xmath134 are well known @xcite . in particular , the chargino - top squark contribution to the amplitude is proportional to @xmath0 . it is possible to obey the experimental constraint on @xmath134 and still have large flavor changing effects in other processes from the chargino - top squark loop by choosing heavy enough chargino and top squark masses @xcite . recently in ref . @xcite the higher order @xmath0 enhanced contributions to @xmath134 were resummed , following the formalism of ref . these higher order contributions can be large and can modify the regions of mssm parameter space excluded by @xmath134 @xcite . because the sm amplitude for @xmath136 is helicity suppressed , contributions from neutral higgs boson exchange in the 2hdm or mssm can compete with the sm amplitude at large @xmath0 . in the mssm with large @xmath0 , the dominant contribution to the amplitude comes from @xmath41 , @xmath42 and @xmath40 exchange with a flavor changing susy bubble in the external @xmath137 quark line , and grows with @xmath138 @xcite . in the mssm with minimal flavor violation the flavor changing susy bubble is a chargino - top squark loop ; if non - minimal flavor violation is present , additional contributions come from a gluino - bottom squark and neutralino - bottom squark loop . all other one - loop susy contributions have at most a @xmath139 enhancement . in the type ii 2hdm with large @xmath0 , penguin and box diagrams give a contribution to the amplitude that grows with @xmath139 and is of the same order as the sm contribution @xcite . in the mssm , @xmath135 can be enhanced by three orders of magnitude over the sm rate because of the @xmath138 enhancement in the amplitude . this has been demonstrated in an eft calculation in ref . @xcite and in diagrammatic calculations in refs . the experimental sensitivity is best in the channel @xmath140 ; here the sm predicts @xmath141 @xcite , while the current experimental bound is @xmath142 ( 95% cl ) @xcite . at large @xmath0 and low @xmath32 , there are regions of susy parameter space in which @xmath143 can exceed its current upper bound @xcite ; thus susy parameter space is already being probed by this decay . future experiments at the tevatron run ii should improve the experimental bound by a factor of 40 with 2 fb@xmath144 of data @xcite , and further at an extended run ii . at the lhc this decay is expected to be observed at the sm rate after three years of running at low luminosity @xcite . @xmath1-@xmath118 mixing has been analyzed in the context of the effective type iii 2hdm in refs . @xcite . in the eft , a tree - level neutral higgs boson exchange diagram with flavor changing couplings at each vertex contributes to @xmath1-@xmath118 mixing . if the tree - level relations for mssm higgs boson masses and couplings are used , then the three tree - level eft diagrams involving @xmath41 , @xmath42 and @xmath40 exchange sum to zero @xcite . however , the tree - level relations for the mssm higgs boson masses and couplings are in general a poor approximation , and when radiative corrections to these masses and couplings are included the neutral higgs mediated contributions to @xmath1-@xmath118 mixing no longer cancel . in the sm , the radiative corrections to the ckm matrix are very small electroweak effects . in the mssm , however , there are @xmath0 enhanced flavor changing loop diagrams that may yield significant corrections to the ckm matrix . in ref . @xcite the down - type quark mass matrix was run down from a high scale and the @xmath0 enhanced susy corrections were included at the scale where the susy particles were integrated out ; diagonalizing the quark mass matrices then yielded the radiatively corrected ckm matrix . here we describe the source of the susy corrections to the ckm matrix in the eft . when we diagonalized the down - type quark mass matrix in eq . [ eq:3genmass ] , the one loop susy corrections were implicitly included . writing the one loop corrections explicitly , we have @xmath145 in the vev basis this becomes , @xmath146 comparing this to eq . [ eq:3genyuks ] , @xmath125 is given by @xmath147 diagonalizing the down - type quark mass matrix diagonalizes @xmath125 . since @xmath148 are not in general diagonal in the same basis as the tree - level yukawa matrix @xmath149 , they yield an additional rotation of the mass matrix . this additional rotation is a correction to the ckm matrix . at present , the ckm elements are not predicted by theory and must be input from experiment . if the ckm elements are measured solely through @xmath150 couplings to quarks , the susy corrections can simply be absorbed into the bare elements and are undetectable . however , if the ckm elements derived from @xmath150-quark - quark , @xmath49-quark - quark , and especially chargino - squark - quark couplings can be compared , the susy corrections may become apparent . we expect the chargino - squark - quark couplings to receive large corrections relative to the @xmath150-quark - quark couplings because susy - breaking effects will lead to different results for the flavor changing susy bubble diagrams in the external down - type quark and down - type squark lines . the susy radiative corrections that lead to flavor changing neutral higgs couplings are enhanced by @xmath0 and can be quite large , as we have seen . as in the flavor conserving case , these @xmath0 enhanced corrections should be resummed . such a resummation has been performed recently in ref . @xcite for higher - order corrections to @xmath134 , following the results in ref . @xcite . however , it is unclear whether this procedure will work in general because it is unclear whether the proof in ref . @xcite that all higher order @xmath0 enhanced terms are automatically resummed if the relation in eq . [ eq : ybmb ] is used is applicable to the flavor changing case . for example , factors of @xmath151 or @xmath152 can in principle arise at higher orders . this issue should be clarified before the resummation procedure is applied to additional flavor changing processes . it is clear that the @xmath0 enhanced susy corrections to the relation between down - type fermion masses and their yukawa couplings has the potential to be a good handle on the mssm at large @xmath0 . these corrections lead to large effects in both higgs physics and @xmath1 physics . because these effects can be quite large , reliable calculations are needed . in higgs boson decays , the external momenta are on the order of the electroweak scale and their effects should be included through diagrammatic calculations ; these are available for the susy - qcd corrections to @xmath71 and calculations are in progress for the susy - electroweak corrections to both @xmath71 and @xmath153 . in addition , resummation of higher order @xmath0 enhanced contributions beyond the one loop level is very important in some regions of parameter space . this resummation is well known and straightforward in the eft approach but should also be incorporated into the diagrammatic calculations . in flavor changing processes such as rare @xmath1 meson decays and ckm matrix renormalization , it is not entirely clear how the resummation should be done . finally , the susy corrections to the ckm matrix may have a significant effect on the @xmath49 and chargino couplings to ( s)fermions relative to the @xmath131 couplings . in short , the @xmath0 enhanced susy corrections to the relation between down - type fermion masses and their yukawa couplings offer a wide range of challenges and opportunities for mssm phenomenology . i am grateful to the members of the university of minnesota theoretical physics institute for organizing an excellent symposium and workshop where this work was presented , and to the other workshop participants for many lively discussions . i also thank u. nierste and t. plehn for fruitful discussions . fermilab is operated by universities research association inc . under contract de - ac02 - 76ch03000 with the u.s . department of energy . h. e. haber , g. l. kane , phys . * 117 * , 75 ( 1985 ) . l. j. hall , r. rattazzi , u. sarid , phys . * d50 * , 7048 ( 1994 ) ; r. hempfling , phys . rev . * d49 * , 6168 ( 1994 ) ; m. carena , m. olechowski , s. pokorski , c. e. m. wagner , nucl . * b426 * , 269 ( 1994 ) . j. f. gunion , h. e. haber , nucl . * b272 * , 1 ( 1986 ) ; * b278 * , 449 ( 1986 ) [ e : * b402 * , 567 ( 1993 ) ] ; j. f. gunion , h. e. haber , g. kane , s. dawson , _ the higgs hunter s guide _ ( addison - wesley , reading , ma , 1990 ) [ e : hep - ph/9302272 ] . s. l. glashow , s. weinberg , phys . rev . * d15 * , 1958 ( 1977 ) ; e. a. paschos , phys . * d15 * , 1966 ( 1977 ) . d. atwood , l. reina , a. soni , phys . rev . * d55 * , 3156 ( 1997 ) . m. carena , s. mrenna , c. e. m. wagner , phys . rev . * d60 * , 075010 ( 1999 ) . f. borzumati , g. r. farrar , n. polonsky , s. thomas , nucl . b555 * , 53 ( 1999 ) . m. carena , h. e. haber , s. heinemeyer , w. hollik , c. e. m. wagner , g. weiglein , nucl . * b580 * , 29 ( 2000 ) . t. appelquist , j. carazzone , phys . rev . * d11 * , 2856 ( 1975 ) . h. e. haber , y. nir , nucl . phys . * b335 * , 363 ( 1990 ) ; h. e. haber , in the proceedings of the workshop on electroweak symmetry breaking , budapest , hungary , july 11 - 13 , 1994 , p. 1 - 16 [ hep - ph/9501320 ] . d. m. pierce , j. a. bagger , k. matchev , r. j. zhang , nucl . * b491 * , 3 ( 1997 ) . g. d. kribs , phys . rev . * d62 * , 015008 ( 2000 ) . l. randall , r. sundrum , nucl . phys . * b557 * , 79 ( 1999 ) ; g. f. giudice , m. a. luty , h. murayama , r. rattazzi , jhep * 9812 * , 027 ( 1998 ) . m. carena , d. garcia , u. nierste , c. e. m. wagner , nucl . b577 * , 88 ( 2000 ) . m. carena , s. mrenna , c. e. m. wagner , phys . rev . * d62 * , 055008 ( 2000 ) . h. e. logan , hep - ph/0012202 ; m. carena , h. e. haber , h. e. logan , s. mrenna , fermilab - pub-00/334-t , in preparation . k. s. babu , c. kolda , phys . lett . * b451 * , 77 ( 1999 ) . h. e. haber , m. j. herrero , h. e. logan , s. pearanda , s. rigolin , d. temes , phys . * d63 * , 055004 ( 2001 ) . a. dabelstein , nucl . phys . * b456 * , 25 ( 1995 ) . j. a. coarasa , r. a. jimnez , j. sol , phys . b389 * , 312 ( 1996 ) . g. l. kane , g. d. kribs , s. p. martin , j. d. wells , phys . * d53 * , 213 ( 1996 ) ; w. loinaz , j. d. wells , phys . * b445 * , 178 ( 1998 ) ; h. baer , j. d. wells , phys . rev . * d57 * , 4446 ( 1998 ) . s. heinemeyer , w. hollik , g. weiglein , eur . j. * c16 * , 139 ( 2000 ) . h. e. haber , m. j. herrero , h. e. logan , s. pearanda , d. temes , in preparation . f. abe et al . [ cdf collaboration ] , phys . lett . * 79 * , 357 ( 1997 ) ; b. abbot et al . [ d0 collaboration ] , phys . 82 * , 4975 ( 1999 ) . j. guasch , j. sol , phys . b416 * , 353 ( 1998 ) . r. a. jimnez , j. sol , phys . lett . * b389 * , 53 ( 1996 ) . h. eberl , k. hidaka , s. kraml , w. majerotto , y. yamada , phys . rev . * d62 * , 055006 ( 2000 ) . j. a. coarasa , d. garcia , j. guasch , r. a. jimnez , j. sol , eur . j. * c2 * , 373 ( 1998 ) ; phys . * b425 * , 329 ( 1998 ) . c. hamzaoui , m. pospelov , m. toharia , phys . rev . * d59 * , 095005 ( 1999 ) . k. s. babu , c. kolda , phys . lett . * 84 * , 228 ( 2000 ) . s. bertolini , f. borzumati , a. masiero , g. ridolfi , nucl . * b353 * , 591 ( 1991 ) ; r. barbieri , g. f. giudice , phys . * b309 * , 86 ( 1993 ) ; n. oshimo , nucl . * b404 * , 20 ( 1993 ) ; m. a. diaz , phys . * b304 * , 278 ( 1993 ) ; y. okada , phys . * b315 * , 119 ( 1993 ) ; r. garisto , j. n. ng , phys . * b315 * , 372 ( 1993 ) ; f. m. borzumati , z. phys . * c63 * , 291 ( 1994 ) ; v. barger , m. s. berger , p. ohmann , r. j. phillips , phys . * d51 * , 2438 ( 1995 ) . h. chankowski , . sawianowska , hep - ph/0008046 . g. degrassi , p. gambino , g. f. giudice , jhep * 0012 * , 009 ( 2000 ) ; m. carena , d. garcia , u. nierste , c. e. m. wagner , hep - ph/0010003 . huang , w. liao , q .- s . yan , s .- h . zhu , hep - ph/0006250 . x. g. he , t. d. nguyen , r. r. volkas , phys . * d38 * , 814 ( 1988 ) ; w. skiba , j. kalinowski , nucl . phys . * b404 * , 3 ( 1993 ) ; y .- b . dai , c .- s . huang , h .- w . huang , phys . * b390 * , 257 ( 1997 ) . m. j. savage , phys . lett . * b266 * , 135 ( 1991 ) . h. e. logan , u. nierste , nucl . phys . * b586 * , 39 ( 2000 ) . f. m. borzumati , c .- s . huang , c. kolda , h. e. logan , in preparation , to be published in the proceedings of the uk phenomenology workshop on heavy flavour and cp violation , durham , uk , 17 - 22 september 2000 .	-6.5 cm fermilab - conf-01/010-t 0.25 cm hep - ph/0102029 5.7 cm in the minimal supersymmetric extension of the standard model ( mssm ) , fermion masses and yukawa couplings receive radiative corrections at one loop from diagrams involving the supersymmetric particles . the corrections to the relation between down - type fermion masses and yukawa couplings are enhanced by @xmath0 , which makes them potentially very significant at large @xmath0 . these corrections affect a wide range of processes in the mssm , including neutral and charged higgs phenomenology , rare @xmath1 meson decays , and renormalization of the ckm matrix . we give a pedagogical review of the sources and phenomenological effects of these corrections .
one of the highly active areas of investigation of hamiltonian nonlinear systems over the past decade has been the examination of nonlinear dynamical lattices of the discrete nonlinear schrdinger ( dnls ) type . chiefly , this development has arisen due to the multitude of applications of pertinent models that have emerged in areas such as nonlinear optics and atomic physics . more specifically , in the optical context , the setting of fabricated algaas waveguide arrays @xcite has been one of the most prototypical ones for the application of dnls models . there , the interplay of discreteness and nonlinearity revealed many interesting features including peierls - nabarro potential barriers , diffraction and diffraction management @xcite , and gap solitons @xcite , among others ; see also the reviews @xcite and references therein . another recent development , which also promoted the analysis of discrete systems in connection with nonlinear optics was the proposal @xcite and creation @xcite of optically induced photonic lattices in photorefractive crystals such as sbn . this paved the way for the observation of a large set of exciting nonlinear wave related phenomena in such crystals . as a representative subset , we mention here the formation of patterns such as dipole @xcite , quadrupole @xcite and necklace @xcite solitary waves , impurity modes @xcite , discrete vortices @xcite , rotary waves @xcite , higher order bloch modes @xcite and gap vortices @xcite , two - dimensional ( 2d ) bloch oscillations and landau - zener tunneling @xcite , wave formation in honeycomb @xcite , hexagonal @xcite and quasi - crystalline lattices @xcite , and recently the study of anderson localization in disordered photonic lattices @xcite . although this setting is mostly studied in the continuum context with a periodic potential ( and sometimes in the presence of the inherent crystal anisotropy ) , it has also spurred a number of studies in the dnls context with the saturable photorefractive nonlinearity @xcite . lastly , another physical realization of such nonlinear dynamical lattices arose over the past few years in atomic physics through the examination of bose - einstein condensates ( becs ) trapped in periodic potentials . there , once again , a reduction of the relevant model can be formulated in the tight - binding approximation within the mean - field limit , reducing the so - called gross - pitaevskii equation with a periodic potential to a genuinely discrete nonlinear schrdinger equation @xcite . in both the nonlinear optical and in the atomic physics setting discussed above , multi - component systems were also examined in recent investigations . more specifically , the first observations of discrete vector solitons in optical waveguide arrays were reported in @xcite , the emergence of multipole patterns in vector photorefractive crystals was presented in @xcite , while numerous experiments with becs were directed towards studies of mixtures of different spin states of @xmath0rb @xcite or @xmath1na @xcite and even ones of different atomic species such as @xmath2k@xmath0rb @xcite and @xmath3li@xmath4cs @xcite . it should be noted that while the above bec experiments did not include the presence of an optical lattice , the addition of such an external optical potential is certainly feasible within the present experimental capabilities @xcite . our aim in the present work is to propose and analyze a family of solutions particular to multicomponent ( in particular , binary , although more - component generalizations are certainly possible ) systems of dnls equations . we dub these proposed solutions `` interlaced '' discrete solitons and vortices , a name stemming from the feature that the profiles of the modes in the two interacting components will have a vanishing intersection of excited sites in the extreme discrete limit of zero coupling between adjacent nodes of the lattice . in these structures , the first component will be excited where the second component is not and vice - versa . in the one - dimensional case , we show how to interlace in a stable fashion simple , as well as more elaborate , bound states of the system @xcite . for such solutions , we consider their existence and stability properties also from an analytical point of view , using as a starting point the anti - continuum limit ( of no - coupling between the sites ) . then we generalize our considerations to higher dimensional settings , showcasing the potentially stable interlacing of more elaborate structures , such as discrete vortices @xcite ( but also of vortices with non - vortical structures ) . we present detailed stability diagrams of such interlaced structures , and also examine their dynamics when they are found to be unstable . our presentation is structured as follows . in section ii , we present the model and general mathematical setup . in section iii , we illustrate both analytically and numerically the properties of such structures in 1d settings . in section iv , we generalize these considerations to a numerical investigation of higher dimensional settings . finally , in section v , we summarize our findings and present our conclusions . we consider a set of coupled dnls equations @xmath5 where @xmath6 is a @xmath7-dimensional index and @xmath8 is the discrete laplacian in @xmath7 dimensions . we look for stationary solutions @xmath9 , @xmath10 through the relation @xmath11 the dynamical equations ( [ dyn1 ] ) then transform into @xmath12 the stability is determined in a frame rotating with frequency @xmath13 for @xmath14 and @xmath15 for @xmath16 , i.e. , we suppose that @xmath17,\qquad v_n(t)=\exp(i\lambda_2t)[v_n+\xi^{(2)}_n(t)].\ ] ] the small perturbations @xmath18 , with @xmath19 can be expressed as @xmath20 leading to the linear stability equations @xmath21 with @xmath22 soliton and vortex solutions are calculated using methods based on the anti - continuous limit . upon calculating these solutions at @xmath23 , we continue them to finite coupling by varying @xmath24 or other parameters ( such as the interspecies nonlinearity strength @xmath25 ) . we are interested in interlaced solitons ( iss ) in 1d lattices and interlaced vortices ( ivs ) in 2d and 3d lattices . the excited sites at @xmath23 are equal to @xmath26 and @xmath27 , except for a phase factor @xmath28 , while @xmath29 at the corresponding excited site . these values are @xmath30 in what follows , we choose @xmath31 and @xmath32 . we also choose @xmath33 as , for @xmath34 interlaced solitons and vortices are unstable for every value of @xmath24 . we consider interlaced solitons which are labeled by @xmath35 , where @xmath36 . this number indicates the `` order '' of the excited state at the anti - continuous limit , whose phase @xmath37 is chosen so that the isolated solitons ( i.e. when @xmath38 ) are stable for any small @xmath24 . for instance , the ground state @xmath39 means @xmath40 and the first excited state @xmath41 will be taken to mean @xmath42 at the ac limit . thus , the state @xmath43 corresponds to @xmath40 , @xmath44 and @xmath45 to @xmath42 , @xmath46 . we first analyze the @xmath43 state , which is stable for @xmath47 . at @xmath48 the iss become unstable through hopf bifurcations ( the value of @xmath49 differs as a function of the rest of the system parameters such as @xmath25 , however the above scenario is robust ) . cascades of this type of bifurcations arise as @xmath24 increases and , when , @xmath50 , the iss become also exponentially unstable . there is a special region for @xmath51 $ ] where the system experiences an inverse hopf bifurcation recovering the stability in a window . the system becomes unstable again through hopf bifurcations for @xmath52 $ ] and exponential instabilities for @xmath53 $ ] . besides , for @xmath54 $ ] there exist windows with only exponential instabilities . fig . [ fig1 ] illustrates all of the above features , by showcasing a typical example of the @xmath43 state , a typical continuation of its principal linearization eigenfrequencies @xmath55 , and a full two - parameter diagram of the stability of this state in the two - parameter plane @xmath56 . for @xmath45 states , the scenario is essentially similar to the @xmath43 case , although , in essence , it is considerably simpler due to the absence of any inverse hopf bifurcations and restabilization windows . [ fig2 ] shows the corresponding features for @xmath45 , as fig . [ fig1 ] for the @xmath43 case . [ cols="^,^,^ " , ] in the present work , we have illustrated the possibility to successfully interlace structures which are stable in each one of the components ( either simple ones , such as single site solitary waves , or more elaborate ones , such as bound states and vortices ) in order to produce stable multi - component interlaced solitons / vortices . we have continued the resulting structures from the anti - continuum limit of no inter - site coupling to finite coupling and illustrated the intervals of stability , as well as the ones of both exponential and oscillatory ( hopf ) instabilities . we have given detailed two - parameter diagrams of the stable ranges of the solutions as a function of the inter - site and inter - component couplings . these revealed that the linear stability of the interlaced structures necessitates sufficiently weak coupling ( typically no larger than 0.4 , with the relevant range decreasing as the inter - component interaction is increased ) and sufficiently weak inter - component interaction ( i.e. , @xmath57 ) . finally , we examined the dynamical evolution of the instability of such interlaced structures , which typically resulted in the destruction of the waveforms , in favor of simpler , more stable dynamical patterns . nevertheless , there is still a number of important open questions for future consideration . for instance , it would be particularly interesting to examine whether it would be possible for the inter - component coupling to actually stabilize structures that are dynamically unstable in the single - component setting . also , it would be useful to possess a systematic classification of the solutions ( interlaced and non - interlaced ones ) available in the multi - component system setting , similarly to the one - component classifications of @xcite . such efforts are currently underway and will be reported in future publications . r. morandotti , u. peschel , j. s. aitchison , h. s. eisenberg and y. silberberg , phys . lett . * 83 * , 2726 - 2729 ( 1999 ) ; h. s. eisenberg , y. silberberg , r. morandotti and j. s. aitchison , phys . lett . * 85 * , 1863 ( 2000 ) . fitrakis , p.g . kevrekidis , h. susanto and d.j . frantzeskakis , phys . e * 75 * , 066608 ( 2007 ) ; v.m . rothos , h.e . nistazakis , p.g . kevrekidis and d.j . frantzeskakis , j. phys . a : math . theor . * 42 * , 025207 ( 2009 ) .	in the present work , we propose a new set of coherent structures that arise in nonlinear dynamical lattices with more than one components , namely interlaced solitons . these are waveforms in which in the relevant anti - continuum limit , i.e. when the sites are uncoupled , one component has support where the other component does not . we illustrate systematically how one can combine dynamically stable unary patterns to create ones such for the binary case of two - components . in the one - dimensional setting , we provide also a detailed theoretical analysis of the existence and stability of these waveforms , while in higher dimensions , where such analytical computations are far more involved , we resort to corresponding numerical computations . lastly , we perform direct numerical simulations to showcase how these structures break up , when exponentially or oscillatorily unstable , to structures with a smaller number of participating sites .
there is now substantial observational evidence@xcite that favors the existence of a smooth exotic cosmic component of energy of negative pressure . going at times under the name of a cosmological constant or quintessence or , at other times , dark energy , which we will adopt here , its true nature remains obscure . the unexpected faintness of high redshift type ia supernovae ( sne ia)suggests that the universe is accelerating today@xcite , relentlessly driven by dark energy . when the sne ia results are combined with observations of the amplitudes of primordial fluctuations in the cosmic microwave background radiation the overall picture seems to be one of an accelerating flat universe . since the standard flat universe , despite its well - known shortcomings , has long been favored on aesthetic and theoretical grounds@xcite , the hope has arisen that the injection of dark energy will cure its ills , particularly in regards of its age of the universe problem . thus a major industry of investigating the constraints imposed by continuously updated astrophysical observations on the dark energy in refined versions of the standard model has flourished in recent times . the present paper is one more contribution in this direction . when it is assumed that the dark energy , viewed in general to be time - dependent , does not interact with matter , the energy equations for nonrelativistic pressureless matter and dark energy decouple leading to conservation of matter and to the dark energy equation @xmath0,where @xmath1 and @xmath2 are the dark energy density and pressure respectively and @xmath3 is the redshift . in this case a solvable cosmological model is obtained if a specific variation of @xmath4 is invoked@xciteor a definite parameterization for the equation of state @xmath5 is suggested @xcite . alternatively if one assumes that the dark energy interacts with matter@xcite , the energy equations for both are coupled and one needs a definite variation for @xmath4,in addition to specifying its equation of state . in this case the conservation of the energy - momentum tensor of the field equations , which holds when matter and the dark energy are noninteracting , is replaced by the conservation of the sum of this tensor and an extra appended tensorial piece representing the time - dependent dark energy . here we follow this line : specifically we assume ( a ) @xmath6,where @xmath7 is the trace of the energy - momentum tensor of cosmic matter of density @xmath8 and pressure @xmath9,and ( b ) a one parameter form for @xmath10 . a variation @xmath11 was introduced by @xcite for the cosmological constant @xmath12,the motivation being to identify the cosmological constant with a lorentz - invariant scalar representing a form of quintessence . this cosmology is reminiscent of similar earlier attempts at identifying @xmath12 with the ricci scalar@xcite . the postulate @xmath13 is interesting because it implies that the cosmological constant vanishes in the radiation - dominated cosmic era of flat cosmology so that the successful standard primordial nucleosynthesis predictions are unaltered . in the matter - dominated era the postulate reduces to @xmath14 where @xmath15 is hubble s parameter . the cosmological constant variation @xmath16 itself was widely discussed in the literature @xcite . in particular @xcite have pointed out that it follows from dimensional arguments consistent with quantum gravity . since such arguments do not depend on the cosmological constant equation of state @xmath17 it is legitimate to regard them as equally valid for dark energy with @xmath18 . extending this postulate to a dark energy with an equation of state of negative pressure we take @xmath19 where @xmath20 is a dimensionless constant . a consequence of this is that the matter density parameter @xmath21 is constant in the model . we take it to be @xmath22 . this is because a matter density parameter around 0.30 seems to be favored by observations indicating that the dark energy accounts for 2/3 of cosmic matter @xcite . in fact @xcite has strongly argued a case for @xmath23 from measurements of the physical properties of clusters , cmb anisotropies and the power spectrum of mass inhomogeneities . for the dark energy equation of state we consider 3 models with the one - parameter forms : ( 1)@xmath24,(2)@xmath25,and ( 3)@xmath26 , where @xmath27 is constant . model ( 1 ) , _ viz _ , @xmath28,is a generalization of the cosmological constant case @xmath29 . strictly speaking a constant @xmath30 is valid for the cosmological constant only . yet models of cosmic evolution driven by nonrelativistic matter and a quintessence component @xmath31,an exotic fluid with an arbitrary equation of state @xmath32 ( @xmath33 ) , have been widely studied @xcite . in a number of these models ( particularly those with tracking solutions),both the dark energy density parameter @xmath34 and @xmath30 vary so slowly with redshift @xcite as to justify the approximate use of an effective equation of state parameter @xmath35 @xcite . more generally , the absence of robust fundamental physics - based dark energy models and the difficulty to observe a time dependence of @xmath30 from cmbr@xciteor from fits to luminosity distances @xcite , admits the possibility of a @xmath30 which is constant in some specified range , and which arises as a model - independent approximation to the dark energy equation of state @xcite.(the cosmology with a dark energy @xmath36 and decoupled from ordinary matter so that @xmath37 has been recently discussed by one of us @xcite and by @xcite ) . on the other hand parameterizations ( 2 ) and ( 3 ) are special cases of the two - parameter forms : @xmath38,and @xmath39 which were proposed by @xcite and @xcite , and by @xciterespectively , and recently studied , together with the case @xmath24 , by@xcite . the form @xmath40 diverges at very high redshifts whereas this difficulty is avoided in the model @xmath41 , where @xmath42 as @xmath43 . but , as argued by @xcite , a safer strategy , which we follow here , is to regard these parameterizations as only valid for low - z ( @xmath44 decoupling redshift @xmath45 ) and describing the late behavior of dark energy . this was done by @xcite who studied these parameterizations and found , on taking as prior @xmath46 in a flat universe , that a constant @xmath47 is preferred by the fit to the gold data for type ia supernovae @xcite . also it is already known that the cosmological constant scenario remains consistent with tight constraints from new cosmic microwave background and galaxy clustering data@xcite . quite generally observations seem to require dark energy with present values @xmath48 and @xmath49 @xcite . with this in mind , and noting that recent sneia observations from hst do not indicate a rapid variation of @xmath50 away from its cosmological constant value , we pursue , for simplicity , the following approach : we consider the preceding three @xmath50 parameterizations and set in them,@xmath51 , @xmath52 . then we investigate the constraints on the dark energy equation of state from recent supernova data and observations of the angular sizes of ultra compact radio sources . in section 2 we present the basic equations of the models . in sections 3 and 4 we examine the constraints on them from supernova and angular sizes data respectively . in discussing angular sizes we compare our results with those from a recent work by @xcite , and also consider the implications of our angular size - redshift relations in the presence of relativistic beaming of the radio sources . section 5 winds up the paper with a discussion of the results and some concluding remarks . we consider a spatially flat frw universe ( @xmath53 is the rw scale factor ) @xmath54 with cold matter of zero pressure and energy density @xmath55 and dark energy of density @xmath4 and pressure @xmath56 . denoting the scale factor today by @xmath57,(subscript @xmath58 denotes present - day quantities),and defining @xmath59,where @xmath3 is the red - shift , einstein s gravitational field equations can , in this case , be written as ( @xmath60 ) @xmath61 where @xmath62 is the deceleration parameter and @xmath63 hubble s constant ( an over - dot denotes time differentiation),with @xmath64 being its present - day value ( @xmath65 is the normalized hubble constant ) . defining the density parameters @xmath66 we deduce from equation([e2 ] ) that @xmath67,valid at all times including @xmath68 . combining equations ( [ e2])-([e4])we obtain @xmath69 in the einstein - de sitter(edes)standard model @xmath70 or @xmath71 so that @xmath72 . for the cosmological constant @xmath12 case @xmath73 so that @xmath74 which admits an accelerating universe scenario provided @xmath75 . in this paper we assume that @xmath19 where @xmath76 is the trace of the matter energy - momentum tensor and @xmath20 a dimensionless constant @xcite . then in the matter - dominated epoch of flat cosmology we have from equations ( [ e2 ] ) and ( [ e4 ] ) that @xmath77 and @xmath78 @xcite . we further set , as was done by @xcite and argued in the introduction , @xmath79 . we then obtain from equation([e5 ] ) , @xmath80 the rest of the paper investigates the consequences of this model for @xmath81 using the different dark energy parameterizations discussed in the introduction . inserting @xmath82 in equation([e6])yields @xmath83 here equation([e6])shows that @xmath85 for @xmath86,implying a cosmic deceleration - acceleration transition at redshift @xmath87 . in this case the solution of equation ( [ e6])for @xmath88 is ( @xmath89 ) : @xmath90 the deceleration - acceleration cosmic transition occurs in this model at @xmath92 so that we must have @xmath93 . in this case the solution of equation([e6 ] ) for @xmath88 is(@xmath89 ) : @xmath94}.\ ] ] for a flat universe the luminosity distance in units of megaparsecs may be defined by @xmath95 where @xmath96 in terms of @xmath97 the predicted distance modulus is @xmath98 we next obtain expressions for @xmath97 and @xmath99 in our models . in calculating @xmath99 we use the widely accepted value for the hubble constant @xmath100 @xcite . from equations([e7 ] ) and ( [ e10 ] ) @xmath101 , \qquad w\neq -1/2,\nonumber\\ d_l&\equiv & c(1+z)^2d(0,z ) = ch_0^{-1}(1+z)ln(1+z ) , \qquad w=-1/2.\end{aligned}\ ] ] hence by equation([e12 ] ) , @xmath102\right\ } , \qquad w \neq -1/2,\nonumber\\ \mu_{p } & = & 43.10 + 5\log\left[(1+z)\ln(1+z)\right ] , \qquad w = -1/2.\end{aligned}\ ] ] from equations([e8])and([e10 ] , @xmath103\ ] ] where @xmath104 is the incomplete gamma function @xcite . then @xmath105 so that @xmath106 equations([e9 ] ) and([e10 ] ) give @xmath107 @xmath108\nonumber\\ & = & ch_0^{-1}\exp(w)(1+z)\int_{\frac{1}{1+z}}^{1}u^{w-\frac{3}{2}}\exp(-wu)du.\end{aligned}\ ] ] hence @xmath109 several astrophysics groups@xcitehave recently updated the original supernova data of @xcite and @xcite that provided the first glimpse into an apparently accelerating universe . in particular @xcitehave published photometric and spectroscopic observations of 23 supernovae in the redshift range @xmath110 . confronting our predictions for @xmath99 with their data as analyzed by the batm ( bayesian adapted template match ) method @xcite , and calculated using @xmath111 , we have minimized with respect to the parameter @xmath27 the @xmath112 statistic : @xmath113 ^ 2}{\sigma_i^2},\ ] ] where the summation is over all 23 data points in table 11 of @xcite and @xmath114 the corresponding uncertainties in the observed distance moduli . we discuss the application of this procedure to our 3 models . using the first of equations ( [ e14])we calculated with the aid of equation ( [ e21 ] ) @xmath115 . for @xmath116 figure 1 shows that the resulting curve has a minimum @xmath117 at @xmath118 , with upper limits @xmath119 and @xmath120 at the 68@xmath121 and 95@xmath121 confidence levels ( c.l . ) respectively ( with 22 degrees of freedom- d.o.f . ) . for @xmath122 corresponding to @xmath123 ( coasting universe ) we obtain , using the second equation in([e14 ] ) , @xmath124 . to discuss the implications of the value @xmath125 for the age of the universe in this model we first note the following . for a flat universe with a hubble constant @xmath126 and contributions to the mass - energy density today of 1/3 and 2/3 of its total value from non - relativistic matter and dark energy respectively , it is observed that the age of the universe is @xmath127 with uncertainty of about @xmath128 @xcite . a consistent age @xmath129 is also determined from cmb anisotropy , independently of @xmath130 @xcite . moreover , computer simulations of globular - cluster stars evolution produce ages of @xmath131 @xcite . these estimates agree with values of @xmath132 obtained by a variety of other methods , e.g. from rates of cooling of old white dwarf stars or from radioactive chronology @xcite . finally , assuming @xmath133 @xcite deduce the constraint @xmath134,in agreement with the product @xcite @xmath135 the observed ages of the universe are therefore consistent with a consensus age of about @xmath136 @xcite . in the present model we have , from equation([e7 ] ) , @xmath137 then @xmath138 gives @xmath139 . at the 68@xmath121 c.l . @xmath140 corresponding to @xmath141 , an estimate accommodated by equation([e22 ] ) . the coasting cosmology ( @xmath122 ) corresponds , as is well known , to @xmath142 . in this model we have evaluated equation([e21])in the range @xmath143 using equation([e18 ] ) for the calculated distance modulus and plotted the results in figure 2 . we note that @xmath112 decreases monotonically as @xmath27 increases from @xmath144 but reaches a minimum @xmath145 at @xmath146 , corresponding to the transition redshift @xmath147 . the value @xmath148 corresponds to the 68@xmath121 c.l . limit . here we used equation([e20])in equation([e21])and plotted @xmath112 versus @xmath149 in figure 3 . the curve has a minimum @xmath150 at @xmath151,corresponding to the transition redshift @xmath152 . the angular size distance of a light source is @xmath153 where @xmath154 is the proper distance of the source . in a flat universe @xmath155 , where @xmath156 is the source s radial coordinate . the angular size - redshift relation @xmath157 where @xmath158 is the source s angular size , and @xmath159 its intrinsic length , measured in _ _ p__arse__c__s ( 1__pc _ _ = @xmath160)and assumed to be redshift - independent , is one of observational cosmology s important tests of cosmological models . like other classical kinematic tests it does not , generally , distinguish between cosmological models at low redshifts @xmath161 where the models are expected to converge . in fact for models with constant @xmath81,one has , for @xmath161 , @xmath162,\ ] ] which is formally the frw result for small redshifts @xcite . but for @xmath163 there is less confidence in the measurements because of possible influences of poorly understood galactic evolutionary effects . however @xcite has argued that ultra - compact radio sources with angular sizes ( measured using vlbi : very long baseline interferometry ) in the milliarcsecond ( _ _ mas__= @xmath164 _ radians _ ) range ( typically less than a hundred parsecs in extent ) are deeply embedded in active galactic nuclei(agn)@xcite and thus sheltered from extra - galactic evolutionary effects . objects of this type have a fleeting existence ( @xmath165),so it is reasonable to assume that characteristic parameters of their population ( e.g. linear sizes)do not change on a cosmological time scale . @xcite showed that the angular size - redshift test for ultra - compact sources favors the einstein - desitter @xmath70 canonical model . but subsequently @xcite demonstrated that the data is compatible with low - density constant-@xmath12 models , indicating that the best choice of cosmological parameters for spatially flat universes was @xmath166 and @xmath167 . in their latter work @xcite , utilized a data set of 337 ultra - compact sources selected by @xcite from a 2.29 ghz survey by @xcitecomprising 917 sources with a correlated flux limit of approximately 0.1 jy ( 1 jy @xmath168 jansky @xmath169 ) . from their study of this compilation they conclude that the canonical model is ruled out by the observed angular diameter - redshift relation . later on @xcite refined the analysis of gurvits original data set @xcite and found for flat universes that @xmath170 . building on @xcite earlier work @xcitecompiled a new data set of 330 compact radio sources which has , subsequently , been used by several authors in order to constrain the parameters of different quintessence cosmological models @xcite . we next give general formulae for the @xmath171 relations in the present models . we write @xmath172 where @xmath173 is the source s characteristic angular scale ( in mas ) . these expressions are ( in mas ) : model 1 : @xmath174 @xmath175 model 2 : @xmath176 model 3 : @xmath177 using @xmath178 we find the small@xmath179 expansions of these equations : model 1 : @xmath180 @xmath181 where from equation([e6 ] ) @xmath182 . thus both these equations formally coincide with equation([e25 ] ) . models 2 and 3 : @xmath183,\ ] ] where @xmath184 in model 2 and @xmath185 for model 3 . in both cases the small-@xmath3 expansion of @xmath186 agrees with equation([e25 ] ) on retaining only the @xmath27-independent part of @xmath81 . the existence of a critical redshift @xmath187 corresponding to a minimum angular size can be qualitatively understood as follows . the reason for it is not just because , in the context of cosmic expansion , was the light received today from a source emitted when the source was closer , but also , more importantly , because at a larger @xmath3 the source of a standard linear size occupies a larger fraction of a large circle . from equations ( [ e26])-([e29 ] ) the redshift @xmath188 satisfies the equations : model 1 : @xmath189 @xmath190 model 2 : @xmath191\int_1^{1+z_m}u^{w-\frac{1}{2}}\exp(-wu)du.\ ] ] model 3 : @xmath192\int_{\frac{1}{1+z_m}}^1u^{w-\frac{3}{2}}\exp(-wu)du.\ ] ] several authors @xcite studied critical redshifts in different models to find out how sensitive @xmath187 is to variation of parameters like @xmath27 . we address this question in 4.3.2 . our object now is to investigate constraints on the parameters @xmath27 and @xmath193 using the data compilation of @xcite for the angular size measurements of milliarcsecond radio sources , observed at frequency @xmath194 , in the redshift range @xmath195 the number of sources , originally 330 , was reduced to those with a spectral index @xmath196 in the range @xmath197 and total luminosity @xmath198 so as to minimize any possible dependence of angular size on @xmath196 and also restrict the intrinsic size of the sources . these criteria were met by 145 sources which were then grouped into 12 bins of 12 - 13 sources per bin . this binned data was used in figure 10 of the paper by @xcite . we attempted the determination of the best values of the models parameters @xmath199 and @xmath27 through a simultaneous minimization with respect to @xmath193 and @xmath27 of the @xmath112 statistic @xmath200 ^ 2}{\sigma_i^2},\ ] ] where @xmath201 denote the predicted angular sizes given by equations ( [ e26])-([e29]),@xmath202 the corresponding observed values of @xcite , and @xmath114 the observation s error of the sample s @xmath203 bin . we have chosen the range of @xmath193 to span the interval @xmath204 \emph{mas}$ ] and @xmath27 in intervals appropriate to each model . we adopt a minimization procedure , for 10 d.o.f , that produces minimum @xmath112 values for @xmath27 values corresponding to different values of @xmath193 in its range of variation . the least value among these @xmath112 minima , denoted @xmath205,corresponds to a minimization of @xmath112 with respect to both @xmath193 and @xmath27 . the results are displayed in tables 1 , 2 , and 3 , and in figure 5 . the values of @xmath27 and @xmath193 corresponding to @xmath205 are as follows . * model * * 1 * _ case _ @xmath116 : using equation ( [ e26 ] ) in equation([e37])with @xmath27 in the range @xmath206 $ ] we obtain ( table 1 ) @xmath207 corresponding to @xmath208 and @xmath209 . figure 4 shows , for @xmath210 , the @xmath211 and @xmath212 confidence contours in the @xmath213 plane . * model * * 1 * _ case _ @xmath214 ( _ coasting cosmology _ ) : from equations ( [ e27 ] ) and ( [ e37 ] ) we deduce ( see figure 5 ) that @xmath215 at @xmath216 _ mas _ and @xmath217 _ mas _ , where the errors produce the @xmath211 and @xmath212 confidence limits respectively . * model 2 * using equations ( [ e28 ] ) and ( [ e37 ] ) and taking @xmath27 in the range [ 0,3 ] we note , in this case , that @xmath112 decreases monotonically from @xmath218 at @xmath219 and @xmath220 to the minimum value @xmath221 at @xmath222 and @xmath223 ( table 2 ) . the value @xmath224 corresponds to @xmath225 . the @xmath211 and @xmath212 confidence contours in the @xmath213 plane , for @xmath226 , are displayed in figure 6 . * model 3 * in this model equations ( [ e29 ] ) and ( [ e37 ] ) reveal , for @xmath27 in the range @xmath227 $ ] , that @xmath112 descends from @xmath228 at @xmath229 and @xmath220 to a minimum @xmath230 at @xmath231 and @xmath232 ( table 3 ) . this value of @xmath27 corresponds to @xmath233 . for @xmath234 , the @xmath211 and @xmath212 confidence contours are shown in figure 7 . equations ( [ e33])-([e36 ] ) can be solved for @xmath187 corresponding to specific values of @xmath27 . the results are plotted in figures 8 - 10 for models 1 - 3 respectively . the redshifts @xmath187 for the best - fit @xmath27 values are shown on the diagrams . we note that in both models 1 and 3 the best - fit @xmath27 value corresponds to @xmath235 . this reiterates the point made by @xcite that the minimum redshift test `` can not by itself differentiate between different cosmological models '' because different scenarios might correspond to the same @xmath187 . the first and third models in this paper are generically quite distinct , yet they have the same @xmath187 . in fact for all three models the values of @xmath187 are quite close . in particular in model 2 the best - fit @xmath27 gives @xmath236 which almost coincides with @xmath237 for coasting cosmology . expectedly for @xmath238 in model 1 we recover ( figure 8) the standard model result @xmath239 . the predictions of the @xmath171 relations ( [ e26])- ( [ e29 ] ) for @xmath193 and @xmath27 corresponding to @xmath205 in each model are plotted , alongside the data of @xcite , in figure 11 . ( table 4 gives @xmath240 and the corresponding @xmath241 values for @xcite data points and the end points @xmath242 of their error bars ) . the inset magnifies the neighborhood of the minima of the curves . the curves , drawn for the best - fit values of @xmath27 and @xmath193,cluster within a narrow spread of @xmath241 and appear to be reasonably consistent with @xcite data . none of the four models appears to be particularly favored over the others by this data . lastly in the work of @xcite a multi - parameter regression analysis of the data with @xmath243 a free parameter yields @xmath244 . this corresponds to @xmath245 , a range within which fall all the @xmath193 values in table 5 . in his 1993 analysis of compact radio sources @xcitefound the angular sizes to be essentially independent of redhift in the interval @xmath246 . in the latter work of @xcitethis feature appeared to persist for median angular sizes of sources with @xmath247 . although according to @xcite , and as mentioned earlier on , it is reasonable to assume that milliarcsecond ultra - compact radio sources are not affected by evolutionary effects , it is still important to treat @xcite data , as emphasized by them , with caution . a widely - accepted model of an ultra compact radio source is one in which a central low luminosity object is straddled by a pair of radio - bright lobes emitting synchrotron radiation . the lobes are sustained by two hot gas jets@xcite . in such a picture differences in the spectral index between the core , the central engine , and its jet components may introduce a `` linear size - redshift '' dependence even in the absence of evolution . @xcitehave however argued that such a dependence is likely to be weak . nevertheless more data at various frequencies going beyond limited source samples can , in particular , help verify the importance of such a dependence and also , more generally , enhance further the credibility of milliarcsecond radio - sources as standard measuring rods . with this goal in mind @xcitehave recently studied the implications of an updated sample of sources in the @xcitecatalogue . we shall shortly discuss their work and compare their results to ours . another aspect of interest is relativistic beaming@xcite , the process by which the dppler effect modifies the appearance of a radio source with lobes . for a lobe whose jet axis is oriented close to the line of sight from the source to the earth relativistic beaming generally increases the apparent radio power of the source and decreases its angular size . the implications of this orientation bias on the frw @xmath171 relation@xcitehave been studied by @xcitewith the aim of evaluating the statistical confidence in the observational data , particularly in relation to estimates of the value of @xmath248 . a similar study was also undertaken by @xcite . we shall consider a simple variation of the @xciteapproach in order to assess the impact of this orientation bias on our calculations . @xcitereexamined the @xcite catalogue of ultra compact radio sources , updating it with respect to both red shift and radio information and replacing the original choice of 337 sources by @xcitewith a sample of 613 objects in the red shift range @xmath249 to @xmath250 . over its red shift range of @xmath251 @xcite 337 objects sample contains sources of luminosity that varies over three orders of magnitude , from 0.01 to 1 jy ( see fig 3 in @xcite ) . moreover the sources in the high red shift range @xmath252 exhibit a smaller dispersion in luminosity as compared to sources with @xmath253 . this suggests that the higher red shift objects have similar linear sizes . this is why @xcite , and also @xcite , ignore objects with @xmath254 . this mismatch between the luminosity of higher and lower z sources is also apparent in the 330 sources sample of @xcite(see fig 3 in @xcite ) . this is why that we also limit ourselves in this paper to @xcite data with @xmath255 ( see fig 10 in @xcite and table 4 in the present paper ) . of the 613 sources in the sample of @xcite 468 have @xmath256 . these were placed in 6 bins with 78 objects in each bin . but @xcite show that the cosmological parameters which best fit the @xmath257 data in this case , including a best - fitting value @xmath258 for the source s intrinsic length , are not very sensitive to the choice of binning . rather than relying too heavily on a particular binning @xcite produce data points corresponding to bin sizes of 76 , 77 and 78 objects and take for their @xmath259 points values that are a composite of these three cases . their resulting data points are : @xmath260 @xmath261 and @xmath262,with standard deviation @xmath263 . using this data , and marginalizing over the sources intrinsic length,@xcite obtain the best - fit parameter values @xmath264 and the curvature parameter @xmath265,where the error bars are @xmath211 confidence limits . these results do not agree with our models . in fact the @xmath266 curve for@xcite results would lie well below the base of figure 11 . we have confirmed this disagreement by doing a @xmath112 analysis of the predictions of our models against @xcite angular size results . we will comment on this matter in the summary and conclusions section . @xcite gave qualitative and quantitative discussions of the effect of relativistic beaming on the frw @xmath171 relation . they consider a simulation sample of compact radio sources , with each source composed of two identical but oppositely directed jets having an angle @xmath267 to the line of sight , where @xmath268 . each source emits an isotropic power - law spectrum with flux density @xmath269^{3+\alpha},\ ] ] where @xmath270 is the source s luminosity , @xmath97 is the luminosity distance , @xmath196 is the spectral index , @xmath271 is the jet speed / c , and @xmath272 is the lorentz factor of the jets . for a quasi - continuous jet formed out of finite - lifetime blobs the appropriate power index above is reduced to @xmath273 @xcite . the inclusion of the source s lobe orientation in the frw @xmath274 relation @xmath275 replaces the source s intrinsic length @xmath159 by @xmath276 where @xmath277 is the projection of @xmath159 in the plane of the sky ( @xmath278 for a source whose jet axis lies along the plane of the sky ) it follows from equation ( [ e38])that the maximum allowed angle in the sample is given by @xcite : @xmath279^{\frac{-1}{3+\alpha}}\right),\ ] ] where @xmath280 denotes the sources flux density limit . @xcitedefine the average lobe orientation angle by @xmath281 since @xmath267 is small one can assume @xmath282 and insert @xmath283 into the @xmath171 relations . the essence of the @xcite quantitative method is to investigate the effect of this modification on the predictions of equation ( [ e39 ] ) . the twin - jet model of @xcite assumes that the measured angular size of the source corresponds to the separation of the jets projected onto the plane of the sky . @xcite considers this feature to be unrealistic because the flux of the receding ( away from us ) jet is very much smaller than that of the advancing(towards us)jet , for example by a factor of up to @xmath284 for a jet lorentz factor @xmath285 of 5 , and may therefore be neglected @xcite . the relativistic beaming hypothesis is primarily based on two parameters : the lorentz factor @xmath285 ( or equivalently the parameter @xmath286 the jet velocity / c ) and the viewing angle @xmath267 . in a study of 25 quasars with @xmath3 between 0.1 and 2.5 ( some quasars with their nonstellar radiation spectrum and evidence of jets and extended emission features also display relativistic beaming effects)@xcite find that @xmath285 increases with @xmath3 if @xmath287 but remains constant at @xmath288 if @xmath289 . with this in mind @xcite take , in their analysis of a simulation data , @xmath290 , although they point out that this choice does not qualitatively affect their results . however a choice of @xmath285 linked to a positive @xmath248,as in the 1995 paper by @xcite , is hard to justify in these days where @xmath291 ( accelerating universe ) is very probable . moreover there are other reasons that might render an arbitrary choice of @xmath285 fraught with uncertainties@xcite . in addition , geometric projection effects of components of jets of galaxies or quasars can lead to dramatically misleading values of @xmath285 . for example the transverse motion of blobs of gas across the sky in a well - studied jet in the galaxy @xmath292 ( at about @xmath293 away from us)appears to be _ superluminal _ , corresponding to a velocity nearly 6 times the speed of light!@xcite . this cosmic illusion of faster - than - light velocities is simply a projection effect , produced by the blobs of gas moving at a very small angle @xmath267 at a near - velocity - of - light speed @xcite . it is worthy of mention that doppler boosting was first discussed by @xciteto explain the apparently one - side jet in@xmath292 . in this paper we choose to relate @xmath285 to @xmath267 by optimizing the relativistic beaming effect . this technique was used by @xcite,@xcite , and @xcite . define the doppler factor which fundamentally characterizes relativistic beaming by ( see equation ( [ e38 ] ) ) : @xmath294^{-1}\ ] ] optimization of relativistic beaming is obtained by requiring @xmath295 . this yields @xcite @xmath296 , or @xmath297 . since the lobe orientation bias replaces @xmath159 by @xmath298 in the @xmath171 relation , @xmath299 in equations ( [ e26])-([e29])is replaced by @xmath300 . in what follows we discuss the constraints of the observational data on @xmath300 . clearly application of the @xmath112procedure of ( 4.3.1 . ) to the modified @xmath171 relations using the binned @xcite data assigns now the best - fit values of @xmath193 to @xmath300 and leaves those of @xmath27 ( and hence of @xmath248)undisturbed . taken together the three models imply ( see table 5 ) @xmath301 mas , or , equivalently , @xmath302 pc . since milliarcsecond ultra - compact radio sources are typically less than a hundred parsec in extent the last inequality would imply that @xmath303 , or @xmath304 . ( if we use @xmath305 we get @xmath306 ) . it is known @xcitethat the most powerful agn jets seem to have lorentz factors of the order of these upper bounds . we have studied in this paper supernova and compact radio sources angular sizes constraints on three cosmological models whose dynamics is driven by non - relativistic matter of density parameter @xmath307 and a smooth time - dependent dark energy component with density @xmath308 ( @xmath309 is the matter energy - momentum tensor ) and equation of state @xmath310 being either constant or redshift - dependent , with @xmath311 in the latter case . the variation @xmath312 implies that the dark energy vanishes in the early universe leaving the standard model s primordial nucleosynthesis predictions intact . this is somewhat reminiscent of the modified general relativity model of @xcite which can be cast in the form of a variable-@xmath12 cosmology with @xmath313 , @xmath314 being the ricci tensor . there the radiation density @xmath315 in the early universe , as in standard flat cosmology . however the friedmann equation is modified so that the standard cosmic expansion rate is altered by the factor @xmath316 where @xmath317 , a constant , satisfies @xmath318 in general relativity . here the postulated dark energy ansatz does not affect the standard cosmic expansion rate of the early universe . it reduces in the matter - dominated phase to @xmath319,a variation that was extensively studied for the cosmological constant @xcite . our main results are summarized in table 5 . for the first model with @xmath30 constant ( implying by equation ( [ e6 ] ) that @xmath320 ) the supernova data of @xcite admits an accelerating universe with @xmath321 and @xmath322 ( @xmath323 ) and a coasting universe with @xmath324(@xmath325 ) . on the other hand @xcite angular sizes data allows a mildly decelerating universe with @xmath326,@xmath287 and characteristic angular scale @xmath327 ( @xmath328 ) and also a coasting cosmology with @xmath329 and @xmath330 ( @xmath331 ) . the value @xmath332 corresponding to a source s intrinsic length @xmath333 is very close to @xmath334 obtained as the best - fit value in an earlier model by @xcite where the scale factor is essentially linear in @xmath335,_viz_,@xmath336 . the results of the constant-@xmath30 model ( and also those of the other models in this paper ) do not extrapolate to the radiation dominated universe since there , as we have pointed out in the introduction , @xmath4 vanishes . nevertheless an accelerated expansion in model 1 may pose problems for structure condensation after matter dominance : structure tends not to form in the presence of cosmic acceleration . however the modest acceleration ( @xmath337 ) in model 1 corresponds to @xmath338 with upper limits @xmath339 and @xmath340 at the @xmath211 and @xmath212 confidence levels respectively . these values of @xmath30 correspond from equation ( [ e6 ] ) to @xmath341 and @xmath342 respectively so that within the limits of the quoted confidence levels a decelerating universe is not excluded . in fact it has been argued @xcite that absorption by intergalactic dust of light travelling over immensely long distances might explain the faintness of extragalactic sne ia obviating the need for a cosmic acceleration based explanation . yet we are inclined to believe that our supernova results for the coasting universe scenario are more robust because of their concordance with those of @xcite which are obtained by a different approach . a flat cosmology with dark energy of constant @xmath30 and matter density parameter @xmath21 held at @xmath46 was also considered by @xcite who however assumed noninteracting dark energy and used the larger set of the supernova `` gold '' data . they find in this case the preferred ( cosmological ) constant value @xmath343 . models 2 and 3,which by construction describe a present - day accelerating universe with @xmath344,are consistent with the data for universes that undergo deceleration - acceleration transitions at redshifts in the range @xmath345 . in the standard @xmath346 cosmological model @xmath347 for @xmath348 , in contrast with the observational value @xmath349 from the sneia analysis of @xcite . predictions of @xmath350 in seven popular quintessence models inspired by supergravity or m1 string theory have recently been studied by @xcite who noted that all of them can , in the low-@xmath3 approximation @xmath351,(@xmath352),mimic the @xmath346 model . here in models 2 and 3 the sneia data give @xmath353 and @xmath354,both values being very close to the observational result @xmath355 of @xcite . on the other hand consistency of the models with the angular size data @xcite yield @xmath356 and @xmath233 , to be compared with @xmath357 from the @xmath346 . inspection of figure 11 and table 5 leads to an important conclusion : the model s angular size - redshift curves drawn in figure 11 fit @xcite data equally well . this is a reflection of the small values and span of @xmath358 as seen in table 5 for the results : for the four models @xmath358 varies only from 4.47 to 4.75 . by comparison in the supernova test the @xmath358 values are considerably larger and lie in the range 15.4 to 19.67 so that one has more faith in the agreement of the models with the angular size data . the predictions of the models for the critical minimum redshift in the angular size - redshift relation give @xmath187 values in the narrow interval [ 1.65,1.72 ] compared to the standard model result @xmath359 . the constant-@xmath360 and the third models have the same @xmath235 whereas the coasting universe and model 2 have @xmath361 ( the same as that of @xcite)and @xmath236 respectively . thus the minimal redshift cannot , by itself , effectively discriminate between these models . we have also compared the predictions of our models with those of the recent work of @xcitein which they revisit the old @xcitecatalogue of ultra compact radio sources and reconsider an angular size - redshift data set in the light of modern preferences of the cosmological parameters . we find that their results do not agree with our models . in a sense this is not surprising since the underlying premises of the two works are different:@xciteapproach is based on using one simple potential to test the hypothesis that vacuum energy is constant . the present models are phenomenological and based on a time - dependent dark energy coupled to matter . hence there is no overlap between the two approaches@xcite . finally , we tested our angular size - redshift relations in the presence of relativistic beaming . relativistic beaming in the frw @xmath171 relation was investigated by @xcite . here we used a simple variation of their method in a two - jet model of radio sources with the advancing ( towards us)jet axis close to the line of sight . in this picture consistency of the models with @xcite data is found for a beaming lorentz factor @xmath362 . such an upper bound is consistent with values of the lorentz factor for powerful agn jets @xcite . to conclude , we have presented dark energy models that are in reasonable agreement with the supernova data of @xciteand in good agreement with the @xcite compact radio source angular size versus redshift binned data . the three models that we have studied are simplified versions of ones recently considered by @xcite in a different context . but as remarked by these authors comparing models for the equation of state of dark energy will remain something of a mug s game until there exists substantially more data at higher redshifts . we thank r.g . vishwakarma for sending @xcite complete data sets on angular sizes of compact radio sources , and j.c.jackson for a useful correspondence . we also acknowledge the financial support of the directorate of scientific research of the university of khartoum . one of us ( ammar ) thanks professor k.r . sreenivasan for hospitality at the abdus salam international center for theoretical physics , trieste , italy , where part of this work was done , and the swedish agency for international development ( sida ) for financial support . ammar also thanks dr . abu bakr mustafa for his hospitality at computerman college for computer studies , khartoum , where also part of this work was done . 99 abdel - rahman , a .- m.m . 1997 , gen . grav . , 29 , 1329 abdel - rahman , a .- m.m.2002,nuovo cim.,b117,311 abdel - rahman , a .- m.m . , & hashim , m.h.a . 2005 , , 298 , 519 abramowicz , m.a.,ellis , g.f.r . , & lanza , a. 1990 , , 361 , 470 abramowitz , m. , & stegun , i.a . 1964 , handbook of mathematical functions , dover publications , inc . alam , u. , sahni , v. , saini , t.d . , & starobinsky , a.a . 2003 , , 344 , 1057 alam , u. , sahni , v. , & starobinsky , a.a . 2004 , jcap , 0406 , 008 alcaniz , j.s . 2002 , , 65 , 123514 al - rawaf , a.s . , & taha , m. o. 1996a , phys . b , 366 , 69 al - rawaf , a.s . , & taha , m. o. 1996b , gen . grav . , 28 , 935 arbab , a.i.1997 , gen.rel.grav.,29,61 arbab , a.i.2003,jcap , 0305,008 arcuri , r.c . , & 1994 , , 50 , 2928 aurich , r. , & steiner , f. 2003 , , 67 , 123511 baccigalupi , c. 2003 , mem . s. a. it . della supplementi societa astronomica italiana ) , 3 , 192 barris , b.j . , 2004 , , 602 , 571 caldwell , r.r . , dave , r. , & steinhardt , p.j . 1998 , , 80 , 1582 carvalho , j.c . , lima , j.a.s . , & waga , i. 1992 , , 46 , 2404 cepa , j.2004 , , 422 , 831 chaisson , e. , & mcmillan , s . 2002 , _ astronomy today _ , ( prentice hall ) chen , g. & ratra , b. 2003 , , 582 , 586 chevallier , m. , & polarski , d. 2001 , int . j. mod . d , 10 , 213 chiba , t. , sugiyama , n. , & nakamura , t. 1997 , , 289 , l5 corasaniti , p.s . , kunz , m. , parkinson , d. , copeland , e.j . , & bassett , b.a . 2004 , , 70 , 083006 dabrowski , y . , lasenby , a . , & saunders , r . 1995 , , 277 , 753 di pietro , e. , & claeskens , j .- f . 2003 , , 341 , 1299 . dicus , d.a . , & repko , w.w . 2004 , , 70 , 083527 efstathiou , g.1999 , , 310 , 842 freedman , w.l . , 2001 , , 553 , 47 freedman , w.l . , & turner , m.s . 2003 , rev . phys . , 75 , 1433 freese , k . , adams , f.c.,frieman , j.a . , & mottola , e . 1987 , nucl.phys.b,287,797 frey , s.,gurvits , l.i.,kellermann , k.i.,schilizzi , r.t . , & pauliny - toth , i.i.k.1997,,325,511 frieman , j.a . , hill , c.t . , stebbins , a. , waga , i. 1998 , , 75 , 2077 gardner , c. l. 2005 , nucl . phys . b , 707 , 278 gurvits , l.i . 1994 , , 425 , 442 gurvits , l.i . , kellermann , k.i . , & frey , s. 1999 , , 342 , 378 huterer , d. , & turner , m.s . 2001 , , 64 , 123527 jackson , j.c . , & dodgson , m. 1996 , , 278 , 603 jackson , j.c . , & dodgson , m. 1997 , , 285 , 806 jackson , j.c . 2004 , jcap , 0411 , 007 jackson , j.c . , & jannetta , a.l . 2006,jcap , 0611 , 002 jackson , j.c . 2007,(private communication ) . jain , d. , dev , a. , & alcaniz , j.s . 2003 , class . grav . , 20 , 4485 johri , v.b . 2004 , astro - ph/0409161 johri , v.b . , & rath , p.k . 2005 , astro - ph/051007 kellerman , k.i . 1993 , , 361 , 134 kneller , j.p . , & strigari , l.e . 2003 , , 68 , 083517 knox , l. , christensen , n. , & skordis , c. 2001 , , 563 , l95 kolb , e.w . , & turner , m.s . 1990 , _ the early universe _ , ( addison wesley ) krauss , l.m . , & schramm , d.n . 1993 , , 405 , l43 krauss , l.m . , & chaboyer , b. 2001 , astro - ph/0111597 krolik , j.h.1998 , active galactic nuclei , ( princeton univ . press ) lima , j.a.s . , & carvalho , j.c . 1994 , gen.rel.grav.,26,909 lima , j.a.s . , & alcaniz , j.s . 2000a , , 357 , 393 lima , j.a.s . , & alcaniz , j.s . 2000b , gen . , 32 , 1851 lima , j.a.s . , & alcaniz , j.s . 2002 , , 566 , 15 lind , k.r . , & blandford , r.d . 1985 , , 295 , 358 linder , e.v . 2003 , , 90 , 91301 majernik , v.2001,phys.lett.a,282,362 majernik , v.2003,gen.rel.grav.,35,1007 melchiorri , a. 2004 , astro - ph/0406652 ostriker , j.p . , & souradeep , t. 2004 , pramana , 63 , 817 oswald , t.d.,et al . 1996 , , 382 , 692 overduin , j.m . , & cooperstock , f.i . 1998 , , 58 , 043506 ozer , m. 1999 , , 520 , 45 padmanabhan , t. , & choudhury , t.r . 2003 , , 344 , 823 peacock , j.a.1999,cosmological physics,(c.u.p . ) peebles , p.j.e . , & ratra , b. 2003 , rev . phys . , 75 , 559 perlmutter , s. , et al . 1999 , , 517 , 565 peterson , b.m.1997 , an introduction to active galactic nuclei , ( c.u.p . ) preston , r.a.,et al.1985,,90,1599 ratra b. , & peebles , p.j.e . 1988 , , 37 , 3406 riess , a.g . , et al . 1998 , , 116 , 1009 riess , a.g . , et al . 2004 , , 607 , 665 sandage , a . 1988 , , 26 , 561 shklovsky , i.s.1964a ,- aj,7,748 shklovsky , i.s.1964b ,- aj,7,972 spergel , d. , & pen , u. 1997 , , 491 , l67 steinhardt , p.j . , wang , l. , & zlatev , i. 1999 , , 59 , 123504 stepanas , p.g . , & saha , p.1995 , , 272,l13 tonry , j.l . , et al.2003 , , 594 , 1 turner , m.s . , & white , m. 1997 , , 56 , 4439 turner , m.s . 2002a , int.j.mod.phys.a,17s1,180 turner , m.s . 2002b , int.j.mod.phys.a,17,3446 turner , m.s . 2002c , , 576,l101 turner , m.s . 2002d , astro - ph/0207297 ubachukwu , a.a.1999 , publ . aust . , 16 , 130 ubachukwu , a.a . & chukwude , a.e.2002 , j.astrophys . astr.,23,235 vermeulen , r.c . & cohen , m.h.1994 , , 430,467 vishwakarma , r.g . 2001 , class . grav . , 18 , 1159 vishwakarma , r.g . 2003 , , 345 , 545 wang , l. , caldwell , r.r . , ostriker , j.p . , & steinhardt , p.j . 2000 , , 530 , 17 weller , j. & albrecht , a. 2002 , , 65 , 103512 wetterich , c . 1995 , , 301 , 321 zhu , z. , & fujimoto , m. 2002 , , 581 , 1 zhu , zong - hong , fujimoto , masa - katsu , & he , xiang - tao 2004 , , 417 , 833 zlatev , i. , wang , l. , & steinhardt , p.j . 1999 , , 82 , 896 ccc\|\|cc\|\|ccc 0.1 & 0&106 & & & 1.25&-0.45&4.72 + 0.2 & 0&85.44&&&1.30&-0.55&4.74 + 0.3&0&66.53&&&1.35&-0.60&4.75 + 0.4&0&50.16&&&1.40&-0.65&4.80 + 0.5&0&36.33&&&1.45&-0.70&4.88 + 0.6&0&25.03&&&1.50&-0.80&4.97 + 0.7&0&16.27&&&1.55&-0.85&5.08 + 0.8&0&10.04&&&1.60&-0.90&5.20 + 0.9&0&6.36&&&1.65&-0.95&5.35 + 0.95&0&5.47 & & & 1.70&-1&5.51 + 1.00&-0.05&5.17&&&1.80&-1&6.28 + 1.05&-0.15&5.02&&&1.90&-1&7.96 + 1.10&-0.20&4.90&&&2.00&-1&10.54 + 1.15&-0.30&4.80&&&2.10&-1&14.02 + 1.20&-0.40&4.76&&&2.20&-1&18.41 + ccc\|\|cc\|\|ccc 0.1&3&104.29&&&1.25&1.00&5.00 + 0.2&3&80.97 & & & 1.30&0.90&4.78 + 0.3&3&60.91 & & & 1.35&0.70&4.61 + 0.4&3&44.12 & & & 1.40&0.60&4.50 + 0.5&3&30.59&&&1.45&0.50&4.47 + 0.6&3&20.33 & & & 1.50&0.40&4.52 + 0.7&3&13.32 & & & 1.55&0.30&4.64 + 0.8&3&9.58 & & & 1.60&0.20&4.84 + 0.85&2.80&8.85 & & & 1.65&0.15&5.07 + 0.9&2.50&8.24 & & & 1.70&0.10&5.39 + 0.95&2.20&7.65 & & & 1.75&0&5.78 + 1.00&1.90&7.08 & & & 1.80&0&6.28 + 1.05&1.70&6.56 & & & 1.90&0&7.96 + 1.10&1.50&6.09&&&2.00&0&10.54 + 1.15&1.30&5.66&&&2.10&0&14.02 + 1.20&1.10&5.30&&&2.20&0&18.41 ccc\|\|cc\|\|ccc 0.1&3&108.87&&&1.25&1.85&4.83 + 0.2&3&89.00&&&1.30&1.55&4.70 + 0.3&3&71.25&&&1.35&1.35&4.61 + 0.4&3&55.63 & & & 1.40&1.15&4.58 + 0.5&3&42.13&&&1.45&0.95&4.61 + 0.6&3&30.77&&&1.50&0.75&4.69 + 0.7&3&21.53 & & & 1.55&0.65&4.82 + 0.8&3&14.42 & & & 1.60&0.55&5.04 + 0.9&3&9.43 & & & 1.70&0.55&6.28 + 1.00&3&6.57 & & & 1.80&0.55&8.59 + 1.05&3&5.94 & & & 1.90&0.55&11.97 + 1.10&2.65&5.80 & & & 2.00&0.55&16.43 + 1.15&2.35&5.28 & & & 2.10&0.55&21.95 + 1.20&2.05&5.03 & & & 2.20&0.55&28.55 + cccccc 1&-0.29&0.785&3.216&0.961&0.461 + 2&-0.20&0.703&2.346&0.840&0.420 + 3&-0.12&0.585&1.504&0.722&0.374 + 4&-0.05&0.550&3.170&0.823&-0.427 + 5&0.04&0.597&1.277&0.704&0.411 + 6&0.10&0.550&2.683&0.765&-0.075 + 7&0.15&0.562&0.634&0.621&0.482 + 8&0.19&0.332&1.354&0.537&-0.107 + 9&0.29&0.389&0.895&0.508&0.196 + 10&0.34&0.690&1.078&0.768&0.573 + 11&0.41&0.525&0.734&0.623&0.433 + 12&0.55&0.633&0.990&0.731&0.535 c\|ccccc\|ccccccc\|c 1&&-0.7&15.4&-&-0.2&&-0.45&1.25&4.72&-&0.05&&1.65 + & & -0.5&19.67&-&0&&-0.50&1.30&4.75&-&0&&1.72 + 2&&1.1&16.5&0.45&-0.5&&0.50&1.45&4.47&1&-0.50&&1.71 + 3&&1.7&16.35&0.42&-0.5&&1.15&1.40&4.58&0.77&-0.50&&1.65	three models of a flat universe of coupled matter and dark energies with different low - redshift parameterizations of the dark energy equation of state are considered . the dark energy is assumed to vary with time like the trace of the energy - momentum tensor of cosmic matter . in the radiation - dominated era the models reduce to standard cosmology . in the matter - dominated era they are , for modern values of the cosmological parameters , consistent with data from sne ia searches and with the data of @xcite for angular sizes of ultra compact radio sources . we find that the angular size - redshift tests for our models offer a higher statistical confidence than that based on sne ia data . a comparison of our results with a recent revised analysis of angular size - redshift legacy data is made , and the implications of our models with optimized relativistic beaming in the radio sources is discussed . in particular we find that relativistic beaming implies a lorentz factor less than 6,in agreement with its values for powerful active galactic nuclei .
an operator @xmath0 is a projector if it is idempotent , i.e. , if it is endowed with property @xmath1 . the projection is along ( or parallel to ) its null space @xmath2 and onto its range @xmath3 . this entails that @xmath4 for @xmath5 and @xmath6 for @xmath7 . if the subspaces @xmath3 and @xmath2 are orthogonal the operator is called an orthogonal projector , which is the case if and only if @xmath0 is self - adjoint . otherwise it is called oblique projector . oblique projectors , though introduced early @xcite , have received less attention than orthogonal projectors . nevertheless , quite recently there has been a renewed interest in relation to their properties and applications @xcite . in particular , oblique projectors have been shown to be of significant relevance to signal processing techniques @xcite . the present effort is very much motivated by problems arising in the area of signal representation outside the traditional orthogonal basis setting @xcite . in such a context a signal @xmath8 , represented mathematically as an element of a vector space , is approximated as a linear expansion of the form [ eq1 ] f^k=_i=1^k c_i v_i . the vectors @xmath9 in are sometimes sequentially fed or chosen according to some optimality criterion . in such situations one needs to be in a position to effectively adapt the coefficients of the linear superposition so as to account for the possibility of changes in the model . this may entail a)increasing the order @xmath10 of the model by incorporating new terms in the expansion b)reducing the order by eliminating some terms in the expansion c)replacing some of the vectors in by different ones . assuming that the signal space is an inner product space , for @xmath11 given in to be the best approximation of a signal @xmath8 in a minimum distance sense , the coefficients in should be calculated in such a way that @xmath11 is the orthogonal projection of @xmath8 onto @xmath12 . this is the main reason for the popularity of orthogonal projectors in the context of approximation techniques . nevertheless , suppose that the observed signal is produced by the interference of two phenomena so that the model becomes [ eq2 ] _ i=1^k c_i v_i + _ i=1^n d_i w_i . if one were interested in discriminating the phenomena by splitting the signal , the component in @xmath12 could be obtained by an oblique projection operation mapping the other component to zero . there is a broad range of applications in which this procedure happens to be of assistance @xcite . thus , we felt motivated to find recursive equations for adapting oblique projectors . some of the equations to be proposed here are inspired by our previous work on recursive biorthogonalization for orthogonal projectors representation @xcite . we have recently been made aware that such a work is closely related to earlier one on recursive generalized inverses @xcite . in spite of the fact that for most numerical implementations a projector is represented by a matrix , we prefer to think of projectors as operators acting by performing inner products . an important reason for this choice is the following : the equations can thereby be applied in general inner product spaces and comprise two very important cases in particular . namely , the euclidean inner product space , where a projector is indeed a matrix , and the space of functions of finite 2-norm . we like to see the proposed recursive equations as generalized gram - schmidt like procedures for generating sequences in inner product spaces . such sequences give rise to oblique projectors onto nested subspaces and , of course , to orthogonal projectors as special case . the paper is organized as follows : section [ sec2 ] introduces the notation along with a discussion on the general construction of oblique projectors . section [ sec3 ] provides the recursive equations for stepwise updating / downdating of such projectors . applications are illustrated in section [ num ] by i)recovering a simulated x - ray diffraction peak from a background and ii)filtering impulsive noise from the register of the motion of a system consisting of the superposition of damped harmonic oscillators . the conclusions are drawn in section [ con ] . as already mentioned we will work in a general inner product space @xmath13 , where the square norm @xmath14 is induced by the inner product that we represent as @xmath15 . given two closed subspaces , @xmath16 and @xmath17 , such that @xmath18 and @xmath19 , the _ oblique projector _ operator onto @xmath20 along @xmath21 will be represented as @xmath22 . then @xmath22 satisfies : ^2&= & + v&= & v , v + w&= & 0 , w . in the particular situation in which @xmath21 happens to be the orthogonal complement of @xmath20 in @xmath13 , i.e. if @xmath23 , the operator is self - adjoint and represents an orthogonal projection onto @xmath20 . we emphasize this special case by using the particular notation @xmath24 . in the sequel the orthogonal projector operator onto a subspace , say the subspace @xmath25 , will be indicated as @xmath26 . let us assume that in general @xmath27 and @xmath28 , with @xmath29 the orthogonal complement of @xmath21 . denoting as @xmath30 the standard orthonormal basis in @xmath31 , i.e. , the inner product @xmath32 with @xmath33 equal one if @xmath34 and zero otherwise , we define the operators @xmath35 and @xmath36 as @xmath37 thus the corresponding adjoint operators @xmath38 and @xmath39 are @xmath40 the operations @xmath41 and @xmath42 indicate that @xmath39 and @xmath38 act by performing inner products in @xmath13 . the inner product is defined in such a way that for @xmath43 and @xmath44 a complex constant the mapping @xmath45 produces a vector in @xmath31 of the form @xmath46 . the operation @xmath47 indicates the inner product in @xmath31 , thereby for @xmath48 the mapping @xmath49 yields a vector in @xmath20 of the form @xmath50 note that the matrix representation of @xmath51 has elements given by the inner products @xmath52 . the operator @xmath53 where @xmath54 denotes the moore - penrose pseudo - inverse , is known to be the oblique projector onto @xmath20 along @xmath21 @xcite . the particular choice [ udef ] u_i = v_i - _ v_i= _ v_i , i=1, ,k produces the expression for @xmath22 used in signal processing applications @xcite . certainly , setting @xmath55 one has the convenient equation [ fo2 ] = ( ^ _ ) ^^ _ that we adopt hereafter . amongst the many properties of oblique projectors that have been studied we shall recall only the basic property needed for our purpose . it follows by applying @xmath56 on both sides of , i.e. , [ pro ] _ = _ ( ^ _ ) ^^_. since @xmath57 and @xmath58 , with the substitution @xmath59 the right hand side of takes the form @xmath60 . such an operator is the orthogonal projector onto @xmath61 . consequently , [ pro2 ] _ = _ , = ( _ ) = \ { _ v_i}_i=1^k . by denoting @xmath62 with @xmath63 the element @xmath64 of matrix @xmath65 , we can express @xmath22 as [ evw ] = _ i=1^k v_i u_i , . furthermore , from , , and [ euw ] _ = _ = _ i=1^k u_i u_i , . because @xmath56 is self - adjoint @xmath66 , and vice versa . on comparing and we see that the dual vectors @xmath67 are the same . this is of enormous assistance to derive the equations for adapting oblique projectors so as to account for the updating or downdating of the projecting subspace @xmath20 . this will allow us to give the proofs of the proposed recursive equations either by verification or by induction . it is appropriate to stress at this point that if we chose @xmath23 we would have @xmath68 and consequently @xmath69 . hence for such special situation @xmath70 and all the recursive equations of the subsequent sections would give rise to orthogonal projectors . in this section we provide the equations for updating and downdating oblique projectors in order to account for the following situations : let us consider that the oblique projector @xmath71 onto the subspace @xmath72 along a given subspace @xmath21 is known . if the subspace @xmath73 is enlarged to @xmath74 by the inclusion of one element , i.e. , @xmath75 , we wish to construct @xmath76 from the availability of @xmath71 . on the other hand , if the subspace @xmath77 is reduced by the elimination of one element , say the @xmath78-th one , we wish to construct the corresponding oblique projector @xmath79 from the knowledge of @xmath71 . the subspace @xmath21 is assumed to be fixed . its orthogonal complement @xmath80 in @xmath81 changes with the index @xmath10 to satisfy @xmath82 , where @xmath83 denotes an orthogonal sum whilst the former is a direct sum , i.e. , @xmath84 . we assume that @xmath71 is known and write it in the explicit form [ evkw ] = _ i=1^k v_i u^k_i , . our aim is to find the vector @xmath85 , and to change the vectors @xmath86 to @xmath87 , so as to obtain [ evkwp ] = _ i=1^k+1 v_i u^k+1_i , . we will show that the duals @xmath88 can be constructed inductively from the dual of a single vector . [ lem1 ] for @xmath89 , with @xmath90 , operator @xmath91 is the oblique projector onto the span of the single vector @xmath92 along @xmath21 . from the definition of @xmath93 ( c.f . it follows that the operator @xmath94 maps every vector in @xmath21 to the zero vector . suppose that @xmath8 is in the span of @xmath92 . then @xmath95 for some constant @xmath44 . since @xmath96 we have @xmath97 moreover @xmath98 , which concludes the proof that @xmath91 is the oblique projector onto the span of @xmath92 along @xmath21 . in order to inductively construct from @xmath99 the duals @xmath100 we have to discriminate two possibilities * @xmath101 , i.e. , @xmath102 * @xmath103 , i.e. @xmath104 let us consider first the case i ) . clearly if @xmath105 the corresponding @xmath106 belongs to @xmath107 , because @xmath108 yields @xmath109 . the proposition below prescribes how to modify the corresponding dual vectors in order to guarantee that @xmath110 . let @xmath105 and vectors @xmath111 in be given . for an arbitrary vector @xmath112 the dual vectors @xmath113 computed as [ utikp ] u_i^k+1= u_i^k- u_k+1 , u_i^ky_k+1 for @xmath114 and @xmath115 produce the identical oblique projector as the dual vectors @xmath116 . we use to explicitly express @xmath76 [ red ] _ i=1^k+1 v_i u_i^k+1 , & = & _ i=1^k v_i u_i^k , , - _ i=1^k v_i u_i^k , u_k+1 y_k+1 , + v_k+1 y_k+1 , + & = & - u_k+1 y_k+1 , + v_k+1 y_k+1 , + & = & - v_k+1 y_k+1 , + v_k+1 y_k+1 , , where the last equality holds because @xmath117 and @xmath118 for @xmath105 . hence , the left hand side of equals @xmath71 . the next proposition considers the case ii ) let vector @xmath119 and vectors @xmath120 in be given . thus the dual vectors @xmath113 computed as [ eq ] u_i^k+1=u_i^k- u_k+1^k+1 u_k+1 , u_i^k , where @xmath121 with @xmath122 , provide us with the oblique projector @xmath76 . in order to organize the proof let us establish the following relations : [ a1 ] q_k+1,v_i&=&0 , i=1, ,k + [ a2 ] q_k+1,v_k+1&=&u_k+1 , v_k+1- v_k+1 , _ v_k+1 = \|\| q_k+1\|\|^2 . the first relation follows from the definition of @xmath123 and the fact that @xmath124 for @xmath114 @xmath125 on the other hand @xmath126 furthermore \|\|q_k+1\|\|^2&= & q_k+1,u_k+1 - q_k+1 , _ u_k+1 = q_k+1 , u_k+1 + & = & q_k+1 , v_k+1 - q_k+1 , _ v_k+1= q_k+1 , v_k+1 + & = & u_k+1 , v_k+1 - v_k+1 , _ v_k+1 . we are now in a position to start the proof of the proposition by induction . from lemma [ lem1 ] we know that @xmath127 is the oblique projector onto @xmath128 along @xmath21 . assuming that @xmath71 is the oblique projector onto @xmath73 along @xmath21 we will prove that @xmath76 is the oblique projector onto @xmath74 along @xmath21 . for this we need to prove that the recursive equation yields the operator @xmath76 satisfying : * @xmath129 * @xmath130 * @xmath131 we begin by using to express @xmath76 as [ obkp ] v_i u_i^k+1 , & = & v_i u_i^k , , - v_i u_i^k , u_k+1 u_k+1^k+1 , + v_k+1 u_k+1^k+1 , + & = & - u_k+1,+ v_k+1 , . for all @xmath132 in @xmath21 it holds that @xmath133 and @xmath134 . then from we conclude that condition iii ) is satisfied . every @xmath135 can be written as @xmath136 . thus , from and using relations and v & = & _ i=1^k c_i v_i + c_k+1v_k+1 - c_k+1 u_k+1 + c_k+1 v_k+1 + & = & _ i=1^k c_i v_i + c_k+1 v_k+1= v , which demonstrates condition ii ) . finally , since from and it follows that @xmath137 , we have ^2 & = & - u_k+1 , + v_k+1 , + & = & - u_k+1 , + v_k+1 , = . [ re1 ] if vectors @xmath138 are linearly independent they are also biorthogonal to the dual vectors arising inductively from the recursive equation . the proof of this property is given in appendix a. [ re2 ] if vectors @xmath138 are not linearly independent the oblique projector @xmath71 is not unique . indeed , if @xmath139 are dual vectors giving rise to @xmath71 then one can construct infinitely many duals as : [ nonu ] y_i= u_i^k+ y_i - _ j=1^k y_j v_j , u_i^ki=1, ,k , where @xmath140 are arbitrary vectors in @xmath13 . we use to write v_i y_i , & = & + v_i y_i , - v_i _ j=1^k u_i^k , v_j y_j , + & = & + v_i y_i , - _ j=1^kv_i u_i^k , v_j y_j , + & = & + v_i y_i , - _ j=1^k v_j y_j , + & = & . it follows from property [ re1 ] that if vectors @xmath138 are linearly independent equation yields the unique duals @xmath141 . suppose that by the elimination of the element @xmath78 the subspace @xmath73 is reduced to @xmath142 . in order to give the equations for adapting the corresponding dual vectors generating the oblique projector @xmath79 we need to consider two situations : * @xmath143 i.e. , @xmath144 * @xmath145 , i.e. , @xmath146 the next proposition addresses i ) . let @xmath71 be given by and let us assume that removing vector @xmath147 from the spanning set of @xmath73 leaves the identical subspace , i.e. , @xmath148 . hence , if the remaining dual vectors are modified as follows : [ opeqb ] = u_i^k+ , the corresponding oblique projector does not change , i.e. @xmath149 . let us recall that @xmath148 implies @xmath150 . hence @xmath151 , as it is seen from the fact that @xmath152 , which implies @xmath153 if and only if for @xmath114 it holds that @xmath154 . this is not true if @xmath148 so that we can use to express @xmath79 as _ ^k v_i , & = & _ ^k v_i u_i^k , + _ ^k + & = & - v_j u_j^k,+ - v_j u_j^k , u_j + & = & - v_j u_j^k,+ - u_j^k , u_j + & = & . finally , proposition 4 addresses ii ) . let @xmath71 be given by and let us assume that the vector @xmath147 to be removed from the spanning set of @xmath73 is not in @xmath155 . in order to produce the oblique projector @xmath79 the appropriate modification of the dual vectors can be achieved by means of the following equation [ duba ] = u_i^k- . using we write : [ equp ] _ ^k v_i , & = & _ ^k v_i u_i^k , - _ ^k + & = & - v_j u_j^k , - + v_j u_j^k , + & = & - . we notice that @xmath156 is the orthogonal projector onto the span of the single vector @xmath157 and denote it as @xmath158 . thus the orthogonal projector onto @xmath159 can be expressed as @xmath160 . applying this operator on the right hand side of we obtain : @xmath161 and , since @xmath162 and @xmath163 , using the fact that @xmath164 it follows that ( _ - _ u^k_j ) ( -_u^k_j ) & = & _ - _ u^k_j - _ u^k_j + _ u^k_j + & = & _ - _ u^k_j = _ _ = _ _ . from the last equation and we gather that _ _ _ ^k v_i , - _ _ = 0 . for every vector @xmath165 we therefore have _ _ ( _ ^k v_i , - ) f = _ _ f = 0 , with @xmath166 . this implies that either @xmath167 is the zero operator or @xmath168 for every @xmath43 . the latter can not be true because from the definition of @xmath167 it is seen that @xmath169 and by hypothesis @xmath170 . hence @xmath167 should be the zero operator , which leads to the conclusion that = _ ^k v_i , . [ re3 ] the case of replacing a vector in @xmath171 , say @xmath147 by @xmath172 , is actually equivalent to augmenting the subspace @xmath173 to @xmath174 after the vector @xmath147 was deleted . some implementation issues arise , though . in order to modify the duals as prescribed in we need to compute a vector @xmath175 . for the sequential enlargement of the projecting subspace , discussed in section [ secup ] , the projector @xmath176 can be sequentially constructed by means of the orthonormal vectors @xmath177 . nevertheless , when replacing vectors sequentially we need to allow for the recalculation of the corresponding projectors . one possibility that could be considered is the recalculation of the orthogonal vectors @xmath178 @xcite . an alternative approach implies to use of the dual corresponding to the deleted vector for orthogonalization purposes . a discussion concerning the implementation of such a procedure is given in @xcite . the examples presented in this section aim at illustrating the application of our recursive construction of oblique projectors for signals in @xmath179 $ ] , the space of square integrable functions on @xmath180 $ ] . for @xmath8 and @xmath181 in @xmath179 $ ] , we define the inner product , according to the previously adopted convention , as @xmath182 where @xmath183 indicates the complex conjugate of @xmath8 . in the examples bellow all the integrals are numerically calculated . here the signal is simulated by emulating a crystallographic problem . it is assumed to be the x - ray diffraction intensity produced by a powder sample of a clay mineral consisting of very flat crystals . each such crystal is formed by the stacking of @xmath184 layers producing a diffraction intensity as given by @xcite i_n(x ) = , where the variable @xmath185 , given in radians , is related to the diffraction angle @xmath186 according to the equation x= 2 . [ xt ] the parameter @xmath187 in is the effective distance between two consecutive layers and characterizes the material . the parameter @xmath188 represents the wavelength of the incident radiation . we denote the diffraction intensity produced by the whole sample as @xmath189 . thus , [ smod ] f_1(x)= _ n=1^k c_n . as already mentioned @xmath184 indicates the possible number of layers forming a single crystal in the sample . the coefficients @xmath190 account for the proportions of crystals consisting of @xmath184 layers . here , for simulating the signal , the coefficients were considered to be @xmath191 the diffraction figure @xmath192 emerges from a background that is modelled as f_2(x)= 50 _ j=1 ^ 3 j e^-j(x- ) . the combined phenomenon gives rise to the signal @xmath193 plotted in the left graph of figure 1 on the interval relevant to the diffraction model , namely @xmath194.$ ] we are interested in extracting the diffraction peak @xmath192 from the background . for this we will construct sequentially oblique projectors onto subspaces @xmath73 given as @xmath195.\ ] ] the final @xmath10-value is to be adjusted . the subspace @xmath21 is here @xmath196.\ ] ] since the order @xmath10 of the diffraction model is assumed unknown , it was adjusted as follows : firstly the order model was sequentially increased ( up to @xmath197 ) and then sequentially downdated . it was observed that the recovering of the signal was not very sensitive to the model order . in a range from @xmath198 to @xmath197 the approximations were totally equivalent . from @xmath199 to @xmath198 changes in the approximations were noticed but the approximations could still be considered ` practically ' equivalent . the recovered peak @xmath200 ( for @xmath198 ) is depicted in the right graph of figure 1 . it happens to coincide , in the scale of the figure , with the graph of the theoretical one ( c.f . ) . the convenience of the proposed adaptive technique in the determination of the order of the diffraction intensity model is clear : otherwise when changing the @xmath10 value as described above the whole projector would have to be recalculated for each different value of @xmath10 . however , the advantage of the proposed technique is even more significant when , as is the case in the next example , overestimation of the order in the signal model may result in the failure to discriminate the signals . [ f1 ] measured in radians and related to the diffraction angle through . graph on the right : diffraction intensity extracted from the background.,title="fig:",width=283 ] measured in radians and related to the diffraction angle through . graph on the right : diffraction intensity extracted from the background.,title="fig:",width=283 ] in this case the signal is considered to be the register of the motion of a system consisting of uncoupled damped harmonic oscillators . the @xmath184-th oscillator is characterized by a frequency of @xmath201 hz . the corresponding equation for its motion @xmath202 as a function of time is given as x_n(t)= e^-t ( n t ) , n=1, ,k . the distribution of frequencies is considered to be @xmath203 so that the motion of the system is registered by the signal [ smod2 ] f_1(t)=_n=1 ^ 100 , t . this signal ( shown in the right graphs of figure 2 ) is corrupted by impulsive noise , which represents a type of electrical noise appearing in some practical situations . the possible pulses are taking from the set of @xmath204 gaussian sparks @xmath205 . hence the corresponding subspace @xmath21 is @xmath206.\ ] ] first a random superposition of @xmath207 pulses is added to the signal @xmath192 to simulate the noisy one plotted in the top left graph of figure 2 . the result after the oblique projections along the subspace @xmath21 given above is shown in the top right graph ( it coincides with the theoretical signal @xmath192 given in ) . the left graph at the bottom of figure 2 corresponds to a different realization of the noise , in this case generated as a random superposition of @xmath208 pulses . the signal , after filtering by oblique projections along @xmath21 , is shown in the right graph ( it also coincides with the theoretical one ) . [ f2 ] + + let us point out that an alternative way of splitting the signal would entail fitting both the signal and noise models , with the consequent increment in the dimension of the problem of determining the corresponding unknown parameters . in the example of this section , for instance , 400 more parameters ( coefficients of the noise model ) would be involved . recursive equations for updating / downdating oblique projectors have been proposed . the updating strategy can be regarded as a generalized gram - schmidt like procedure for generating a sequence giving rise to oblique projectors along a fixed subspace . the downdating strategy modifies such sequence to account for the removal of some elements . the equations are of the same nature as those for producing orthogonal projectors , but involve different vectors . orthogonal projectors arise within this framework as a particular case . the proposed technique has been applied to the problem of discriminating signals produced by different phenomena . the applications that have been considered assume that the signal model is determined by physical considerations and only the order of the model is to be adjusted . the task of setting the right order model is facilitated by the recursive nature of the proposed equations . in the two examples considered here the signal splitting is not very sensitive to the order of the signal model . however , an important difference between the two examples is the following : while in the first example an excessive overestimation of order model ( maximum possible number of layers present in a crystal ) does not prevent the extraction of the diffraction peak from the background , an excessive overestimation of the order of the model in the second example ( maximum possible frequency of an oscillator ) may produce the failure to separate the signal from the impulsive noise . the reason being that for very high frequencies the angle between the signal subspace and the noise subspace becomes very small , which generates an ill posed problem . the recursive feature of the proposed equations turns out to be even more important in those situations in which the signals splitting is achieved by stepwise selection of each component of the signal model . this is the subject of a recent work @xcite , where the present approach is shown to be of significant assistance . support from epsrc ( ep@xmath209d062632@xmath2091 ) is acknowledged . [ apen ] let us recall that if vectors @xmath210 are linearly independent , all the duals @xmath211 are generated by the recursive equation . we need to show that such vectors satisfy : @xmath212 * @xmath220 and @xmath221 + in this case @xmath222 ( c.f . ) . hence , from the recursive equation , we have @xmath223 * @xmath224 and @xmath225 + now v_l+1 , u^l+1_i & = & v_l+1 , u^l_i - v_l+1 , u^l_i + & = & v_l+1 , u^l_i so that , since @xmath226 ( c.f . ) , @xmath227 * @xmath224 and @xmath228 + this implies @xmath229 * @xmath220 and @xmath230 + in this case @xmath231 l. rebollo - neira , recursive bi - orthogonalisation approach and orthogonal projectors , math - ph/0209026 ( 2002 ) ; l. rebollo - neira , new topics in mathematical physics research , nova science publisher , new york , ( 2006 ) , ch . on non - orthogonal signal representation . m. andrle , l. rebollo - neira , experiments on orthogonalization by biorthogonal representations of orthogonal projectors , journal of computational and applied mathematics , in press , corrected proof available on line ( 2006 ) .	a generalization of the gram - schmidt procedure is achieved by providing equations for updating and downdating oblique projectors . the work is motivated by the problem of adaptive signal representation outside the orthogonal basis setting . the proposed techniques are shown to be relevant to the problem of discriminating signals produced by different phenomena when the order of the signal model needs to be adjusted . = 1
the existence of a minimum stable orbit for material around a black hole is a key prediction of general relativity in the strong field limit . for schwarzschild black holes this is at @xmath0 ( where @xmath1 ) , while for maximally rotating kerr black holes ( spin parameter @xmath2 ) this reduces to @xmath3 . bright accretion flows give a way to observationally test such ideas as well as to directly estimate the spin of the black hole . optically thick material in keplerian orbits should emit a quasi - thermal spectrum with maximum temperature emission from the largest luminosity / smallest area regions i.e. set by the last stable orbit ( shakura & sunyaev 1973 ) . as these models are thermal , there is a clear prediction that their maximum temperature , @xmath4 , should increase with total disc luminosity as @xmath5 , if the emission is from a constant area i.e. if there is a constant inner disc radius . strong observational support for this comes from the galactic black hole ( gbh ) binary systems . these can show disc dominated spectra ( often termed the high / soft state ) where @xmath5 over large changes in luminosity ( ebisawa et al . 1993 ; kubota , makishima & ebisawa 2001 ; kubota & makishima 2004 ) . the fixed radius inferred from the proportionality constant of this relation is consistent with a keplerian disc ( stress - free inner boundary condition ) in a low to moderate spin spacetime ( @xmath6 ) for all gbh for which this experiment can be performed ( gierliski & done 2004 ; li et al . 2005 ; shafee et al . 2005 ; davis , done & blaes 2005 ) . this ties in with the spins expected from models of stellar collapse . these predict a maximum birth spin of @xmath7 from the combination of slow rotation of the pre - supernovae core and angular momentum losses through gravitational wave radiation during collapse ( e.g. gammie , shapiro & mckinney 2004 ) , and there is not enough mass transfer in the binary evolution to cause appreciable spin up from accretion ( king & kolb 1996 ) . however , there is growing evidence that the accretion disc inner radius can also be _ larger _ than @xmath8 , due to the inner disc evaporating into hot , geometrically thick accretion flow . these models were developed in response to observations of gbh at low luminosity ( generally termed the low / hard state ) , where the expected accretion disc emission is much weaker and at lower temperature than predicted by an optically thick disc extending down to the last stable orbit ( e.g esin et al . 2001 ; frontera et al . 2001 ) . instead , the x - ray spectra peak at @xmath9100 kev ( see e.g. the review by zdziarski & gierliski 2004 ) , requiring that the continuum is formed in hot , optically thin plasma , which is not strongly cooled by compton scattering of seed photons ( e.g. zdziarski & gierliski 2004 ) . these cooling limits are very difficult ( barrio , done & nayakshin 2003 ; though not impossible e.g. beloborodov 1999 ) to reconcile with a cool disc extending down to the minimum stable orbit . other observational advantages of truncated disc configuration is that it can provide a geometric explanation for the major x - ray spectral state transitions seen in gbh by a physical change in the accretion flow structure . at low luminosities the inner cool disc is replaced by a hot accretion flow . the transition radius between the two phases is far from the hole , few photons from the disc illuminate the hot flow and the spectrum is hard . as the transition radius moves inwards the disc penetrates further into the hot flow , so more seed photons cool the flow and its spectrum softens until the disc extends down to the last stable orbit ( esin et al . this changing transition radius also gives a changing size scale to associate with the changing characteristic frequencies seen in the variability power spectra ( see e.g. the review by done 2001 ) . theoretically , such hot inner flows are generically produced by a range of analytic approximations to the accretion flow equations . in these , the protons are heated by gravity , while the electrons gain energy only through coulomb collisions and cool predominantly via compton scattering ( shapiro , lightman & eardley 1975 ) , advection ( narayan & yi 1995 ) , outflows ( blandford & begelman 1999 ) , convection ( abramowicz & igumenshchev 2001 ) or jets ( falcke , krding & markoff 2004 ) . numerical simulations of the accretion flow at low densities , including the self consistently generated magnetic dynamo viscosity ( balbus & hawley 1991 ) , appear to show some form of inner hot flow , though its properties are more complex than any of the approximate analytic solutions above ( hawley & balbus 2002 ) . another argument for the truncated disc comes from the quasi - periodic oscillations ( qpos ) observed in the power density spectra of black hole binaries . during the transition from the low / hard to the high / soft spectral state their frequency typically increases ( e.g. cui et al . 1999 ; rossi et al . 2004 ) indicating a decreasing size - scale in the accretion flow . if the qpo frequency is directly related to the transition radius between the cold disc and the hot inner flow ( stella & vietri 1998 ; titarchuk , osherovich & kuznetsov 1999 ; psaltis & norman 2000 ) then its behaviour is consistent with a decreasing truncation radius as the source approaches the soft state . clearly , the point at which the cool disc truncates is of great importance to our understanding of the nature of the accretion flow , and hence whether we can use observations of compact objects to test strong gravity . fortunately there is an independent constraint on this from the shape of the fluorescent iron line produced by x - ray illumination of the cool disc . special and general relativistic effects are both important as the disc material is in very high velocity ( keplerian ) orbits in a strong gravitational field . all these effects get stronger as the disc extends closer to the black hole , so the width of the line gives a measure of the inner edge of the cool fluorescing material ( fabian et al . 1989 ; 2000 ) . thus for the accretion flow model above in which there is a variable truncation radius between a disc and inner hot flow , the line width should be correlated with spectral hardness . hard spectra ( generally seen at low @xmath10 ) should have narrow lines , softer spectra should have broader lines . there are indications that this correlation is seen both in agn ( lubiski & zdziarski 2001 ) and in gbh ( ycki , done & smith 1998 ; churazov , gilfanov & revnivtsev 2000 ; ibragimov et al . 2005 ) . however , some of the more recent results from the line shape seen in the gbh do not fit easily into this picture . there are reported lines _ in the low / hard state _ ( i.e hard x - ray spectra , with qpo frequencies which are not at their maximum values ) which are so broad as to require a disc around a maximally spinning black hole , illuminated by highly centrally concentrated radiation field , perhaps implying direct tapping of the spin energy of the black hole or at least substantial continuous stress across the last stable orbit ( miniutti , fabian & miller 2004 ; miller et al . 2002 ; miller et al . plainly this would rule out the inner hot flow model if there is no alternative explanation for the line width . the clearest example of this discrepancy is the _ bepposax _ data of the bright low / hard state of xte j1650500 . the line width in this spectrum is also in conflict with the moderate spin derived from the luminosity - temperature relation obtained when this object shows disc dominated spectra ( gierliski & done 2004 ) . thus the extent of the disc inferred from the iron line width in xte j1650500 seems at odds with the broad band spectral shape , the qpos , and the disc continuum emission . here we re - examine the _ bepposax _ data on xte j1650500 and show that while these data are indeed consistent with an extreme relativistic line , they can be fit equally well ( or better ) using a model of a ` normal ' relativistic line ( i.e. from a truncated disc with standard gravitational energy release ) , together with blueshifted highly ionized narrow absorption features ( predominantly resonance lines of he- and h - like iron ) . these absorption features distort the blue edge of the line , especially where the reflection spectrum is ionized and/or from high inclination material . if the presence of such material is confirmed then this removes the _ only _ observational evidence against a truncated disc interpretation of the low / hard state in gbh . we extract the xte j1650500 _ bepposax _ archival data taken on 2001 september 11 for both the medium - energy concentrator spectrometer ( mecs ) and phoswich detector system ( pds ) instruments , using the standard ( september 1997 ) background , auxiliary response ( mecs ) and redistribution files ( mecs and pds ) . we use the same energy ranges as miniutti et al . ( 2004 ) , i.e. 2.510 kev ( mecs ) and 13200 kev ( pds ) , with 2 per cent systematic errors added in quadrature to the pds data . we checked the publicly available crab pds data and obtain reduced @xmath11 for a single power law of energy index 1.13 , showing that this prescription is an adequate description of the response uncertainties . = 0.8 we also extract the _ rxte _ proportional counter array ( pca ) and high - energy x - ray timing experiment ( hexte ) spectra from the entire outburst . we calculate the x - ray colours and estimate the bolometric luminosity following the method of done & gierliski ( 2003 ) . we used the distance of 4 kpc ( tomsick et al . 2003 ) and mass of 7.3 m@xmath12 ( orosz et al . 2004 ) for the luminosity estimate . we also extracted the power - density spectra from pca observations in the 1/128 to 64 hz frequency band and fitted them by a multiple lorentzian model . we use xspec version 11.3.1 ( arnaud 1996 ) for spectral fitting . the error of each model parameter is given for a 90 per cent confidence interval . we fix the absorption column at @xmath13 @xmath14 in all fits ( miller et al . 2002 ; miniutti et al . [ fig : xte ] shows the estimated bolometric luminosity , the frequency of the strong low - frequency qpo and x - ray colours obtained from the _ rxte_. the vertical line marks the time of the _ bepposax _ observation analysed in this paper . the qpo frequency during the _ bepposax _ observation is plainly not the maximum seen from this source . any association of the qpo with the inner radius of the cool disc clearly requires that the disc does _ not _ extend down to the last stable orbit , but truncates _ above _ this point . the x - ray colours show that the _ bepposax _ data were taken while the source is still smoothly softening , in a position which is associated with the softest low / hard states . given that the spectrum softens further _ after _ the _ bepposax _ observation , models of the broad band spectra with a moving inner disc predict that the disc truncates _ above _ the last stable orbit . thus both the qpos and broad band spectral shape indicate that the innermost parts of a cool disc are not present . this is in conflict with the extreme iron line smearing _ in these same data _ which requires that the disc extends down to the last stable orbit of a high spin black hole . with the mecs data we can reproduce the extreme relativistic broadening of miniutti et al . we use their phenomenological model of a smeared edge and iron line with extreme kerr spacetime distortions ( the laor model in xspec ) on a continuum consisting of a disc spectrum and power law . this gives an adequate fit ( @xmath16 ) to the data , and the parameters of the strong iron line ( equivalent width of @xmath17 ev ) do indeed point to a very small inner disc radius of @xmath18 with extreme central concentration of the illumination , @xmath19 where @xmath20 . a similarly extreme line is seen using the same model in the _ xmm - newton _ data taken a few days after the _ bepposax _ observation ( miller et al . 2002 ) . the xspec laor code is written for the specific case of calculating the relativistic effects on a line which can be approximated as an initial delta function . this is identical to the transfer function , so relativistic effects on an arbitrary initial spectrum can be calculated by simply convolving the initial spectrum with this transfer function . we recoded the laor model into a convolution model ( hereafter called conline ) in xspec and use this to smear a full reflected spectrum ( the pexriv model in xspec with additional narrow gaussian line ) to replace the phenomenological smeared edge / line features in the previous fit . convolution means that the spectrum outside of the observed energy band can be important so we always extend the energy range used by xspec to 0.250 kev . we follow miller et al . ( 2002 ) and miniutti et al . ( 2004 ) and fix the inclination angle to @xmath15 and assume solar abundances . we constrain the radial emissivity index to be between 3 and 5 as suggested by these previous fits . this gives @xmath21 , i.e. slightly worse than the phenomenological approximation for reflected continuum emission . the best - fitting model is reflection dominated , with @xmath22 , i.e. for a model in which the intrinsic continuum is hidden from view . however , this is poorly constrained , with the error range allowing @xmath23 . by contrast the line parameters are fairly well constrained , with equivalent width of @xmath24 ev , and inner disc radius of @xmath25 and @xmath26 . thus the line parameters do change with the addition of a reflected continuum , most significantly in reducing the equivalent width , although the smearing is also affected , with the illumination being not quite so centrally concentrated . detailed results on the line equivalent width ( and to a lesser extent the shape ) derived without including a reflected continuum should be treated with caution . nonetheless , the small disc inner radius is very similar to that in the phenomenological fits ; there is a very broad feature in the data between 47 kev as shown in fig . [ fig : mecs]a . the reflected component from the pexriv fits in the previous section is strongly ionized . this means that the intrinsic iron line emission from the disc is _ not _ narrow . it should instead be strongly comptonized , and this comptonization should also affect the edge in the reflected continuum ( ross , fabian & young 1999 ) . the neglect of compton scattering below 12 kev in pexriv means that this code has an edge which which is artificially deep and sharp for a highly ionized disc . we replace the pexriv reflected continuum and narrow gaussian line with reflected spectra of d. ballantyne from a constant density ionized disc ( hereafter cdid ; ballantyne , iwasawa & fabian 2001 , based on ross & fabian 1993 ) which are publicly available as xspec table models . these calculate the reflected spectrum from an x - ray illuminated slab of constant density , including compton up- and down - scattering below 12 kev , with self consistent iron ( and other elements ) line calculation for solar abundance material . these give @xmath27 for a highly smeared , ionized slab , somewhat worse than the ( unphysical ) fits above , but with two fewer free parameters as the line energy and width are now calculated self - consistently with the reflected continuum . the reflection emission is highly ionized , @xmath28erg cm s@xmath29 , and again the spectrum is consistent with pure reflected continuum , although the errors are wide and only restrict @xmath30 . the best fit again picks out extreme relativistic smearing , with @xmath31 and @xmath32 . however , the significance of detection of relativistic smearing is reduced with respect to the reflection models which do not include comptonization of the line and edge features . the pexriv based reflection models give @xmath33 for removing all smearing from the model , while the cdid reflection models give @xmath34 . = 0.98 figs . [ fig : mecs](b ) and ( c ) show these fits with extreme relativistic smearing , and no relativistic smearing , respectively . it is clear that the _ intrinsic _ reflected emission alone mainly the edge in the reflected continuum rather than the line itself , as there is very little line emission at these high ionizations can match most of the broad feature in the data . this can have a strong impact on the detection significance of relativistic smearing and some impact on the detailed parameters . plainly it is important to use an accurate description of ionized reflection . however , there is considerable uncertainty in these models as they depend on the ( unknown ) vertical and radial ionization structure of the disc . the cdid models assume that the reflecting material can be approximated by a slab of constant ( vertical and radial ) density . yet a disc is in hydrostatic equilibrium so its structure _ responds _ to the incident radiation ( nayakshin , kazanas & kallman 2000 ; ballantyne , ross & fabian 2001 ) . the upper layers are heated to the compton temperature of the illuminating flux , so they expand , becoming less dense and more ionized . for hard power law illumination the compton temperature is large , so the upper layers are completely ionized , forming a skin on top of the disc , but for softer illuminating spectra ( such as that of the data used here ) the upper layers of the disc are still visible so the difference between the hydrostatic and constant ionization reflection models is much smaller ( done & nayakshin 2001 ) . as well as this vertical structure , there should also be a _ radial _ dependance of the ionization structure of the disc , and the full reflected emission should be the emissivity weighted integral of all of these different radii regions with vertical ionization structure . we use the xion code ( nayakshin et al . 2000 ) to model reflection from a _ disc _ of material in hydrostatic equilibrium from power law illumination , i.e. including both radial and vertical ionization gradients . this can be used to produce the total emission ( direct plus reflected ) , in which the contribution of reflection is hardwired into the spectrum for a given geometry . however , in order to give a comparable number of degrees of freedom as in the cdid fits , we use xion to calculate the reflected emission _ only _ , and add a power law as the incident continuum so that the fraction @xmath35 is allowed to vary . we use the magnetic flares geometry to calculate the reflected emission from 10100 schwarzschild radii from a disc accreting at @xmath36 ( see fig . [ fig : xte ] ) . the xion code includes relativistic smearing , but only for non - rotating black holes , so we switch this off and use the conline convolution model as before . this gives a somewhat worse fit than before , with @xmath37 . a better fit can be obtained by setting @xmath38 , with @xmath39 . this is due to the higher disc density expected at higher mass accretion rates . both models give @xmath40 and @xmath41 and again the best fit is reflection dominated . the inclination of the system is is not well known , but recent studies of the companion star indicate that its _ minimum _ value is @xmath42 ( orosz et al . 2004 ) , showing that the assumption of @xmath15 in the previous fits is probably too low . the filled squares in fig . [ fig : inc]a show the increase in @xmath43 with inclination for the fits for the cdid reflection models . plainly the lower inclination fits are preferred by the data . however , we note that there is no minimum in @xmath43 at @xmath15 as seen from pexriv based reflection models ( miller et al . 2002 ) , so removing the only reason for using this inclination . these fits become very unstable so we fix @xmath44 and assume that the continuum is purely reflected emission ( i.e. no intrinsic continuum visible ) in order to produce the plots . however , we checked that these assumptions do not qualitatively affect the results , unless we allow for extremely centrally concentrated emission . we can obtain good fits with @xmath45 at higher inclinations up to @xmath46 , but only with @xmath47 . at higher inclinations , the highly smeared solution becomes a worse fit to the data because of the increasing blueshift of the relativistic transfer function due to the larger projected line of sight velocities ( e.g. fabian et al . the fit compensates for this by increasing the gravitational redshift i.e. by decreasing the disc inner radius and/or by increasing @xmath48 . since this figure is made for constant @xmath44 then this requires that the inner radius decreases . however , this does not give enough gravitational redshift to compensate for the increased line broadening and most of the increase in @xmath43 comes from this increased blue extent of the spectral features . these extremely smeared fits then have a _ worse _ @xmath43 than the very different fits in which the line is mostly _ unsmeared _ , as described in the section above ( see fig . [ fig : mecs]c ) . this solution is mostly independent of inclination as the intrinsic shape of the reflected emission below 10 kev is independent of inclination ( magdziarz & zdziarski 1995 ) , as are the relativistic effects for large inner disc radii . because these two solutions , the extremely smeared and almost unsmeared , are far apart in parameter space , they give a double minimum in @xmath43 space . we show the @xmath43 for the mostly unsmeared fits as the open triangles in fig . [ fig : inc](a ) , and it is plain that these are a much better description of the data for inclinations above @xmath15 than the highly smeared fits ( filled squares ) . fig . [ fig : inc](b ) shows the corresponding disc inner radius inferred for each solution as a function of inclination . plainly , as the best fit switches between these two very different solutions , there is a corresponding dramatic jump in inferred disc inner radius . the residuals to the ( mostly ) unsmeared fit for an inclination of @xmath15 in fig . [ fig : mecs](c ) show a clear feature at @xmath9 7.5 kev ( the smeared fit in fig . [ fig : mecs]b also shows a similar feature ) . this looks like a fairly narrow absorption feature , and including a narrow , gaussian line ( with best - fitting energy @xmath49 kev and equivalent width of @xmath9100 ev ) improves the fit by @xmath50 , making it _ significantly better_. the relativistic smearing reduces _ dramatically _ , now being entirely consistent with @xmath44 , and with a larger disc inner radius , @xmath51 . = 0.47 motivated by the fits above , we consider how such absorption lines can occur . observationally , strong resonance absorption lines are often seen from high inclination galactic binary systems , generally from iron k@xmath52 he- and h - like iron at 6.7 and 7.0 kev ( e.g. ueda et al 1998 ; kotani et al . 2000 ; lee et al . 2002 ; ueda et al . however , if this absorbing material is outflowing then substantial blueshifts could be observed , and are seen in agn ( pounds et al . if the feature at @xmath53 kev is an absorption line from iron then this implies a velocity of @xmath90.09@xmath54 or 0.13@xmath54 depending on whether the identification is with the h or he - like ion . we use the xstar photoionization models which are publicly available as xspec multiplicative model tables . we choose the ones calculated for power law illumination of a constant density photoionized column with turbulent velocity of 100 km s@xmath55 . this gives an improved @xmath43 for _ all _ inclinations with the cdid reflection models ( see fig . [ fig : inc]c ) , and dramatically changes the inferred smearing of the reflected emission ( fig . [ fig : inc]d ) . the relativistic line is now _ always _ consistent with the illumination of @xmath44 expected from gravitational energy release , and the inner disc radius is greater than @xmath56 . the existence of several narrow features in the ionized absorber between 6.67 kev means that @xmath43 space is complex , with multiple local minima . however , the column in all fits is @xmath57 @xmath14 of ionized material ( @xmath58 erg cm s@xmath55 ) outflowing at @xmath90.10.15@xmath54 , with the best - fitting redshift of @xmath59 . this model for an inclination of @xmath15 is shown in fig . [ fig : mecs](d ) . this effect is only seen when using the newer models of reflection from ionized material , which include the self - consistent comptonization of the iron line and edge by the hot upper layers of the disc . [ fig : mod ] ( grey , or green in colour , line ) shows the best - fitting model using the pexriv reflected continuum and narrow gaussian line , convolved with the laor relativistic profile . the sharp drop at @xmath97.2 kev seen in the data can be matched by the sharp blue wing of the relativistically smeared , intrinsically _ narrow _ line . by contrast , the black line in fig . [ fig : mod ] shows the best - fitting model using the extremely smeared , highly ionized cdid reflection . the upper layer of the irradiated reflecting material is hot , so comptonizes much of the line and edge , blending them together and making an intrinsically rather smooth spectrum ( ross et al . 1999 ) . convolution of this spectrum with relativistic effects makes it even broader , so there is no sharp change in curvature produced by the blue wing of the line . since the data does have a sharp drop at 7.27.8 kev then these more physically self consistent models of reflection give a worse fit . the significant residuals in the spectrum at @xmath97.5 kev then are better fit by including outflowing absorption , predominantly resonance iron k lines at rest frame energies of 6.7 and 6.95 kev . the drop seen in the data then no longer drives the relativistic line profile to extreme parameters and the derived smearing is completely compatible with a truncated disc . = 0.35 there is a clear broad residual associated with iron in these _ bepposax _ data from xte j1650500 . simple models of reflection , which ignore comptonization of the line , require that the relativistic effects are extreme , implying a spinning black hole and perhaps direct extraction of the spin energy . however , these models also require that the reflected emission is highly ionized . this means that comptonization of the line and edge is important , making these features _ intrinsically _ broad and blended together . the strength and significance of relativistic effects are then determined by the detailed modelling of the intrinsic reflected emission , which in turn depends on the ( poorly known ) vertical and radial structure of the disc . nevertheless , the current best models of ionized reflection always pick out parameters where the innermost radius of the emitting material is much closer than @xmath56 , and with radial emissivity rather steeper than expected from gravitational energy release alone . however , the smearing derived from these ( best available ) reflection models are _ not _ robust to changes in inclination , nor to the addition of ionized absorption features . while either ( or both ) of these effects can resolve the discrepancy of the inner disc radius derived from the line shape with that inferred from the overall evolution of spectral and timing features , a further conflict remains which is that the fits favour a reflection dominated solution . to get more constraints on the amount of reflection requires data at higher energies , so we include the simultaneous pds data to extend the spectrum up to 200 kev . however , the spectrum is now much more sensitive to details of the continuum modelling due to the wide bandpass , and it is obvious that a power law can no longer be used . even an exponentially cutoff power law is not a good approximation to the shape of a thermally comptonized spectrum over such a wide bandpass . the rollover at energies close to the electron temperature , @xmath60 , is rather sharper than an exponential ( gierliski et al . 1997 ) . also , there are indications that the spectra can be more complex than predicted by just a single temperature thermal compton component , with signatures of both thermal and non - thermal electrons even in the low / hard state ( mcconnell et al . 2000 ; ibragimov et al . 2005 ) . we use the eqpair continuum model ( coppi 1999 ) in order to model the curvature of the comptonized continuum . this computes the comptonization of seed photons by a steady state electron distribution which is _ calculated _ self - consistently from balancing an initial heating rate with compton and coulomb cooling processes . the code allows the input heating to be either non - thermal , with electrons injected with a power - law distribution of @xmath61 , between lorentz factors @xmath62 and @xmath63 , or thermal , or some combination of both . however , even if _ all _ the electron heating is non - thermal , the self - consistent electron distribution ( and hence output comptonized spectrum ) can be mainly thermal with only a small non - thermal tail if coulomb collisions dominate the electron cooling . this happens where the power injected in the hard electrons , @xmath64 ( parameterized as a compactness @xmath65 ; see e.g. coppi 1999 ) is larger than that injected in soft photons , @xmath66 . conversely , where the soft photons dominate then compton cooling is most important , and the electron distribution retains a mainly power - law shape with only a small thermal component ( poutanen & coppi 1998 ) . as well as calculating the curvature of the comptonized spectrum at high energies , eqpair also computes the low - energy cutoff due to the seed photon distribution . this cutoff is typically at a few @xmath67 , so is _ within _ the observed bandpass for the mecs data . this curvature can make a large difference to the inferred disc temperature and luminosity ( see e.g. done , ycki & smith 2002 ) , and could potentially distort the inferred spectrum in the red wing of the iron line . we assume that the seed photons are the same as the observed disc photons in all the fits below , and fix @xmath68 . this continuum complexity means that we can not easily use the cdid reflection models since these are tabulated assuming a cutoff in the continuum at a fixed energy of 100 kev . the assumed high energy continuum shape affects the reflected emission at energies above @xmath69 kev due to the predominance of compton down - scattering . instead we use the reflection model in eqpair ( which is based on pexriv but reflecting the self - consistent continuum from eqpair ) , together with a gaussian line . we relativistically smear both reflected continuum and line with the laor kernel , now extending the bandpass for the convolution calculation from 0.21000 kev , and assume an inclination of @xmath15 . purely thermal electron heating gives an ( almost ) acceptable fit , with @xmath70 for a reflection dominated ( @xmath71 ) , highly smeared fit ( @xmath72 , @xmath73 ) . however , purely non - thermal injection models give a significantly better fit , with @xmath74 ( there is one extra free parameter which is the spectral index of the power law electron injection ) . the subtle difference in continuum curvature means that the solution is no longer reflection dominated , with @xmath75 though it is still strongly smeared ( @xmath76 and @xmath77 ) . the mecs and pds spectrum and residuals for this best - fitting non - thermal injection model are shown in fig . [ fig : mecspds](a ) . section [ sec : complex_reflection ] showed that there are large differences in the model spectra for ionized reflection . while the cdid reflection models concentrate on the spectra below @xmath912 kev , computing the effects of photoionization and compton up and down - scattering at low energies , they do not accurately treat compton scattering at higher energies , nor include angle dependent effects , nor allow for different illuminating continua . this contrasts with the pexriv models of reflection which have very approximate ionization balance ( done et al . 1992 ) and neglect comptonization below 12 kev , but treat comptonization at higher energies very accurately , and give the reflected emission as a function of observed inclination angle . pexriv is based on the convolution kernels given by magdziarz & zdziarski ( 1995 ) , so can be extended to work with _ any _ illuminating continuum . plainly what is needed is a combination of these models , using the better cdid photoionization / compton scattering of reflection at low energies and the pexriv accurate compton down - scattering at higher energies . the most flexible way to incorporate this is as a convolution model in xspec , so that the reflected spectrum can be calculated for any arbitrary continuum . we have coded this by finding the pexriv cross - section at 10 kev which gives the same 1214 kev reflected spectrum gradient as the cdid models for a given ionization parameter and illuminating spectral index . the corresponding pexriv greens functions are used to form the spectrum above 12 kev . below 12 kev the greens functions are assumed to collapse to delta functions , given by the ratio of reflected to incident flux in the cdid models , rescaled to match the normalization of pexriv at 14 kev . this model is hereafter called refbal . we use this together with the eqpair continuum model , and get a somewhat worse fit than for the pexriv models , with @xmath78 for an illuminating continuum from non - thermal injection . this is shown in fig.[fig : mecspds](b ) and by contrast with the pexriv reflection , it requires a larger fractional contribution from reflection , with @xmath79 . however , both pexriv and cdid based models require that reflection is highly smeared , with @xmath80 = 23 . we then include the narrow absorption features from xstar ionized absorption . as with the mecs data alone , this dramatically reduces the derived relativistic smearing to being consistent with a disc truncated _ above _ @xmath81 for the cdid reflection models . however , the fit still has @xmath82 i.e. is reflection dominated , although there is a significant contribution from the intrinsic continuum component as well . the inclusion of the pds data up to 200 kev plainly gives more constraints on the amount of reflection than the mecs data alone . with data only up to 10 kev the amount of reflection is very poorly defined , and the spectrum can be pure reflected emission with no intrinsic continuum . including the higher energy data shows that there is most probably true continuum present , though the derived reflected fraction is sensitive to details of how the reflected and continuum emission are modelled . with the pexriv based reflection models , the data are consistent with @xmath83 as expected for a truncated disc , while the more physical cdid based reflection models have substantially more reflection present , with @xmath84 . both these require extreme smearing and emissivity , but like the mecs - only fits , the smearing is again sensitive to the inclusion of narrow absorption lines from an ionized outflow for the cdid reflection models . the bright low / hard state spectrum seen by _ bepposax _ from the transient black hole xte j1650500 is certainly consistent with extremely smeared , reflection dominated emission as claimed by miniutti et al . however , it is also consistent with much less smearing if there is some ionized , outflowing absorption which gives narrow resonance lines at @xmath97.5 kev . reducing the amount of smearing makes the reflected iron line consistent with the expectations of truncated disc models , removing the conflict between the position of the inner edge of the cool disc derived from the broad - band spectral shape and qpos . this is _ directly testable _ with high resolution data around the iron line . such absorbing material is _ often _ seen in x - ray binaries , especially those at fairly high inclination angles ( ueda et al . 1998 ; kotani et al . 2000 ; lee et al . 2002 ; ueda et al . 2004 ) , including xte j1650500 ( miller et al . the one difference here is that the outflowing material has velocity @xmath94.5@xmath85 km s@xmath55 , much higher than the @xmath9500 km s@xmath55 typically seen in these objects . a low velocity wind would be launched from fairly large radii in the disc , as expected for line driving from the uv emitting region of the gbh disc ( proga & kallman 2002 ) . by contrast , the much higher velocities required here would need to be associated with a launch from the inner accretion disc , where line driving is not effective due to the high ionization . this suggests rather that the wind would have to be magnetically driven , and/or be associated with the jet . we note that these data are taken shortly after the time of the maximum in radio emission ( corbel et al . 2004 ) , so the inferred very fast outflow could be associated with the jet ejection . we can estimate the mass outflow rate . for a constant - velocity outflow the mass outflow rate at a radius @xmath86 is @xmath87where @xmath88 is the number density of the outflowing material , @xmath89 is its velocity , @xmath90 is the fractional solid angle in which the outflow is directed and @xmath91 is the proton mass . in a steady state @xmath92 is constant for all radii . the column density measured by the observer is @xmath93 therefore @xmath94 for @xmath95 @xmath14 , @xmath96 , @xmath97 and @xmath98 we find @xmath99 g s@xmath100 ( assuming accretion efficiency of 0.1 and black hole mass of 10 m@xmath12 ) . this is only about 1 per cent or less of estimated inflow mass accretion rate and constitutes a very feasible number . our spectral fitting also shows that this absorption does _ not _ make the data consistent with all the expectations of a truncated disc . the best current models of reflection based on irradiation of a slab of constant density give reflection fractions which are much larger than expected for a truncated disc geometry . this is not necessarily the case using the older pexriv reflection models , which can give much smaller reflection fractions . the difference is due to comptonization of the line and edge in the reflected emission as it travels through the upper layers of the disc ( ross et al . this effect is neglected in pexriv , so the iron edge is rather sharp . on the other hand , the cdid reflection is not calculated for the disc temperatures appropriate for galactic binaries i.e. 0.51 kev . in such hot material , collisional processes alone will ionize iron up to the k shell ( davis et al . by contrast , the cdid models do not include collisional ionization , and thus require intense illumination to ionize iron up to the k shell to match the data . this pulls the temperature of the upper layers to the compton temperature of 1.52 kev , for spectra with @xmath101 1.92 ( ross et al . 1999 ; nayakshin et al . this is a factor 23 larger than the observed disc temperature in these data . thus while the pexriv models underestimate the width of the line and edge by having no compton scattering below 12 kev , the cdid models may overestimate it for galactic binaries . from our spectral fits , the derived reflected fraction is sensitive to the detailed shape of this edge . better models of reflection for conditions appropriate to galactic binary discs will enable this effect to be quantified , removing this as a source of uncertainty . the amount of reflection is also sensitive to the detailed form assumed for the continuum . this is not easily defined even with broad bandpass data . obviously , the wider the bandpass , the more constraints there are on the form of the continuum curvature , but also the more the spectrum is sensitive to the detailed form of continuum assumed . broad bandpass data suggest that the spectrum has convex curvature which can be modelled by an additional soft excess component ( e.g. di salvo et al . 2001 ; ibragimov et al . 2005 ) . this makes the intrinsic continuum at higher energies rather harder than expected from the observed 210 kev emission , reducing the amount of reflection required to match the data at high energies . including such a component in the model makes the reflected fraction almost completely undetermined and consistent with a truncated disc with the refbal and ionized absorber models . such additional continuum complexity is not required statistically from the spectral data alone ( even with the moderate spectral resolution and broad bandpass of _ bepposax _ ) . however , energy resolved _ variability _ could give additional constraints as such data can fairly unambiguously pinpoint the seed photon energy seen by the comptonizing plasma ( gierliski & zdziarski 2005 ) , so could disentangle the shape of the soft emission . most detections of extreme broad iron lines in gbh are in the very high state ( alternatively named the steep power law state ) rather than in the low / hard state examined here ( see e.g. miller et al . this ( typically rather high mass accretion rate ) spectral state shows strong disc emission as well as a strong , steep high - energy tail , while the power spectrum of the variability shows a strong qpo which is close to its maximum frequency ( e.g. mcclintock & remillard 2006 ) . thus both spectral and timing properties imply that the inner disc does extend close to the black hole ( quite how close is difficult to determine due to the complexity of the spectra : kubota & done 2004 ; done & kubota 2005 ) . thus there are no obvious conflicts as for the low / hard state . there are objects where the disc dominated spectra imply moderate spin ( from the @xmath5 relation ) while the iron line smearing implies extreme spin but these could be explained by the accretion flow extending below @xmath8 in the very high state ( reynolds & begelman 1997 ) . however , there _ is _ a more subtle conflict with the extreme line in these data which is that the very high state spectra typically show strongly comptonized emission . the broad band spectral shape seen in e.g. the simultaneous _ asca_-_rxte_-osse data from xte j1550 - 564 implies that the inner disc is almost completely covered by material with optical depth 23 ( done & kubota 2005 ; gierliski & done 2003 ) . thus the spectral shape implies that very little of the inner disc is seen directly , so it _ can not _ make much contribution to the line emission . yet the _ same _ _ asca _ data require an extreme broad line when fit by pexriv based models ( miller et al . again , we suggest that resonance line absorption in an outflowing wind can modify the derived line width and remove these conflicts in the very high state spectra . to test this , we fit the publicly available standard products gas imaging spectrometer data , as this does not suffer from the pileup problems of the solid - state imaging spectrometer for such a bright source . the relativistic smearing required for a pexriv reflected continuum and narrow line are extreme , with @xmath102 and @xmath103 , with @xmath43 = 796.5/742 , similar to the parameters found by miller et al . ( 2004b ) . replacing this with cdid reflection gives similar extreme relativistic smearing but a worse fit ( @xmath43 = 824.6/744 ) . including xstar ionized absorption improves the fit ( @xmath43 = 805.4/741 ) , and as in xte j1650500 this reduces the relativistic effects to the point where @xmath104 ( no upper limit ) with @xmath48 fixed at 3 . the _ extreme _ broad lines inferred in some agn are similarly potentially affected by absorption complexity changing the line parameters . here there are as yet no obvious problems with this interpretation , especially as extreme spin might be _ expected _ given the hierarchical growth of black holes through accretion ( volonteri et al . 2005 ) . nonetheless , if the apparently extreme gbh iron lines are actually consistent with low spin , normal illumination rather than extraction of rotational energy from an extreme spin black hole , then it is likely that the agn discs behave similarly . we note that the recent _ chandra _ spectrum of the most secure extreme relativistic line , that in mcg6 - 30 - 15 , reveals the presence of ionized absorption at a similar column and ionization as that required here to reduce the relativistic smearing to levels consistent with a truncated disc ( young et al . young et al . ( 2005 ) checked that this absorption did not change the extreme smearing deduced from a laor description of the line , but this is unsurprising due to the unphysically sharp blue wing of the line produced in such models ( see fig . [ fig : mod ] ) . our fitting showed that the relativistic line parameters only changed significantly when using the cdid highly ionized reflection models as these have a much smoother intrinsic spectrum around the iron line / edge due to comptonization . we urge testing the robustness of the extreme relativistic effects to the inclusion of ionized absorption with more physical ionized reflection models . the truncated disc / hot inner flow geometry is very successful in qualitatively ( and sometimes quantitatively ) explaining many disparate properties of the low / hard state in ghb . if the radius at which the hot flow evaporates from cool disc progressively decreases as the source approaches the transition to the soft state then this can account for the observed evolution in spectral and timing properties . however , if there really are detections of substantial amounts of iron line emission from cool disc - like material down at the last stable orbit _ in the low / hard state _ , then the truncated disc idea is simply wrong . some of the extreme line emission might be explained by clumps from a disrupted disc embedded in the hot flow , but this would give rise to only a small amount of line since the covering fraction of the clumps would be rather less than from a disc , and these are embedded in a hot flow with optical depth of around unity , so compton scattering suppresses the line . we show that the bright low / hard state spectrum seen by _ bepposax _ from the transient black hole xte j1650500 can indeed be interpreted as requiring strong , extremely smeared reflected emission from a disc - like structure extending down to @xmath105 as shown by miniutti et al . these data clearly illustrate the conflict between this and the position of the inner edge of the cool disc derived from the broad band spectral shape and qpos . the small inner radius inferred from the line width also conflicts ( though more weakly ) with the moderate spin inferred both from theoretical models of stellar collapse and from the observed @xmath5 relation in this object . we show that _ all _ these conflicts can be removed if relativistic effects on the line and reflected continuum are overestimated due to the presence of absorption line features in the spectrum . the derived line width is then easily consistent with normal emissivity and a disc truncated at @xmath91020 @xmath106 ( so giving no information on the black hole spin ) . the reflection feature are most probably shaped by strong , but not necessarily extreme , gravity . we suggest that this effect is robust , i.e. that moderate columns of highly ionized absorbing material ( predominantly resonance lines of h- and he - like iron ) can _ always _ reduce the inferred relativistic smearing of a highly ionized reflected spectrum . this could then remove the more subtle conflict seen in the very high state gbh spectra , where the disc should be down close to the last stable orbit ( so strong relativistic smearing is expected ) but the spectral shape shows that the inner disc is mostly covered by an optically thick , hot corona . we stress that the presence of absorption at this level is directly _ testable _ with high resolution data . while there are no such direct conflicts for agn , we note that the most secure extreme line candidate , mcg6 - 30 - 15 , has highly ionized reflection ( e.g. wilms et al . 2002 ) and a recent _ chandra _ hetg spectrum shows absorption at the same column and ionization ( young et al . 2005 ) as inferred here for xte j1650500 . we urge a re - analysis of these data to see whether the extreme line parameters are robust to the presence of absorption when using the more physical reflection models which include self - consistent comptonization of the line and edge as opposed to models in which the line is intrinsically narrow . mg and cd thank pparc for support through a fellowship and senior fellowship , respectively .	there is growing evidence from both spectral and timing properties that there is a truncated inner accretion disc in low mass accretion rate galactic black hole systems . the detection of _ extremely _ smeared relativistic iron lines in some of these systems is the _ only _ current piece of evidence which conflicts with this geometrical interpretation of the low / hard state . here we show that the line width in the _ bepposax _ data of a bright low / hard state of the transient black hole xte j1650500 is indeed consistent with extreme relativistic effects . however , the relativistic smearing can be _ significantly _ reduced if there is also resonance iron k line _ absorption _ from an outflowing disc wind . the iron line smearing is then completely compatible with a truncated disc , so gives no information on the black hole spin . = -0.5 cm [ firstpage ] accretion , accretion discs black hole physics x - rays : binaries x - rays : individual : xte j1650 - 500
globular clusters are made up of a very large number of stars with varying mass but identical age , composition , and distance . this makes them a rich and productive application of the theory of stellar structure and evolution . detailed stellar evolution calculations are done numerically using computer programs which incorporate previously calculated nuclear reaction rates and opacities , approximations to complex phenomena such as convection , and assumptions about the chemical composition . to compare theoretical models to observations , moreover , requires converting physical quantities such as luminosity and surface temperature to the observational system of magnitudes and colors using empirical relations or the results of separate stellar atmosphere models . the results of theoretical stellar evolution calculations , therefore , depend upon a set of prior assumptions . to assess the reliability of theoretical models of globular clusters stars one must study how uncertainties in these assumptions of stellar evolution theory propagate to the predictions of the theory . in this work we are interested specifically in studying uncertainties in stellar evolution theory along the red giant branch ( rgb ) of metal - poor globular clusters . the rgb is the region in a color - magnitude diagram comprising low - mass stars that have already expended the hydrogen fuel at their center , and are burning hydrogen in a spherical shell around an inert helium core . stars in this stage of evolution grow larger , brighter , and redder over time . an excellent review of rgb evolution is found in @xcite , including a discussion of uncertainties in the theoretical models and observational tests of the theory . much of our work was guided by this review . after a low - mass star has consumed all the hydrogen at its center , the envelope of the star expands and it moves across the color - magnitude diagram from the main sequence region towards lower effective temperature and redder colors . stars in this stage are said to be on the subgiant branch . initially , hydrogen burning continues through a thick shell covering a region of mass @xmath6 , while outside the hydrogen - burning shell the stellar envelope has the original chemical composition with a convective region in the star s outer layers . as the star progresses along the subgiant branch , the hydrogen - burning shell narrows to a mass of @xmath7 while the convective region grows steadily deeper . the convective region eventually reaches material previously processed by the hydrogen - burning core during the main sequence phase , and containing a higher abundance of helium which immediately is mixed throughout the convective region . eventually , the stellar luminosity begins to grow as the effective temperature continues to drop , and the star is said to be at the base of the rgb . the star progresses up the rgb as the hydrogen - burning shell moves outward through the star leaving behind an increasingly massive core of helium . the star s radius continues to grow larger , its effective temperature lower , and its luminosity brighter . the lower boundary of the convective region , which reached a maximum depth near the base of the rgb , recedes steadily . a discontinuity in the chemical composition is left at the maximum depth reached by the convective envelope , since convection mixes hydrogen from the envelope into the partially depleted region left behind by the hydrogen - burning core of the main sequence phase . when the hydrogen - burning shell reaches this discontinuity , the sudden increase in available fuel causes the stellar luminosity to drop temporarily , and the star s evolution pauses before continuing its progression up the rgb . this point is called the rgb bump . there is a very strong correlation between stellar luminosity and the mass of the helium core along the rgb . the luminosity grows as the core mass grows , and since the rate at which mass is added to the helium core is itself proportional to the luminosity , this means that throughout red giant evolution the helium core mass and the luminosity increase at an ever faster rate . the one exception to this is at the rgb bump , where the sudden change in chemical composition causes the growth in luminosity to pause temporarily , while the helium core continues to gain mass . the probability of observing a star in a given luminosity range on the rgb is inversely proportional to the rate at which stars evolve at that luminosity . in observed globular cluster color - magnitude diagrams , therefore , we find that the number of observed stars steadily decreases along the rgb , except at the red giant bump . in fact , the differential luminosity function of a globular cluster , showing the number @xmath8 of observed stars as a function of magnitude , descends along the rgb as a nearly straight line in the magnitude-@xmath9 plane . the slope of this line indicates how the rate of evolution increases along the rgb . globular cluster luminosity functions provide an important test of the accuracy of stellar evolution models . the number of stars observed as a function of luminosity indicates the relative timescale of stellar evolution , which in turn conveys information about the internal chemical structure of stars . a great deal of attention has been paid , for example , to the total number of stars on the rgb compared to the main sequence turn - off region . several authors @xcite have found a discrepancy in this quantity between observed luminosity functions for the globular clusters m5 and m30 and the predictions of stellar evolution theory . @xcite take this discrepancy to suggest the presence of rapidly rotating cores in red giants , while @xcite take it to indicate the possibility of a chemical mixing process deep in the stellar interior . the magnitude of the rgb bump in observed globular clusters serves to indicate the depth of the chemical discontinuity left by the convective envelope at its point of deepest extent near the base of the rgb . in one of the first extensive studies of the red giant bump , @xcite compared observational determinations of the bump magnitude in 11 globular clusters to theoretical predictions . they found that , taken relative to the horizontal branch magnitude , theoretical values of the bump magnitude are higher than observed values by about 0.4 mag . however , more recent studies using updated stellar evolution models and improved observations @xcite have found no discrepancy between theory and observations . @xcite have studied how uncertainties in the most important individual stellar evolution parameters impact the magnitude of the red giant bump . they look at the equation of state , mixing length , mass loss along the rgb , opacities , and the @xmath0-band bolometric correction . in addition , @xcite studied the impact of element diffusion in detail . the effects of overshooting from the convective envelope have been considered by @xcite . this paper presents a comprehensive study of incorporating all the relevant uncertainties to firmly establish the agreement between standard stellar models and observations of the red giant bump . the magnitude of the rgb bump depends upon critically upon the maximum depth of the convection zone . in contrast , the enhancement in the observed number of stars in the bump depends upon the size of the chemical discontinuity . @xcite introduced the parameter , which measures the relative number of stars in the rgb bump . @xcite found that was a robust prediction of stellar evolution theory , as changes in the opacity , equation of state , and nuclear cross sections changed the predicted value of by a few percent . their work found fair agreement between the observed values and those predicted by standard stellar evolution theory . more recently @xcite used hst data to determine in 54 galactic globular clusters . they found this data was in good agreement with standard stellar evolution models . this paper presents a comprehensive analysis of how uncertainties in the inputs into stellar evolution theory effect the predictions for and . section [ method ] provides an overall of the monte carlo approach used in the paper , while [ uncertain ] presents a detailed analysis of the uncertainties in standard stellar evolution theory . the theoretical luminosity functions are presented in [ theolf ] . section [ rgbb ] discusses the results , while the results are presented in [ rbmp ] . a summary of the main results is given in since it is not feasible to use analytic error propagation formulas with the complex calculations of stellar evolution , one must use numerical methods to study how uncertainty in the parameters of stellar evolution theory propagate to the predictions of the theory . a simple approach is to consider one parameter at a time , performing calculations with several different values of the parameter to examine how the predictions of the theory are affected ( e.g. * ? ? ? the effects involved are generally small , so changes in the parameters produce a roughly linear response in results of the calculations . however , this method does not describe how uncertainties in many different parameters interact to produce a combined uncertainty in the predictions of the theory . to investigate this requires a larger set of calculations , in which all of the significant parameters vary simultaneously . we use a monte carlo approach to the problem . describing the estimated uncertainty in each parameter with a probability distribution , we run a large number of independent stellar evolution calculations in which the value of each parameter is drawn randomly from its corresponding probability distribution . for specific quantitative predictions of the theory , then , the results of all the runs can be combined into histograms showing the most probable value and the distribution of uncertainty around it . this method has been used previously in a series of papers examining uncertainties in theoretical globular cluster ages @xcite . the work here is essentially an extension of this method to study theoretical uncertainties in rgb evolution . stellar evolution is strongly affected by the abundance of heavy elements in a star , indicated by the relative iron abundance [ fe / h ] . because the iron abundance ( or metallicity ) of a star can be determined from observations , it is treated as a known quantity in our simulation rather than as a source of theoretical uncertainty . we therefore conduct our analysis at five different metallicity values spanning a range typical of old galactic globular clusters . for each set of randomly generated parameters in the monte carlo simulation , we calculate an evolutionary sequence using [ fe / h ] = @xmath101.0 , @xmath101.3 , @xmath101.6 , @xmath102.0 , and @xmath102.4 . the results are analyzed separately for each metallicity . our stellar evolution calculations are made using the dartmouth stellar evolution program @xcite , which is a descendant of the yale rotating stellar evolution code @xcite . to explore the evolution of metal - poor stars along the rgb , we calculated evolutionary sequences for a variety of masses and metallicities starting from previously constructed zero - age main sequence models . evolution was carried up to the tip of the rgb , which we defined as the point where energy generated by the triple - alpha reaction reached 1% of the star s total luminosity . the equations of stellar structure were solved to an accuracy of a few parts in @xmath11 . with these tolerances , the code requires about 6000 7500 individual stellar models to evolve the highest metallicity stars to the tip of the rgb , and for the lowest metallicity stars it requires 3500 4000 models . each stellar model comprises about 600 spherical shells on the main sequence , and approximately 1500 shells along the rgb . the greatest difficulty in the monte carlo approach is that it is computationally very demanding . this study required 45,000 evolutionary sequences , roughly 230 billion stellar models , and approximately 600 cpu days on a beowulf cluster of 2.0ghz athalon processors . in all the stellar evolution calculations we use the standard dsep equation of state , which includes the debye - hckel correction of @xcite for electrostatic interactions in the fully ionized plasma . the dsep equation of state in the fully ionized regime ( @xmath12 k ) is essentially the relativistic ideal gas law with the electrostatic interaction , radiation pressure , and electron degeneracy effects included . in the partially ionized regime it is a saha equation including single ionization of hydrogen and metals , and double ionization of helium . although a more sophisticated equation of state is now available in the opal tables of @xcite , @xcite have found that using the opal tables in place of the standard dsep equation of state does not change the evolution of metal - poor stars significantly . based on their work , we consider the equation of state to be an insignificant source of uncertainty in metal - poor red giant evolution . similarly , all of our calculations use the standard treatments of @xcite and @xcite to evaluate plasma screening effects on the nuclear reactions . more recent treatments of plasma screening are available ( e.g. , * ? ? ? . however , @xcite report that their numerical experiments found the new formulations to have no impact on metal - poor red giant evolution . we therefore do not consider plasma screening as a source of uncertainty . the other important parameters of stellar evolution theory opacities , nuclear reaction rates , composition , surface boundary conditions , neutrino cooling rates , and the treatments of convection and diffusion represent possible sources of uncertainty , and so they are varied throughout the monte carlo simulation . the treatment of these parameters is detailed in the next section . there are 18 parameters of stellar evolution theory that we consider as possibly significant sources of uncertainty . some of these , like the primordial helium abundance , are simple numerical values that vary through the monte carlo simulation . others , though , like the opacities and nuclear reaction rates , enter the stellar evolution calculations as tables or formulas . these quantities are varied by multiplying the table or formula by an overall coefficient , and then varying the coefficient through the monte carlo simulation . in the case of high - temperature opacities , we use different but correlated multiplicative factors in different regions of the table . for the bolometric corrections , which are logarithmic quantities , we add rather than multiply by an overall uncertainty term . during the monte carlo simulation , the value of each parameter is drawn randomly from a probability density distribution that we choose to reflect the estimated uncertainty in the parameter . in general , we use gaussian distributions to reflect uncertainties that are statistical in nature , with a well - defined most probable value . for systematic uncertainties we use uniform distributions in which the probability density is constant within some range and zero outside of it . for the surface boundary conditions there are two independent formulations to choose from , and here we use a binary distribution in which one or the other formulation is chosen with equal probability . the 18 parameters varied in the monte carlo simulation are listed in table [ table1 ] along with the probability density distributions used in the monte carlo simulation . in many cases , we adopt a distribution from one of the previous papers in which monte carlo simulations were used to investigate stellar evolution models at the main sequence turn - off point and on the subgiant branch @xcite . for others , we review the literature directly to estimate an appropriate distribution , and these are discussed in detail . * convective overshoot. one of the weaknesses of the mixing length approximation used in standard stellar evolution models is that it fails to properly describe the boundary between convective and radiative regions in a star . assuming no composition gradient , this boundary is established by the schwartzschild condition as the point where the temperature gradient in the star is equal to the adiabatic temperature gradient . the mixing length approximation assumes that all convective motion is confined to this region of instability . in reality , convective material reaches the boundary of the unstable region with a non - zero velocity , and therefore tends to penetrate across the schwartzchild boundary into the stable radiative region . this phenomenon is called convective overshoot , and it has been demonstrated in hydrodynamical simulations of convection ( e.g. , ? ? ? * ) . the extent to which overshooting actually occurs in stars , however , is unclear . since the stars in our simulation do not have convective cores , we are concerned only with overshooting at the base of the surface convection zone . the stellar evolution calculations model convective overshooting by assuming that the region in which the chemical composition is homogenized by convective mixing extends some distance below the schwartzschild boundary into the stable radiative region . temperature gradients , however , are taken to be unaffected by overshooting . the single parameter in this treatment is the depth by which convective mixing reaches into the radiative region , and this distance is expressed as a fraction of the pressure scale height at the schwartzschild boundary . an upper limit on convective overshooting in metal - poor stars is set by observed lithium abundances . lithium breaks down in the high temperatures of stellar interiors , beginning near the bottom of the surface convection zone . a large amount of convective overshooting would carry lithium more quickly from the surface to the interior , and lithium depletion would occur at a faster rate . overshooting depths greater than about 0.2 @xmath13 would be inconsistent with the relatively high lithium abundances observed in metal - poor stars . in our simulation , therefore , we draw values for the overshooting depth from the uniform distribution 0.0 0.2 @xmath14 * diffusion coefficients. models of stable astrophysical plasmas predict that helium and heavy elements in a star s radiative regions should settle toward the center of the star over time , while hydrogen rises toward the surface . the extent to which this actually occurs , however , is subject to considerable uncertainty . there is clear evidence from helioseismology that element diffusion occurs in the sun @xcite . studies of surface abundances in metal - poor stars , though , indicate that in these stars diffusion for some reason does not occur in the outer layers @xcite . our stellar evolution calculations incorporate element diffusion using the treatment of @xcite . because diffusion is not seen in the outer layers of metal - poor stars , we include a modification introduced by @xcite which suppresses diffusion near the surface . by comparing observed iron abundances in globular clusters to the iron abundance predicted by their stellar evolution models , @xcite estimate that whatever process inhibits diffusion in metal - poor stars must act over an outer region of at least 0.005 @xmath15 . on the other hand , the fact that diffusion is not found to be inhibited in the sun suggests that the process inhibiting diffusion extends no lower than the bottom of the solar surface convection zone , which has a mass of @xmath16 . therefore @xcite conclude that the process inhibiting diffusion likely acts over a surface region with mass somewhere between 0.005 and 0.02 @xmath15 . to reflect this in our calculations we follow @xcite in setting the diffusion coefficients to zero in the outer 0.005 @xmath15 layer of the star . in the interior region defined by @xmath17 where @xmath18 is the star s total mass , the diffusion coefficients are set to the standard values of @xcite . in the middle region @xmath19 , the coefficients are ramped from zero to the standard values . to reflect the uncertainty in this treatment of diffusion , we multiply the diffusion coefficients overall by a factor drawn from the uniform distribution 0.5 1.3 . this range is chosen based on an estimated 30% uncertainty in the theoretical calculations of diffusion velocities ( see e.g. ? ? ? * ) , but the lower end of the range is extended to account for the fact that several physical processes in the interior of a star might slow the rate of diffusion , while no such processes would increase it . * triple - alpha reaction rate. the point at which helium ignition occurs in a star s evolution depends on the rate of the helium - burning triple - alpha reaction , in which three helium nuclei fuse to form a single nuclei of @xmath20 . our calculations use the rate of @xcite , with an uncertainty estimated at 15% ( e.g. ? ? ? we therefore multiply the standard @xcite rate by a factor drawn from the gaussian distribution @xmath21 . * plasma neutrino cooling rate. neutrino cooling in the stellar core becomes an important effect in the evolution of stars on the upper rgb . energy lost through the emission of neutrinos tends to slow the rise in temperature as the helium core grows , allowing the core to reach a higher mass before helium burning begins . significant energy loss can occur through several different neutrino processes in a stellar plasma ; for the helium core of stars on the upper rgb the most important is the plasma process , in which a plasmon ( a quantized electromagnetic oscillation in the plasma ) spontaneously decays into a neutrino - antineutrino pair @xcite . we use the plasma neutrino cooling rates of @xcite , who have calculated numerical rates based on the treatment of electromagnetic dispersion relations in plasmas provided by @xcite , and have also derived an analytic approximation to the numerical results for use in stellar evolution calculations . they find their analytic approximation accurate to within 4% in regimes where the plasma process is dominant and within 5% everywhere . published neutrino cooling rates are generally calculated under the standard assumption that neutrinos have no magnetic dipole moment . a non - zero magnetic moment for neutrinos would tend to increase the plasma neutrino emission rates and could impact stellar evolution significantly . in fact , @xcite and @xcite have used stellar evolution calculations incorporating a non - zero neutrino magnetic moment to constrain the magnetic moment through comparison to globular cluster observations . although the existence of a non - zero magnetic moment could significantly impact stellar evolution at the tip of the rgb , this possibility is considered at the moment to be outside of standard physics , and so we do not incorporate it into our analysis . instead , we consider the error in the @xcite analytic approximation formula to be the dominant source of uncertainty in the neutrino cooling rates , and to reflect this in the monte carlo simulation we multiply the @xcite formula by a factor drawn from the gaussian distribution @xmath22 . conductive opacities. in the deep interior of rgb stars , thermal conduction by electrons serves as an important means of energy transfer . the thermal conductivity under these conditions therefore serves to establish the temperature gradient in red giant interiors and determine when the core becomes hot enough to ignite helium . conductive opacities are one of the most significant sources of uncertainty on the upper rgb , so a close examination of our current understanding in this area is appropriate . there are essentially two treatments of conductive opacity available for use in red giant models : the tabulated values of @xcite and the more recent calculations of @xcite . however , as @xcite point out , the @xcite calculations were made for conditions characteristic of white dwarfs and neutron stars ; red giant cores do not fall within their range of validity . specifically , the @xcite results are presented as an analytic fitting formula in terms of the parameter @xmath23 which characterizes the strength of the electrostatic interaction between ions in a plasma . here @xmath24 is the electron charge , @xmath25 is the atomic number , @xmath26 is the boltzmann constant , and @xmath27^{1/3}$ ] is the ion - sphere radius with @xmath28 the number density of ions . the domain where @xmath29 represents matter in the liquid metal phase , while @xmath30 corresponds more typically to a boltzmann gas . the fitting formula of @xcite intended for application to dense matter in the liquid metal phase , is valid only in the range @xmath31 , while in red giant cores @xmath32 is considerably lower . applying the results of @xcite to rgb stars therefore requires extrapolating the fitting formula from liquid metal conditions to boltzmann gas conditions , and it is not clear how much error this extrapolation introduces . because of this , @xcite conclude that the @xcite values are to be preferred at present . the treatment of conductive opacity in our stellar evolution calculations is based on the @xcite tables , along with the relativistic extension provided by @xcite for higher densities . specifically , we employ a set of analytic fitting formulas from @xcite ; for @xmath33 , we use sweigart s fit to the @xcite tables , while for @xmath34 we use sweigart s fit to the canuto relativistic opacities . for @xmath35 we use a ramp between the non - relativistic and relativistic formulas , also provided by @xcite . for conditions appropriate to red giant stars , we find that sweigart s formula are accurate to about 10% when compared to the @xcite tabulations . we estimated the uncertainty in conductive opacities by directly comparing values from the @xcite analytic formula to the tabulated values of @xcite . @xcite have plotted both these sets of opacities for conditions typical of red giant cores . their plot shows that here the @xcite values run @xmath36 @xmath37 lower than the @xcite values . since the @xcite fitting formula was not designed to be accurate in this regime , though , it is possible that this comparison could exaggerate the difference between the two treatments . to check this , we also compared the two sets of opacities in a density - temperature region where both treatments are valid . we make the comparison at densities typical of red giant cores ( @xmath38 @xmath39 ) . the acceptable temperature range is then bounded by two restrictions in the @xcite analysis . on the upper end there is the restriction that @xmath40 . on the lower end is the requirement that a high - temperature classical limit apply to the treatment of the ionic system . this is expressed by the condition @xmath41 where the parameter @xmath42 is defined @xmath43 with @xmath44 being the fermi wavenumber of the electrons and @xmath45 the mass of an ion . @xcite found that the high - temperature classical limit is adequate as long as @xmath46 . in the range @xmath47 , though , their results show that a quantum correction not included in the @xcite analysis tends to reduce the conductive opacities by up to 25% . in figure [ copacfig ] we have plotted the conductive opacities at densities of @xmath48 and @xmath49 for temperatures where @xmath40 and @xmath50 , and we show the temperatures corresponding to @xmath51 . the agreement between the different treatments is better at the higher temperatures , where the classical treatment of ions by @xcite is most valid and also where the temperatures approach those typically found in red giant cores ( @xmath52 @xmath53 k ) . in the region where @xmath54 the treatments agree to within 25% , with the @xcite values running 5 25% below the @xcite values . taking all of this into consideration , we estimate that current values for the conductive opacity in red giant cores are uncertain by about 20% at the 1-@xmath55 level . we therefore multiply our standard values ( obtained as described above from sweigart s fit to the @xcite tables and the relativistic canuto treatment ) by a factor drawn from the gaussian distribution @xmath56 . bolometric corrections. in order to compare our stellar evolution models to observations , it is necessary to transform the physical quantities used in stellar evolution calculations to those quantities measured by observers . specifically , we must transform the theoretically calculated bolometric luminosity , effective temperature , and surface gravity to predict the observational magnitudes for specific pass bands . generally , this transformation is made using color tables based on theoretical treatments of stellar atmospheres , possibly calibrated by empirical data . purely empirical relationships between magnitude and effective temperature are also available based on effective temperatures measured for nearby stars , but these relationships are valid only for a restricted range of metallicities and evolutionary stages . the transformations used in our analysis are based on the color tables constructed for the revised yale isochrones @xcite . these transformations are an empirical recalibration of the theoretical colors and bolometric corrections of @xcite and @xcite . in this study we will examine the @xmath0-band magnitude of the rgb bump , so we must estimate the uncertainty in the @xmath0-band bolometric corrections . @xcite compare the @xmath0-band bolometric corrections @xmath57 from several theoretical , empirical , and semi - empirical sources and find that with a consistent choice for the solar @xmath58 , all the sources agree to within 0.05 mag along a typical globular cluster isochrone . our own analysis of different @xmath57 from different sources confirm this result . therefore , for @xmath58 , we draw the uncertainty term from the uniform distribution @xmath590.05 mag . to explore uncertainties in and , we conducted a monte carlo simulation with 1120 independent sets of randomly chosen stellar evolution parameters , from each set generating luminosity functions at several different metallicities , ages , and assumed initial mass functions . generating luminosity functions requires a series of theoretical evolutionary tracks calculated for stars over a range of masses . in this work we use the nine masses @xmath45 = 0.55 , 0.63 , 0.70 , 0.75 , 0.80 , 0.85 , 0.90 , 0.95 , and 1.00 @xmath15 . each evolutionary track is calculated from the zero - age main sequence to the tip of the rgb . the range of masses was chosen to allow reliable luminosity functions for ages between 10 and 20 gigayears . from a set of evolutionary tracks at appropriate masses one can calculate theoretical isochrones at a variety of ages . we use a modified version of the isochrone generating program used to construct the revised yale isochrones @xcite . the program uses the method of equivalent evolutionary points , locating in each evolutionary track a set of points defined in terms of the central helium abundance ( on the main sequence ) or the helium core mass ( on the rgb ) and interpolating among them to generate the isochrones . from these isochrones , @xmath0-band luminosity functions are calculated . uncertainty in the bolometric correction is incorporated at a later stage of the analysis . luminosity functions are normalized to 1000 stars on the zero - age main sequence and use a bin size of 0.04 mag . generating a luminosity function requires an initial mass function ( imf ) describing the relative number of stars as a function of mass created during the formation of the cluster . the imf is generally assumed to have the form of a simple declining power law , @xmath60 . in the classic study , @xcite found that observational data suggested an initial mass function with exponent @xmath61 = 2.35 ; more recent studies , though , find evidence that the initial mass function for globular clusters varies from cluster to cluster and that metal - poor clusters in general have less steep imfs than metal - rich clusters . because stars on the rgb represent a narrow range of masses , though , the luminosity function in the red giant region is less affected by the imf than are the main sequence and subgiant branch . in particular , we expect that and will be unaffected by the imf . this is tested by generating luminosity functions using the @xcite value for the imf exponent as well as values from extremes of the reasonable range : @xmath61 = 0.00 , 2.35 , and 4.00 . sample luminosity functions are shown in figure [ lffig ] . these luminosity functions were calculated using values for the stellar evolution parameters from the center of the uncertainty ranges discussed in section [ uncertain ] . luminosity functions are shown for a single metallicity at ages of 11 , 13 , and 15 gigayears using the salpeter value for the imf exponent , @xmath61 = 2.35 . the entire set of monte carlo luminosity functions ( 1140 ) for a salpeter imf , an age of 13 gyr and @xmath2$}}= -2.4 $ ] and @xmath2$}}= -1.0 $ ] are shown in figure [ mclffig ] . in making this figure , the luminosity functions have been normalized so that they all have the same number of stars at @xmath62 , which is approximately 1.5 mag below the turn - off . to precisely locate the peak of the rgb bump in our calculated luminosity functions , we find the point in the luminosity function which corresponds to the most populated bin in the bump region , and then calculate the parabola passing through that point and the adjacent points on either side . the vertex of the parabola is taken as the peak of the bump , and the magnitude of the vertex as the bump magnitude . as discussed in section [ uncertain ] , the uncertainties in the @xcite bolometric corrections taken into consideration by adding to the bump magnitude a term drawn randomly from the uniform range @xmath63 mag . for each of the 1120 monte carlo realizations at a given age and metallicity we generate ten such uncertainty terms , so that 1120 independent luminosity functions are used to produce 11,200 realizations of the bump magnitude . these realizations are independent in the bolometric correction , though not in other stellar evolution parameters . the increase in the number of points does not significantly improve the numerical accuracy of the monte carlo simulation , but it does provide smoother histograms . studies of the rgb bump ( e.g. ? ? ? * ) often focus on the difference in magnitude between the bump and the zero - age horizontal branch level , @xmath64 . observationally , this quantity is more reliable than @xmath65 since it avoids uncertainties in the distance modulus and in the calibration of observational data . the theoretical evaluation of @xmath66 , however , is subject to uncertainties in the luminosity of the horizontal branch . in order to confine our analysis to uncertainties on the rgb , we will consider the absolute magnitude @xmath67 of the rgb bump rather than @xmath66 . table [ table2 ] shows the median values of the bump magnitude obtained in the monte carlo simulation , along with 68% and 95% confidence limits . both the median values and the confidence limits depend strongly on the cluster metallicity . in contrast , the choice of cluster age serves merely to shift the distributions of bump magnitude without affecting their shape , so the same confidence limits apply at all ages . in order to show the shape of the bump magnitude distributions , we plot the results for age 13 gigayears as histograms in figure [ bmphist ] . both table [ table2 ] and figure [ bmphist ] show that the uncertainty in theoretical determinations of the bump magnitude increases with metallicity . at [ fe / h ] = @xmath101.0 the distribution of bump magnitudes obtained in the monte carlo simulation is about 80% wider than at [ fe / h ] = @xmath102.4 . at all metallicities the distributions show a slight asymmetry , falling off more gradually toward dimmer magnitudes . the effect of cluster age on the bump magnitude is essentially the same at all metallicities . an increase of 1 gigayear in age corresponds to a shift towards dimmer bump magnitudes of 0.03 0.04 mag , not a large effect compared to the width of the magnitude distributions . even at the lowest metallicity , where uncertainties in stellar evolution theory leave the tightest distribution of bump magnitudes , the shift in magnitude between age 11 gyr and age 15 gyr is only 60% of the distribution width at the 68% confidence level . the effect of metallicity on the bump magnitude is much more significant . changing the cluster metallicity by @xmath68 to 0.2 dex leads to a shift in the bump magnitude comparable to the 68% theoretical confidence limits in table [ table2 ] . since observational measurements of [ fe / h ] are uncertain by about @xmath590.15 dex , this means that when we compare observational and theoretical determinations of the bump magnitude the uncertainties in stellar evolution theory are about as important as observational uncertainties in the metallicity scale . figure [ bmpfeh ] shows the bump magnitude as a function of metallicity , compared to other theoretical and observational results . error bars are shown representing the 68% confidence limits at age 13 gyr , but we omit the error bars at ages 11 and 15 gyr for clarity . the theoretical relation of @xcite for clusters of age 15 gigayears is also shown , as well as observational results for 19 galactic globular clusters drawn from @xcite . @xcite present their results in terms of the global metallicity [ m / h ] , and they assume a different value than we do for the solar heavy element abundance @xmath69 . to make their treatment of metallicity agree with ours , we shift their results to the [ fe / h ] scale assuming [ @xmath61/fe ] = 0.45 ( the median value in our monte carlo simulation ) and @xmath69 = 0.018 . the @xcite relation is somewhat lower than our results for age 15 gyr , especially at the highest and lowest metallicities . the difference , however , is well within the theoretical uncertainties . the @xcite data points show the @xmath67-band magnitude of the red giant bump in 19 galactic globular clusters observed with the _ hubble space telescope_. we use the @xcite metallicities for each cluster , and convert from apparent to absolute magnitude using the rr lyrae calibration of @xcite : @xmath70 + 1.9)$ ] . the uncertainty in the absolute magnitude of the bump is dominated by an uncertainty of @xmath590.12 mag in the rr lyrae calibration , while the uncertainty in globular cluster metallicities is about @xmath590.15 dex . the agreement between the @xcite observations and our theoretical models is excellent . @xcite themselves found a good agreement between their observational values of @xmath71 and the models of @xcite assuming an alpha - capture overabundance [ @xmath61/fe ] = 0.30 for [ fe / h ] @xmath72 @xmath101.0 . our results show , similarly , a good agreement between observations and theoretical values of @xmath65 calculated with a median alpha - capture overabundance [ @xmath61/fe ] = 0.45 . moreover , the results of our monte carlo uncertainty analysis show that differences between theoretical and observed values of the bump magnitude lie within the theoretical uncertainties in stellar evolution as well as within the observational uncertainties in the globular cluster distance modulus and the metallicity scale . it is clear that there is currently no discrepancy between the @xmath0-band bump magnitude as observed in galactic globular clusters and as calculated in standard stellar evolution models . to analyze the individual influence of each continuously varying stellar evolution parameter on , all 1120 monte carlo realizations are plotted on a graph of versus parameter value . the dependence of on a given parameter value is then characterized with a straight line fit . specifically , the parameter range is divided into 20 bins containing 56 monte carlo realizations each , and for each bin the median value of and the 68% confidence limits are determined . a straight line is fit to the median points of each bin using the simple least - squares method with each bin weighted equally . straight lines are also fit to the upper and lower 68% confidence points in each bin . table [ table3 ] presents the slope of the linear dependence for the most significant stellar evolution parameters as well as the total change in bump magnitude as the parameter varies across its range of uncertainty , either between the endpoints of a uniform distribution or between the 68% confidence limits of a gaussian distribution . the impact of surface boundary conditions is explored by comparing the monte carlo distribution of bump magnitudes obtained using the eddington @xmath73 relation to that using the @xcite relation , and table [ table3 ] records the difference between the median bump magnitudes obtained with each of the two treatments . the three most significant sources of uncertainty in the bump magnitude are the alpha - capture abundance , the mixing length , and the low - temperature opacities . all three of these parameters have a stronger influence at higher metallicities , which explains why the overall uncertainty in the bump magnitude increases with metallicity . other significant sources of uncertainty are the surface boundary conditions , convective overshoot , the @xmath74 reaction rate , and the helium diffusion coefficients , the impact of these parameters does not vary significantly with [ fe / h ] , and so the values given in table [ table3 ] are averaged across all five metallicities . all of the other stellar evolution parameters explored in the monte carlo simulation are negligible , impacting the bump magnitude at the level of 0.05 mag or less . @xcite and @xcite have already studied the impact of a few of these stellar evolution parameters on theoretical bump magnitudes . exploring the influence of mixing length , @xcite find a dependence @xmath75 mag with @xmath76ml in units of @xmath13 and with the metallicity unspecified , a result that is in rough agreement with our more detailed analysis . @xcite consider the impact of diffusion on the bump luminosity , and find that including helium and heavy element diffusion according to the @xcite coefficients lowers the bump luminosity by @xmath77 mag compared to models with no diffusion . the impact of diffusion in our monte carlo simulation is comparable . it is interesting that convective overshoot represents only a secondary source of uncertainty in the bump magnitude . after one of the first studies of the rgb bump @xcite found a substantial discrepancy of 0.4 mag between observational and theoretical determinations of @xmath78 , @xcite suggested that this discrepancy could be resolved by including convective overshoot with extension 0.7 @xmath13 in the theoretical models . more recent comparisons using newer observations and theoretical models @xcite have found that the discrepancy is resolved even without convective overshoot . in addition , though , our results show that with an upper limit of 0.2 @xmath13 set by observed lithium abundances in metal - poor stars , the question of whether to include convective overshoot in stellar evolution models is in fact of only secondary importance , much less significant than the alpha - capture abundance , mixing length , and low - temperature opacities . finally , we consider the confidence limits in the bump magnitude obtained when the most significant parameters are held fixed at the center of their uncertainty range , indicating the accuracy that theoretical bump magnitudes would have if one of the important stellar evolution parameters were known precisely . table [ table4 ] shows the 68% confidence limits for fixed values of the alpha - capture overabundance , mixing length , and low - temperature opacities . the median bump magnitudes in table [ table2 ] for each metallicity and age still apply . particularly important are the confidence limits obtained for fixed values of [ @xmath61/fe ] , both because this parameter is the greatest source of uncertainty and because it is in principle an observable quantity . the confidence limits in table [ table4 ] indicate that if the alpha - capture abundance could be precisely determined for a given cluster , the uncertainty in the theoretical bump magnitude for that cluster would be reduced by about 35% . improved spectroscopic observations of globular clusters allowing direct measurements of the alpha - capture abundance will therefore be the most important step toward improving the accuracy of theoretical bump magnitudes . the relative enhancement of the number of stars in the rgb bump can be measured using the parameter introduced by @xcite . it measures the relative evolutionary timescale for stars in the bump region to stars on the rgb which have not yet reached the bump region . the parameter was measured in each of the mc luminosity functions , using the definition of @xcite : the ratio between the star counts in the bump region @xmath79mag to the star counts fainter on the rgb : @xmath80 . in agreement with @xcite we find that is a robust prediction of stellar evolution theory , with little uncertainty . the parameter has a small dependence on and age , and the results from the mc simulation are summarized in table [ table5 ] . it was found that , at a given and age , the @xmath81 uncertainty in the predication of is @xmath4 . parameters which effect the most are , in order of significance , @xmath82 $ ] , mixing length , the high temperature opacities , helium diffusion and the low temperature opacities . the relatively modest effect of each of these parameters on is summarized in table [ table6 ] . comparison between observed and predicated values of serves as an excellent test of standard stellar evolution theory . @xcite used hst data to determine in 54 galactic globular clusters , of which 40 are metal - poor ( @xmath2$}}\le -1.0 $ ] ) and can be compared to our mc results . to this data set , we have added additional determinations of from studies which presented high quality luminosity functions of the globular clusters m3 @xcite , m5 @xcite , m10 @xcite , m12 @xcite . each of these observational studies presented luminosity functions based upon observations of @xmath83 stars in a given cluster . the comparison between these data and the mc models is shown in figure [ rbmpfeh ] , assuming the @xcite metallicity scale . for this metallicity scale our theoretical predication has a reduced @xmath84 when compared to the observations . this implies that our models provide a reasonable fit to the observations . formally , this value of a reduced @xmath85 implies that the models have a 14% probability of correctly describing the data . if the @xcite metallicity scale is used , then the reduced @xmath86 ( 6.5% propability ) and the models provide an acceptable fit to the data . the observed luminosity function of m10 was found to disagree with the predications of standard stellar evolution theory @xcite , due to the fact that there are too many stars on the rgb compared to the main - sequence . from the @xcite luminosity function , m10 has @xmath87 , compared to a predicted value of @xmath88 on the @xcite metallicity scale and @xmath89 on the @xcite metallicity scale . thus , the excess number of stars on the main - sequence has not lead to an anomalous number of stars in the bump region on the rgb . we use a large monte carlo simulation to investigate theoretical uncertainties in the rgb luminosity function , specifically in the @xmath0-band magnitude of the red giant bump and excess number of stars in the bump . we find excellent agreement between theoretical and observational values for the bump magnitude , with the uncertainty in theoretical values comparable to the scatter in observational values . metallicity has a very significant effect on the bump magnitude , while the cluster age impacts the bump magnitude at a level lower than the theoretical uncertainties . the most important sources of uncertainty in the predication of the bump magnitude are the alpha - capture overabundance [ @xmath61/fe ] , the mixing length , low - temperature opacities , and the treatment of surface boundary conditions . theoretical uncertainties in stellar evolution models have little impact on the excess number of stars in the bump region . this is a robust prediction of standard stellar evolution is found to be in reasonable agreement with observations . he mass fraction ( y ) & n / a & @xmath90 & unif . + @xmath82 $ ] & n / a & @xmath91 dex & unif . + mixing length & n / a & @xmath92 & gaus . + convective overshoot & n / a & @xmath93 & unif . + atmospheric @xmath73 & grey or @xcite & 50/50 & binary + low - t opacities & @xcite & @xmath94 & unif . + high - t opacities & @xcite & @xmath95 & gaus . + & & @xmath96 & + diffusion coefficients & @xcite & @xmath97 & unif . + @xmath98 & @xcite & @xmath99 & gaus . + & & @xmath100 & unif . + @xmath101 & @xcite & @xmath102 & gaus . + @xmath103 & @xcite & @xmath104 & gaus . + @xmath105 & @xcite & @xmath106 & gaus . + @xmath74 & @xcite & @xmath106 & gaus . + @xmath107 & @xcite & @xmath108 & gaus . + @xmath109 & @xcite & @xmath110 & gaus . + triple-@xmath61 reaction rate & @xcite & @xmath106 & gaus . + neutrino cooling rate & @xcite & @xmath111 & gaus . + conductive opacities & @xcite & @xmath112 & gaus . + @xmath58 & @xcite & @xmath113 mag & unif . + @xmath114 & @xmath115 & @xmath116 & @xmath117 & @xmath118 & @xmath119 & & @xmath120 & @xmath121 + @xmath122 & @xmath123 & @xmath124 & @xmath125 & @xmath126 & @xmath127 & & @xmath128 & @xmath129 + @xmath130 & @xmath131 & @xmath132 & @xmath133 & @xmath134 & @xmath135 & & @xmath136 & @xmath137 + @xmath138 & @xmath139 & @xmath140 & @xmath141 & @xmath142 & @xmath143 & & @xmath144 & @xmath145 + @xmath146 & @xmath147 & @xmath148 & @xmath149 & @xmath150 & @xmath151 & & @xmath152 & @xmath153 + lccrc & @xmath91 dex & @xmath114 & @xmath154 & @xmath155 + & & @xmath122 & @xmath156 & @xmath157 + & & @xmath130 & @xmath158 & @xmath159 + & & @xmath138 & @xmath160 & @xmath161 + & & @xmath146 & @xmath162 & @xmath163 + mixing length & @xmath164 & @xmath114 & @xmath165 & @xmath166 + & & @xmath122 & @xmath167 & @xmath168 + & & @xmath130 & @xmath169 & @xmath170 + & & @xmath138 & @xmath171 & @xmath172 + & & @xmath146 & @xmath173 & @xmath174 + low - t opacities & @xmath175 & @xmath114 & @xmath176 & @xmath177 + & & @xmath122 & @xmath178 & @xmath179 + & & @xmath130 & @xmath180 & @xmath181 + & & @xmath138 & @xmath182 & @xmath183 + & & @xmath146 & @xmath184 & @xmath185 + convective overshoot & @xmath186 & avg . & @xmath187 & @xmath188 + atmospheric @xmath73 & eddington / & avg . & & @xmath188 + & krishna - swamy & & & + @xmath74 & @xmath189 & avg . & @xmath180 & @xmath188 + he diffusion coef . & @xmath190 & avg . & @xmath191 & @xmath192 + lcc @xmath82 $ ] & @xmath114&@xmath193 + & @xmath122&@xmath194 + & @xmath130&@xmath195 + & @xmath138&@xmath196 + & @xmath146&@xmath197 + mixing length & @xmath114&@xmath198 + & @xmath122&@xmath196 + & @xmath130&@xmath199 + & @xmath138&@xmath200 + & @xmath146&@xmath201 + low - t opacities & @xmath114&@xmath202 + & @xmath122&@xmath203 + & @xmath130&@xmath204 + & @xmath138&@xmath205 + & @xmath146&@xmath206 + @xmath114 & @xmath207 & @xmath208 & @xmath208 & @xmath209 & @xmath134 & @xmath134 + @xmath122 & @xmath210 & @xmath211 & @xmath211 & @xmath135 & @xmath208 & @xmath208 + @xmath130 & @xmath212 & @xmath213 & @xmath214 & @xmath215 & @xmath215 & @xmath211 + @xmath138 & @xmath5 & @xmath216 & @xmath213 & @xmath213 & @xmath214 & @xmath214 + @xmath146 & @xmath5 & @xmath216 & @xmath213 & @xmath213 & @xmath213 & @xmath213 + lccrc & @xmath91 dex & @xmath114 & @xmath217 & @xmath218 + & & @xmath122 & @xmath219 & @xmath220 + & & @xmath130 & @xmath221 & @xmath222 + & & @xmath138 & @xmath223 & @xmath224 + & & @xmath146 & @xmath225 & @xmath226 + mixing length & @xmath164 & avg . & @xmath227 & @xmath228 + high - t opacities & @xmath229 & @xmath138 to @xmath114 & @xmath230 & @xmath231 + he diffusion & @xmath190 & @xmath138 to @xmath114 & @xmath232 & @xmath233 + low - t opacities & @xmath94 & avg . & @xmath234 & @xmath235 +	a monte carlo simulation exploring uncertainties in standard stellar evolution theory on the red giant branch of metal - poor globular clusters has been conducted . confidence limits are derived on the absolute @xmath0-band magnitude of the bump in the red giant branch luminosity function ( ) and the excess number of stars in the bump , . the analysis takes into account uncertainties in the primordial helium abundance , abundance of alpha - capture elements , radiative and conductive opacities , nuclear reaction rates , neutrino energy losses , the treatments of diffusion and convection , the surface boundary conditions , and color transformations . the uncertainty in theoretical values for the red giant bump magnitude varies with metallicity between @xmath1 mag at @xmath2$}}= -2.4 $ ] and @xmath3 mag at @xmath2$}}= -1.0 $ ] . the dominant sources of uncertainty are the abundance of the alpha - capture elements , the mixing length , and the low - temperature opacities . the theoretical values of are in good agreement with observations . the uncertainty in the theoretical value of is @xmath4 at all metallicities studied . the dominant sources of uncertainty are the abundance of the alpha - capture elements , the mixing length , and the high - temperature opacities . the median value of varies from 0.44 at @xmath2$}}= -2.4 $ ] to @xmath5 at @xmath2$}}= -1.0 $ ] . these theoretical values for are in agreement with observations .
the forces of electromagnetic origin that arise between electrically neutral , unpolarized but polarizable objects are commonly known as dispersion forces @xcite . they were first addressed within the context of quantum electrodynamics ( qed ) by casimir and polder @xcite , who showed that they may be attributed to the vacuum fluctuations of the electromagnetic field . in accordance with the different nature of the interacting objects , one may distinguish between three types of dispersion forces , namely the forces between atoms in the following referred to as van der waals ( vdw ) forces , the forces between atoms and macroscopic bodies in the following referred to as casimir - polder ( cp ) forces , and the forces between macroscopic bodies in the following referred to as casimir forces . dispersion forces play a major role in the understanding of many phenomena , and they can be a useful or disturbing factor in modern applications . apart from being crucial for the understanding of many structures and processes in biochemistry @xcite , they are responsible for the remarkable climbing skills of some gecko @xcite and spider species @xcite ; the construction of atomic - force microscopes is essentially based on dispersion forces @xcite , while they are also responsible for the problem of sticking in nanotechnology @xcite . in particular , cp forces between atoms and macroscopic bodies are needed for an understanding of the adsorption of atoms and molecules to surfaces @xcite ; they can be used in atom optics to construct atomic mirrors @xcite , while they have also been found to severely limit the lifetime of atoms stored on atom chips @xcite . the study of cp forces which were first predicted for the idealized situation of a ground - state atom interacting with a perfectly conducting plate @xcite has since been greatly extended . various planar geometries like the semi - infinite half space @xcite , plates of finite thickness @xcite , two - layered plates @xcite or planar cavities @xcite have been treated , the most general planar geometry being the planar multilayer system with an arbitrary number of layers @xcite . systems with spherical @xcite or cylindrical symmetries @xcite have also been considered . it should be mentioned that some theoretical approaches ( in particular , those based on normal - mode quantization , e.g. , refs . @xcite ) require a separate treatment for each specific geometry , whereas others ( in particular , the methods based on linear response theory , e.g. , refs . @xcite ) lead to general expressions that are geometry - independent . recently , the problem has been studied within the frame of macroscopic qed in dispersing and absorbing media and an exact derivation of a general expression for the cp force has been given @xcite . although the problem of calculating cp forces ( or equivalently , the respective cp potentials ) is thus formally solved , explicit evaluation requires knowledge of the ( classical ) green tensor for the body - assisted electromagnetic field , which is ( analytically ) available only for a very limited class of geometries . in particular , inhomogeneous bodies or bodies of exotic shapes have not yet been treated . nevertheless , it shall be demonstrated in this paper that the general solution in combination with a born expansion of the green tensor may serve as a starting point for the systematic study of a wide class of geometries . furthermore , the born expansion helps making general statements about two fundamental issues regarding cp forces . first , it may answer the question of whether and to what extent cp forces are additive . second , it can be used to clarify the microscopic origin of cp forces . it is known that up to linear order in the electric susceptibility the force between an atom and a macroscopic body that may be regarded composed of atom - like constituents can be obtained by summation of two - atom ( microscopic ) vdw forces @xcite ; an analogous relation between the casimir force and cp forces can be established @xcite . it is also known that pairwise summation fails at higher order in the susceptibility @xcite , where many - atom interactions begin to play a role @xcite ; in fact , it has been shown that an infinite series of many - atom interactions must be included in order to derive the cp force between an atom and a semi - infinite dielectric half space microscopically @xcite . the article is organized as follows . in sec . [ sec1 ] the born expansion of the cp potential of an atom placed within an arbitrary arrangement of locally , linearly , and causally responding isotropic dielectric bodies is given . the results are then used to elucidate the relation to microscopic descriptions ( sec . [ sec3 ] ) , and to study some specific geometries ( sec . [ sec2 ] ) . finally , a summary is given in sec . consider a neutral , non - polar , ground - state atomic system @xmath0 such as an atom or a molecule ( briefly referred to as atom in the following ) at position @xmath1 which is placed in a free - space region within an arbitrary arrangement of linear dielectric bodies . the system of bodies is characterized by the ( relative ) permittivity @xmath2 , which is a spatially varying , complex - valued function of frequency , with the kramers - kronig relations being satisfied . within leading - order perturbation theory , the cp force on the atom due to the presence of the bodies can be derived from the cp potential ( see , e.g. , ref . @xcite ) @xmath3 according to @xmath4 ( @xmath5 @xmath6 @xmath7 ) . in eq . ( [ eq1 ] ) , @xmath8 is the ground - state polarizability of the atom in lowest order of perturbation theory [ @xmath9 @xmath6 @xmath10 @xmath11 @xmath12 , ( bare ) atomic transition frequencies ; @xmath13 @xmath6 @xmath14 , electric - dipole transition matrix elements of the atom ] , and @xmath15 is the scattering part of the classical green tensor of the body - assisted electromagnetic field , @xmath16 [ @xmath17 , vacuum part ] , which is the solution to the equation @xmath18 \tens{g}(\vec{r},\vec{r}',\omega ) = \delta(\vec{r}-\vec{r}')\tens{i}\ ] ] ( @xmath19 , unit tensor ) together with the boundary condition @xmath20 suppose now that @xmath21 with the green tensor @xmath22 , which is the solution to @xmath23 \overline{\tens{g}}(\vec{r},\vec{r}',\omega ) = \delta(\vec{r}-\vec{r}')\tens{i } , % , \ ] ] being known . a ( formal ) solution to eq . ( [ eq5 ] ) can then be given by the born series @xmath24\nonumber\\ & \quad\times \overline{\tens{g}}(\vec{r},\vec{s}_1,\omega)\cdot \overline{\tens{g}}(\vec{s}_1,\vec{s}_2,\omega)\cdots \overline{\tens{g}}(\vec{s}_k,\vec{r}',\omega),\end{aligned}\ ] ] as can be easily verified using eq . ( [ eq7.1 ] ) together with @xmath25 \sum_{k=1}^\infty\bigl(\frac{\omega}{c}\bigr)^{2k } \biggl[\prod_{j=1}^k\int\mathrm{d}^3s_j\ , \chi(\vec{s}_j,\omega)\biggr]\nonumber\\ & \quad\times \overline{\tens{g}}(\vec{r},\vec{s}_1,\omega)\cdot \overline{\tens{g}}(\vec{s}_1,\vec{s}_2,\omega)\cdots \overline{\tens{g}}(\vec{s}_k,\vec{r}',\omega)\nonumber\\ & \qquad=\bigl(\frac{\omega}{c}\bigr)^2\chi(\vec{r},\omega ) \tens{g}(\vec{r},\vec{r}',\omega).\end{aligned}\ ] ] combining eqs . ( [ eq1 ] ) , ( [ eq4 ] ) , and ( [ eq8 ] ) , we find that the cp potential can be expanded as @xmath26 where @xmath27 is the cp potential due to @xmath28 , and @xmath29 \nonumber\\ & \hspace{1ex}\times\mathrm{tr}\bigl [ \overline{\tens{g}}(\vec{r}_a,\vec{s}_1,iu)\cdot \overline{\tens{g}}(\vec{s}_1,\vec{s}_2,iu)\cdots \overline{\tens{g}}(\vec{s}_k,\vec{r}_a , iu)\bigr]\end{aligned}\ ] ] is the contribution to the potential that is of @xmath30th order in @xmath31 . the born expansion of the cp potential as given by eqs . ( [ eq9.1])([eq9 ] ) can be used to ( systematically ) calculate the potential in scenarios where a basic arrangement of bodies for which the green tensor is known is ( weakly ) disturbed , e.g. , by additional bodies or inhomogeneities such as surface roughness . let us apply eq . ( [ eq9.1])([eq9 ] ) to the case of arbitrarily shaped , weakly dielectric bodies , so that we may let @xmath32 @xmath6 @xmath33 , and hence @xmath34 where @xmath35 is given by eq . ( [ eq9 ] ) , with @xmath36\frac{e^{-\frac{u\rho}{c}}}{\rho } \nonumber\\ & = \frac{1}{3}\bigl(\frac{c}{u}\bigr)^2\delta(\bm{\rho})\tens{i } + \tens{h}_\mathrm{v}(\vec{r},\vec{r}',iu)\end{aligned}\ ] ] being the vacuum green tensor ( see , e.g. , ref . @xcite ) , where @xmath37,\ ] ] @xmath38 ( @xmath39 @xmath6 @xmath40 ; @xmath41 @xmath6 @xmath42 ; @xmath43 @xmath6 @xmath44 ) . combining eqs . ( [ eq9 ] ) and ( [ eq15 ] ) [ together with eqs . ( [ eq16])([eq16.2 ] ) ] , one easily finds that to linear order in @xmath45 the cp potential reads @xmath46 where @xmath47 in this approximation the cp force is simply a volume integral over attractive central forces , as is seen from @xmath48 = -\frac{\hat{\vec{r}}}{r^7}\bigl [ 6g_2(ur / c)-(ur / c)g^\prime_2(ur / c)\bigr]\nonumber\\ & \quad=-\frac{4\hat{\vec{r}}}{r^7 } \bigl[e^{-2x}(9\!+\!18x\!+\!16x^2\!+\!8x^3\nonumber\\ & \quad\hspace{8ex}+\!3x^4\!+\!x^5)\bigr]_{x = ur / c}\end{aligned}\ ] ] ( @xmath49 @xmath6 @xmath50 , @xmath51 @xmath6 @xmath52 ) . in the retarded ( long - distance ) limit , i.e. , @xmath53 where @xmath54 @xmath6 @xmath55 is the minimum distance of the atom to any of the bodies , @xmath56 @xmath6 @xmath57 @xmath58 @xmath59 is the lowest atomic transition frequency , and @xmath60 is the lowest resonance frequency of the dielectric material , the exponential factor in @xmath61 effectively limits the @xmath62-integral in eq . ( [ eq17 ] ) to a region where @xmath63 so eq . ( [ eq17 ] ) reduces to @xmath64 in the nonretarded ( short - distance ) limit , i.e. , @xmath65 where @xmath66 @xmath6 @xmath67 is the maximum distance of the atom to any body part , @xmath68 @xmath6 @xmath69 @xmath58 @xmath59 is the highest atomic transition frequency , and @xmath70 is the highest resonance frequency of the dielectric material , the factors @xmath71 and @xmath72 effectively limit the @xmath62-integral in eq . ( [ eq17 ] ) to a region where @xmath73 @xmath58 @xmath74 @xmath75 @xmath33 , so we may set @xmath76 resulting in @xmath77 the second - order contribution @xmath78 can be separated into a single - point term and a two - point correlation term , @xmath79 as can be seen from eqs . ( [ eq9 ] ) and ( [ eq15 ] ) for @xmath30 @xmath58 @xmath80 . the single - point term @xmath81 which arises from the @xmath82-function in eq . ( [ eq15 ] ) , differs from the first - order contribution @xmath83 according to @xmath84 hence its asymptotic retarded and nonretarded forms can be obtained by applying the replacement ( [ eq24.3 ] ) to eqs . ( [ eq21 ] ) and ( [ eq24 ] ) , respectively . the two - point correlation term is derived to be @xmath85 where @xmath86,\end{aligned}\ ] ] with the abbreviations @xmath87 having been introduced [ recall eqs . ( [ eq16.1 ] ) and ( [ eq16.2 ] ) ] . note that the two - point contribution to the cp force , eq . ( [ eq24.4 ] ) , is a double spatial integral , the integrand of which can be attractive or repulsive , depending on the angles in the triangle formed by the vectors @xmath88 , @xmath89 , and @xmath90 . in the retarded limit , where the inequalities ( [ eq19 ] ) hold , the @xmath62-integral is again effectively limited to a region where the approximations ( [ eq20 ] ) are valid , so eq . ( [ eq24.4 ] ) reduces to @xmath91 we introduce the notation @xmath92 and perform the @xmath62-integral with the aid of the relation @xmath93 exploiting the triangle formula @xmath94 [ which is a trivial consequence of eqs . ( [ eq24.6])([eq24.8 ] ) ] by adding the expression @xmath95\nonumber\\ & + 7\bigl[\alpha^4(\beta^2+\gamma^2)+\beta^4(\gamma^2+\alpha^2 ) + \gamma^4(\alpha^2+\beta^2)\bigr]\nonumber\\ & + 12(\alpha^3\beta^3+\beta^3\gamma^3+\gamma^3\alpha^3 ) + 12\alpha\beta\gamma(\alpha^3+\beta^3+\gamma^3)\nonumber\\ & + 52\alpha\beta\gamma\bigl[\alpha\beta(\alpha+\beta ) + \beta\gamma(\beta+\gamma)+\gamma\alpha(\gamma+\alpha ) \bigr]\nonumber\\ & + 138\alpha^2\beta^2\gamma^2\bigr\}\,\end{aligned}\ ] ] to eq . ( [ eq24.9 ] ) , the result may be written in the form @xmath96,\end{aligned}\ ] ] where @xmath97,\\ \label{eq24.17 } & f_3(\alpha,\beta,\gamma)= \nonumber\\&\quad -1 - 39\frac{\sigma_2}{\sigma_1 ^ 2}+17\frac{\sigma_3}{\sigma_1 ^ 3 } + 72\frac{\sigma_2 ^ 2}{\sigma_1 ^ 4 } -75\frac{\sigma_2\sigma_3}{\sigma_1 ^ 5 } + 20\frac{\sigma_3 ^ 2}{\sigma_1 ^ 6}\end{aligned}\ ] ] [ recall eqs . ( [ eq24.6])([eq24.8 ] ) as well as eq . ( [ eq24.10 ] ) ] . in the nonretarded limit , where the inequalities ( [ eq22 ] ) hold , the @xmath62-integral in eq . ( [ eq24.4 ] ) is effectively limited to a region where @xmath98\end{aligned}\ ] ] [ recall eq . ( [ eq24.12 ] ) ; note that @xmath99 @xmath100 @xmath101 , so eq . ( [ eq24.4 ] ) reduces to @xmath102 [ recall eq . ( [ eq24.6])([eq24.8 ] ) ] . in order to gain insight into the microscopic origin of the cp potential as given by eq . ( [ eq1 ] ) , let us suppose that the susceptibility @xmath31 is due to a collection of atoms of polarizability @xmath103 and apply the well - known clausius - mosotti formula ( see , e.g. , ref . @xcite ) @xmath104 where @xmath105 is the number density of the ( medium ) atoms for @xmath106 @xmath58 @xmath107 . since @xmath31 is the fourier transform of a ( linear ) response function , it must satisfy the condition @xmath108 which , with respect to eq . ( [ eq28 ] ) , implies that the inequality @xmath109 must hold . substituting the susceptibility from eq . ( [ eq28 ] ) into the born series of the cp potential @xmath110 as given by eqs . ( [ eq9.1])([eq9 ] ) , taking into account that the green tensor @xmath111 can be decomposed as @xmath112 [ recall eqs . ( [ eq7.1 ] ) and ( [ eq15 ] ) ] , and recalling the inequality ( [ eq28.1.2 ] ) , it can be shown after some lengthy calculation that the born series can be rewritten as an expansion of @xmath110 in terms of many - atom interaction potentials @xmath113 ( see app . [ appc ] ) , @xmath114 u_{ab\ldots b}(\vec{r}_a,\vec{s}_1,\ldots,\vec{s}_l),\end{aligned}\ ] ] where @xmath115.\end{aligned}\ ] ] here the symbol @xmath116 introduces symmetrization with respect to @xmath117 according to the rule @xmath118\nonumber\\ & = \!\!\sum_{\pi\in \overline{p}(j)}\!\ ! \mathrm{tr}\bigl [ \overline{\tens{h } } ( \vec{r}_{\pi(1)},\vec{r}_{\pi(2)},\omega)\cdots \overline{\tens{h}}(\vec{r}_{\pi(j)},\vec{r}_{\pi(1)},\omega)\bigr].\end{aligned}\ ] ] the sum in eq . ( [ eq28.5 ] ) runs over the maximal number of permutations @xmath119 @xmath120 @xmath121 @xmath122 @xmath123 [ @xmath123 being the permutation group of the numbers @xmath124 that can not be obtained from one another via ( a ) a cyclic permutation or ( b ) the reverse of a cyclic permutation ( cf . [ appc ] ) . the potential @xmath125 is nothing but the ( microscopic ) vdw potential describing the mutual interaction of a ( test ) atom @xmath0 at position @xmath126 and @xmath127 ( medium ) atoms at different positions @xmath128 . the very general equation ( [ eq1 ] ) , which follows from qed in causal media , gives the cp potential of an atom in the presence of macroscopic dielectric bodies in terms of the atomic polarizability and the ( scattering ) green tensor of the body - assisted maxwell field , with the bodies being characterized by a spatially varying dielectric susceptibility that is a complex function of frequency . equation ( [ eq28.1.1 ] ) clearly shows that when the susceptibility is of clausius - mosotti type , i.e. , eq . ( [ eq28 ] ) [ together with the inequality ( [ eq28.1.2 ] ) ] applies , then the cp potential is in fact the result of a superposition of all possible microscopic many - atom vdw potentials between the atom under consideration and the atoms forming the bodies . note that for the vacuum case , @xmath28 @xmath6 @xmath33 , a similar conclusion has been drawn by combining normal mode quantization with the ewald - oseen extinction theorem @xcite . moreover , from the derivation given in app . [ appc ] it can be seen that when eqs . ( [ eq28.1.1 ] ) and ( [ eq28.11 ] ) [ together with the inequality ( [ eq28.1.2 ] ) ] hold , then the susceptibility must necessarily have the form of eq . ( [ eq28 ] ) . from the above it is clear that in order to establish the identity ( [ eq28.1.1 ] ) to all orders in @xmath129 , it is crucial to employ the exact relation ( [ eq28 ] ) between macroscopic susceptibility and microscopic atomic polarizability rather than its linearized version @xmath31 @xmath58 @xmath130 , which is is known to be sufficient for finding a correspondence between macroscopic and microscopic potentials to linear order in @xmath45 ( or @xmath129 , respectively ) @xcite . it should be pointed out that in more general cases where the susceptibility is not of the form ( [ eq28 ] ) the result of applying the born expansion can not be disentangled into spatial integrals over microscopic vdw potentials in the way given by eq . ( [ eq28.1.1 ] ) together with eq . ( [ eq28.11 ] ) . obviously , the basic constituents of the bodies can no longer be approximated by well localized atoms . according to eq . ( [ eq7 ] ) , the expansion in eq . ( [ eq28.1.1 ] ) does not necessarily refer to all bodies . hence from eq . ( [ eq28.11 ] ) it follows that the many - atom vdw potential on an arbitrary dielectric background described by @xmath28 reads @xmath131,\end{aligned}\ ] ] the derivation being unique when requiring the vdw potentials to be fully symmetrized . in particular , for @xmath132 @xmath58 @xmath80 [ where @xmath133 $ ] @xmath58 @xmath33 , so the sum in the r.h.s of eq . ( [ eq28.5 ] ) contains only one term and symmetrization is not necessary ] , eq . ( [ eq28.12 ] ) agrees with the result that can be found by calculating the change in the zero - point energy of the system in leading - order perturbation theory @xcite . in the simplest case of vacuum background , i.e. , @xmath28 @xmath6 @xmath33 , @xmath134 in eq . ( [ eq28.12 ] ) becomes @xmath135 [ recall eqs . ( [ eq16])([eq16.2 ] ) ] , leading to agreement with earlier results @xcite . bearing in mind the general relation ( [ eq28.1.1 ] ) between the cp potential and many - atom vdw potentials , explicit expressions for the two- and three - atom vdw potentials on vacuum background can easily be obtained from formulas given in sec . [ sec1 ] together with eq . ( [ eq28 ] ) . from eqs . ( [ eq17 ] ) , ( [ eq21 ] ) , and ( [ eq24 ] ) one can infer the well - known result @xcite @xmath136 ( @xmath49 @xmath6 @xmath137 ) , recall eq . ( [ eq18 ] ) , which reduces to @xmath138 in the retarded limit , @xmath139 with @xmath54 @xmath6 @xmath140 and @xmath141 @xmath6 @xmath142 , and to @xmath143 in the nonretarded limit , @xmath144 with @xmath66 @xmath6 @xmath140 and @xmath145 @xmath6 @xmath146 . similarly , eq . ( [ eq24.4 ] ) implies that @xmath147 ( @xmath148 @xmath6 @xmath149 , @xmath150 @xmath6 @xmath151 for @xmath152 @xmath58 @xmath153 ) , recall eq . ( [ eq24.5 ] ) , in agreement with the result found in refs . @xcite . equations ( [ eq24.14 ] ) and ( [ eq24.19 ] ) show that eq . ( [ eq35 ] ) simplifies to @xmath154\end{aligned}\ ] ] [ @xmath155 @xmath6 @xmath156 , recall eq . ( [ eq24.10 ] ) and eqs . ( [ eq24.15])([eq24.17 ] ) ] in the retarded limit [ eq . ( [ eq33.1 ] ) with @xmath54 @xmath6 @xmath157 @xmath158 @xmath159 and @xmath141 @xmath6 @xmath160 , and reduces to the axilrod - teller potential @xcite @xmath161 } { 64\pi^4\varepsilon_0 ^ 3r_{12}^3r_{23}^3r_{31}^3 } \nonumber\\&\qquad \times\int_0^{\infty } \mathrm{d } u\ , \alpha_a(iu)\alpha_b(iu)\alpha_c(iu)\end{aligned}\ ] ] in the nonretarded limit [ eq . ( [ eq34.1 ] ) with @xmath66 @xmath6 @xmath162 @xmath158 @xmath159 and @xmath145 @xmath6 @xmath163 . let us use the born series given in sec . [ sec1 ] to calculate the ( leading contributions to the ) cp potential for some specific geometries and begin with a ground - state atom placed on the symmetry axis of a homogeneous dielectric ring of susceptibility @xmath164 , having radius @xmath165 , ( circular ) cross section @xmath166 ( where @xmath167 @xmath75 @xmath165 ) , and volume @xmath168 @xmath58 @xmath169 , the atom being separated from the center of the ring by a distance @xmath170 ( fig . [ fig1 ] ) . from fig . [ fig1 ] we see that @xmath171 @xmath172 @xmath173 @xmath6 @xmath174 for @xmath167 @xmath75 @xmath175 , so an evaluation of the ( trivial ) volume integral in eq . ( [ eq17 ] ) results in the first - order cp potential @xmath176 which is attractive , as expected . in the retarded limit [ eq . ( [ eq19 ] ) with @xmath54 @xmath58 @xmath174 ] eq . ( [ eq25 ] ) reduces to @xmath178 [ cf . ( [ eq21 ] ) ] , while in the nonretarded limit [ eq . ( [ eq22 ] ) with @xmath66 @xmath58 @xmath174 ] one easily finds @xmath179 [ cf . ( [ eq24 ] ) ] . in both limiting cases the cp potential thus reduces to simple asymptotic power laws in @xmath180 , where as usual the leading ( inverse ) power is increased by one when going from the nonretarded to the retarded limit . the first ( single point ) second - order correction term @xmath181 can simply be obtained from eqs . ( [ eq25])([eq27 ] ) by means of the replacement ( [ eq24.3 ] ) , while the calculation of the second ( two - point ) term @xmath182 is a lot more difficult due to the factor @xmath183 @xmath11 @xmath184 . we find ( see app . [ appb ] ) @xmath185 in the retarded limit and @xmath186 in the nonretarded limit . recalling eqs . ( [ eq24.1 ] ) and ( [ eq24.3 ] ) , eqs . ( [ eq26 ] ) and ( [ eq27.10 ] ) imply that up to quadratic order in @xmath45 we have @xmath187\ ] ] in the retarded limit , while eqs . ( [ eq27 ] ) and ( [ eq27.11 ] ) show that in the nonretarded limit @xmath188.\end{aligned}\ ] ] the uncertainty in the magnitude of the contribution quadratic in @xmath45 is due to the approximations made when calculating @xmath189 ( cf . app . [ appb ] ) . however , irrespective of these approximations , eqs . ( [ eq27.12 ] ) and ( [ eq27.13 ] ) show that the leading non - additive correction @xmath190 to the linear result @xmath191 does not change the powers in the asymptotic retarded and nonretarded distance laws ( @xmath192 and @xmath193 , respectively ) , but merely modifies the constants of proportionality . a similar result has been found when studying a dielectric half space @xcite . the explicit evaluation of multiple spatial integrals [ being the main difficulty when evaluating @xmath194 can in fact be avoided in many cases by an appropriate decomposition of the body of interest , as shall be demonstrated in the following . to that end , let us decompose the body described by @xmath31 [ recall eq . ( [ eq7 ] ) ] into smaller bodies numbered by @xmath195 , so that @xmath196 where @xmath197 0 \mbox { for } \vec{r}\notin v_n . \end{array}\right.\ ] ] substituting eqs . ( [ eq10 ] ) and ( [ eq11 ] ) into eq . ( [ eq9 ] ) , and slightly rearranging the terms , we obtain @xmath198 where @xmath199 with @xmath200 is the sum of all @xmath127-body contributions of order @xmath30 in @xmath45 . in eq . ( [ eq11.2 ] ) , @xmath201\nonumber\\ & \times \mathrm{tr}\bigl [ \overline{\tens{g}}(\vec{r}_a,\vec{s}_1,iu ) \cdot\overline{\tens{g}}(\vec{s}_1,\vec{s}_2,iu)\cdots \overline{\tens{g}}(\vec{s}_k,\vec{r}_a , iu)\bigr],\end{aligned}\ ] ] and the notation latexmath:[\ ] ] in analogy to the procedure outlined in ref . @xcite , the retarded limit may conveniently be treated by introducing the new integration variable @xmath325 @xmath58 @xmath329 , transforming integrals according to @xmath330 applying the approximation ( [ eq20 ] ) , and carrying out the @xmath325-integrals . application of this procedure to eqs . ( [ a9 ] ) , ( [ a10 ] ) , and ( [ a11 ] ) leads to @xmath331 where we have introduced the definitions ( [ eq13.17 ] ) and ( [ eq13.18 ] ) . combining eqs . ( [ a14])([a16 ] ) in accordance with eq . ( [ eq13 ] ) and using eqs . ( [ a3 ] ) and ( [ a4 ] ) , we arrive at eq . ( [ eq13.15 ] ) . the asymptotic behavior of eqs . ( [ a9])([a11 ] ) in the nonretarded limit may be obtained by transforming the integral according to @xmath332 retaining only the leading power of @xmath333 , carrying out the @xmath334-integral and discarding higher - order terms in @xmath335 ( cf . @xcite ) , resulting in @xmath336 recall eqs . ( [ eq13.17 ] ) and ( [ eq13.18 ] ) . upon using eq . ( [ eq13 ] ) , eqs . ( [ a18 ] ) ([a20 ] ) lead to eq . ( [ eq13.16 ] ) , where we have neglected the term proportional to @xmath337 in consistency with the nonretarded limit and used eqs . ( [ a5 ] ) and ( [ a6 ] ) .	within leading - order perturbation theory , the casimir - polder potential of a ground - state atom placed within an arbitrary arrangement of dispersing and absorbing linear bodies can be expressed in terms of the polarizability of the atom and the scattering green tensor of the body - assisted electromagnetic field . based on a born series of the green tensor , a systematic expansion of the casimir - polder potential in powers of the electric susceptibilities of the bodies is presented . the born expansion is used to show how and under which conditions the casimir - polder force can be related to microscopic many - atom van der waals forces , for which general expressions are presented . as an application , the casimir - polder potentials of an atom near a dielectric ring and an inhomogeneous dielectric half space are studied and explicit expressions are presented that are valid up to second order in the susceptibility . * pacs * 12.20.-m quantum electrodynamics 34.50.dy interactions of atoms and molecules with surfaces ; photon and electron emission ; neutralization of ions 34.20.-b interatomic and intermolecular potentials and forces , potential energy surfaces for collisions 42.50.nn quantum optical phenomena in absorbing , dispersive and conducting media
the solar wind ( sw ) plasma , an ideal laboratory for the study of collisionless plasma dynamics , is mostly found in a turbulent state @xcite . subproton - scale ( `` dissipation range '' ) turbulence in the sw has become a major research topic over the past decade , both for in - situ satellite measurements @xcite and for numerical @xcite and theoretical @xcite studies . spacecraft observations provide important constraints on turbulent spectra , revealing the presence of breaks in the electromagnetic fluctuations around the proton kinetic scales @xcite . at subproton scales , typical slopes for the magnetic energy spectrum are found to be in the range @xmath2 $ ] , while preliminary results about its electric counterparts are in the range @xmath3 $ ] . from a theoretical point of view , possible explanations for the observed spectra are the development of a kinetic alfvn wave ( kaw ) cascade and/or a whistler cascade @xcite . however , the predicted energy spectra are the same for the two cases and thus auxiliary methods have been suggested in order to identify the exact nature of turbulent fluctuations @xcite . observational evidence points towards a kaw - dominated scenario for a @xmath4 plasma @xcite ( @xmath1 is the ratio between the thermal and the magnetic pressures ) , although contradictory results have also been reported @xcite . theoretical studies , on the other hand , have suggested that oblique kaws and whistlers could coexist as the plasma parameters vary in space and time @xcite . so far , numerical simulations have focused only on one scenario at a time , not on a possible coexistence or a transition between those cascades , leaving such a question as an open problem in sw turbulence research . in this letter , we wish to tackle the fundamental question of a possible dependence of the physics of subproton - scale kinetic turbulence on the plasma @xmath1 parameter by carrying out high - resolution 2d3v simulations of forced plasma turbulence as described by a hybrid vlasov maxwell ( hvm ) model with fluid electrons . while not retaining electron kinetic effects , this approach allows for both kaws and whistlers to be present , and it fully captures the ion kinetic physics . besides , this 2d3v setting allows us to include large `` fluid '' scales while still fully resolving subproton scales , which is not currently possible in 3d3v due to computational limitations . due to the intrinsic anisotropy of the turbulent mhd cascade , and to the strong damping of the parallel modes via resonances @xcite , we also expect such `` 2.5d '' simulations to retain some important dynamical features of the fully 3d case . in the hvm model , fully kinetic ions are coupled with massless fluid electrons @xcite . the hvm equations normalized with respect to the ion mass @xmath5 , the ion gyrofrequency @xmath6 , the alfvn speed @xmath7 and the ion skin depth @xmath8 are given by @xmath9 where @xmath10 is the ion distribution function , @xmath11 and @xmath12 are the electric and magnetic fields , respectively , and @xmath13 is the current density . we assume quasi - neutrality @xmath14 . the number density @xmath15 and the ion mean velocity @xmath16 are computed as the velocity moments of @xmath17 . an isothermal equation of state is assumed for the scalar electron pressure @xmath18 , with a given initial electron - to - ion temperature ratio @xmath19 . @xmath20 is a @xmath21-correlated in time , external forcing that injects momentum in the system with a prescribed average power density @xmath22 . its correlation tensor in fourier space reads @xmath23 $ ] , where brackets denote ensemble averaging , @xmath24 is a wave vector , @xmath25 is a scalar function depending on the amplitude of the wavenumber only , and @xmath26 and @xmath27 , respectively , quantify the relative degrees of incompressibility and compressibility of the forcing . in all simulations presented in this letter , we use @xmath28 . while it may overestimate the actual compressible component of the driving in the sw context , this choice can be justified by the lack of scale separation in the simulations between the driving and ion scales , at which a mixture of compressible and incompressible fluctuations is found in the solar wind ( see , e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , and by the desire to not artificially direct energy into a particular mode at large scale . our numerical implementation of this forcing is a direct transposition of a widely used hydrodynamic technique @xcite . equations ( [ eq : hvm_vlasov])-([eq : hvm_maxwell ] ) are solved in a 2d3v phase space using an eulerian algorithm @xcite , with fully three - dimensional vector fields . the initial condition is a maxwellian plasma in a constant perpendicular magnetic field @xmath29 . the system is initially perturbed by random , 3d , large - scale , small - amplitude magnetic fluctuations , @xmath30 ( with wavenumbers @xmath31 in the range @xmath32 ) . the driving procedure and amplitude ( in code units ) is identical for all the cases documented below . the average power input of @xmath33 is @xmath34 and the forcing acts on the smallest wave numbers of the system , @xmath35 , thus injecting energy only at the largest scales admitted by our numerical box . in the following , we consider three different initial plasma beta values ( @xmath36 , @xmath37 and @xmath38 ) and a temperature ratio @xmath39 , i.e. , typical of sw parameters . we use @xmath40 uniformly distributed grid points to discretize a squared simulation box with @xmath41 , corresponding to a resolution @xmath42 . doubly periodic boundary conditions are imposed , and the spectral domain spans a perpendicular wavenumber range @xmath43 . spectral filters @xcite on the electromagnetic fields are applied during the simulation , in order to avoid spurious numerical effects at the smallest scales : this determines the cutoff in the energy spectra at @xmath44 . the velocity domain is limited by @xmath45 in each @xmath46-direction , with @xmath47 uniformly distributed grid points , so @xmath48 . the time step is constrained by the cfl conditions @xcite . we first investigate the spectral properties of the statistically quasi - steady turbulent state and whether they reproduce the phenomenology expected for kaws or whistlers . the analysis is performed at about @xmath49 , where @xmath50 is the outer - scale nonlinear time ( estimated from a kolmogorov argument ) . in this regime , the average modulus of the in - plane magnetic field , @xmath51 , remains relatively low , @xmath52 . nevertheless , larger values ( up to @xmath53 ) are observed locally in space and time , and coherent magnetic structures are formed . on the one hand , non - negligible in - plane magnetic fluctuations allow for finite @xmath54 ( @xmath55 is the local unit vector along @xmath12 ) , i.e. , for parallel kinetic effects and oblique waves with non - zero @xmath56 . on the other hand , the local in - plane magnetic field turns out to be randomly oriented in the fully turbulent regime , and the spectra are globally isotropic in the @xmath57-plane . therefore a shell - averaging technique can be adopted in that spectral plane ( @xmath58-reduction ) , without being polluted by any preferential direction . moreover , spectra are time - averaged over about @xmath59 . in fig . [ fig : spectra_comparison ] we plot the total magnetic and electric energy spectra , @xmath60 and @xmath61 . the @xmath62 spectrum at @xmath63 exhibits a slope close to @xmath64 , although this result should be treated with caution because of the vicinity of the injection scale and of the small extent of the range ( especially at @xmath65 ) . and @xmath61 ( left and right panel , respectively ) , for @xmath66 , blue , green and red color ( grey scale ) , respectively.,scaledwidth=105.0% ] and @xmath61 ( left and right panel , respectively ) , for @xmath66 , blue , green and red color ( grey scale ) , respectively.,scaledwidth=105.0% ] at @xmath67 , the spectral index changes for all three cases and lies between @xmath68 and @xmath69 , in general agreement with spacecraft observations @xcite . on the one hand , at @xmath36 and @xmath37 , the @xmath70 ( not shown ) and @xmath62 spectra for @xmath67 appear to be fitted better with a @xmath68 slope ( fig . [ fig : spectra_kaw - or - whistler_a ] ) , while the @xmath71 ( and @xmath72 , at @xmath73 ) spectrum is well fitted by a @xmath74 slope ( fig . [ fig : spectra_kaw - or - whistler_b ] ) . a @xmath68 slope would be in agreement with theory for fluctuations forming two - dimensional structures ( * ? ? ? * ; coherent structures are indeed visible in our simulations ) , whereas @xmath74 is the prediction of the standard theories of kaw and whistler turbulence @xcite . on the other hand , at @xmath65 , all the spectra are steeper , fitted by a @xmath69 slope . steepening of the spectra are possibly due to features not included in the standard theories , such as compressibility and/or wave damping effects ( * ? ? ? * ; * ? ? ? * ; see also fig . [ fig : damping_kaw - whistler ] ) . in the electric energy , @xmath75 , at @xmath65 and @xmath76 , a power law steeper than @xmath64 is seen at @xmath77 , whereas a spectral index between @xmath78 and @xmath79 is observed at @xmath67 . at @xmath73 , instead , a bump is present at @xmath80 that makes the spectrum appear steeper , with a slope of @xmath81 and it only partially agrees with the other two cases at @xmath82 ( fig . [ fig : spectra_comparison ] ) . in all the @xmath1 cases , the @xmath75 spectrum at @xmath77 is dominated by the mhd term , @xmath83 , whereas at @xmath67 it is dominated by the hall term , @xmath84 ( cf . ( [ eq : hvm_ohm ] ) ) . the electron pressure term , @xmath85 , is always found to be sub - dominant with respect to @xmath86 . finally , the electric energy overcomes its magnetic counterpart at @xmath87 , regardless of the @xmath88-scale position . we now investigate the nature of turbulent fluctuations in the different @xmath1 regimes . first , we compare the levels of magnetic and density spectra , @xmath62 and @xmath89 ( with @xmath90[1+{\beta_{\rm i}}(1+\tau)/2]$ ] ) , a method to distinguish between kaw ( @xmath91 ) and whistler ( @xmath92 ) turbulence @xcite . second , we check if the relation @xmath93 ( with @xmath94 ^ 2 $ ] ) between the density and parallel magnetic spectra , expected for kaw fluctuations , is satisfied @xcite . we stress that the two methods are not conclusive if taken separately , but are complementary to each other and must therefore be inspected accordingly . in fig . [ fig : spectra_kaw - or - whistler_a ] we compare the magnetic and the normalized density spectra , @xmath62 and @xmath89 , as obtained in our simulations . the main result is that the turbulence is mediated by magnetosonic / whistler ( ms / w ) fluctuations at @xmath36 , whereas the dynamics at @xmath73 appear to be dominated by alfvn wave / kinetic alfvn wave ( aw / kaw ) turbulence . at @xmath65 , instead , there is a signature of a transition at @xmath95 , from a ms to a kaw regime . in order to confirm this scenario , in fig . [ fig : spectra_kaw - or - whistler_b ] we show the comparison between @xmath71 and @xmath96 . in particular , at @xmath36 , a significant disagreement between the two quantities remains even at @xmath44 , thus providing a confirmation of the whistler - dominated regime inferred from fig . [ fig : spectra_kaw - or - whistler_a ] . at @xmath73 , we find @xmath93 through the entire @xmath58 range , thus confirming the kaw - dominated scenario . at @xmath65 , @xmath71 and @xmath96 differ by more than an order of magnitude for @xmath97 , whereas the relation @xmath93 holds well for @xmath67 . this supports the interpretation of a transition from an ms dynamics at large scales to a kaw regime at smaller scales , for @xmath65 ( see also sec . [ subsec : interpret ] and fig . [ fig : damping_kaw - whistler ] ) . further evidence leading to the above conclusions is provided by inspecting the magnetic compressibility , @xmath98 , and by the predominantly perpendicular heating of the ions at low @xmath1 , @xmath99 ( not shown here ) . we finally caution that these results may be dependent on the details of how the turbulence is driven . while we concluded from several test simulations ( not shown ) that the qualitative results presented here do not depend significantly on the resolution and/or on the forcing amplitude , we also found that ms / w waves are not excited in complementary test simulations conducted with a purely incompressible ( perhaps somewhat idealized ) driving ( @xmath100 , @xmath101 ) . studying the detailed dependence of this kind of turbulence on the driving lies outside the scope of the present paper , but may also be relevant to the solar wind context and will therefore be worth exploring in the future . + there are several examples of wave - supporting turbulent systems where linear physics leaves an imprint on the nonlinear dynamics even in strong turbulence regimes @xcite . while it may not apply quantitatively in such regimes , linear theory may still provide some interesting physical insights into the dynamics at work in that case . , for the aw / kaw ( black ) and ms / w ( cyan ) branches at @xmath102 ( left panel ) and @xmath103 ( central panel ) , for a propagation angle of @xmath104 , and at @xmath105 ( right panel ) , for @xmath106.,scaledwidth=115.0% ] , for the aw / kaw ( black ) and ms / w ( cyan ) branches at @xmath102 ( left panel ) and @xmath103 ( central panel ) , for a propagation angle of @xmath104 , and at @xmath105 ( right panel ) , for @xmath106.,scaledwidth=115.0% ] , for the aw / kaw ( black ) and ms / w ( cyan ) branches at @xmath102 ( left panel ) and @xmath103 ( central panel ) , for a propagation angle of @xmath104 , and at @xmath105 ( right panel ) , for @xmath106.,scaledwidth=115.0% ] a possible interpretation for the transition reported above is in terms of the linear properties of the magnetosonic / whistler ( ms / w ) and of the alfvn / kinetic alfvn ( aw / kaw ) modes . in fig . [ fig : damping_kaw - whistler ] we display the ratio of the damping rate to the real frequency , @xmath107 , for the aw / kaw and the ms / w branches of the hvm system , eqs . ( 1)-(3 ) , within our simulation parameters . a representative propagation angle of @xmath108 has been estimated by @xmath109 . at @xmath36 , the aw / kaw is weakly damped for @xmath110 and undergoes a complete resonant absorption as @xmath111 for @xmath112 . the ms / w mode is instead practically undamped for @xmath110 , except for a well - separated series of peaks representing the crossing of the resonant surfaces @xmath113 ( @xmath114 , @xmath115 , @xmath116 , @xmath117 ) . then , for @xmath112 , the peaks form a quasi - continuum of wave damping , but still more than one order of magnitude lower than that of the aw / kaw counterpart . this would suggest a complete absorption of kaws for @xmath118 at @xmath36 , leaving this regime whistler - dominated . at @xmath73 , the frequency - normalized damping rates of the two modes are comparable and an extrapolation to the turbulent state is not obvious ( a comparison of just @xmath119 would show a slightly higher damping of the ms / w branch , but still of the same order of magnitude ) . however , in this regime the aw / kaw mode is not completely absorbed anymore by the ion cyclotron resonance , consistent with a kaw - dominated cascade inferred from the simulations . at @xmath65 , instead , the frequency - normalized damping rates exhibit a transition at @xmath120 : for @xmath121 , the aw / kaw branch is more damped than the ms / w counterpart , whereas at @xmath122 the contrary holds ( this transition is much more pronounced in the pure damping rates , @xmath119 ) . this reflects the behavior shown in fig . 23 at @xmath65 , from which a transition from an ms - regime at @xmath123 to a kaw - dominated scenario for @xmath122 was inferred . we point out that the electron damping on both the aw / kaw and the ms / w modes is missing in the hvm system . this represents a limitation of this model , which should be properly investigated as appropriate numerical resources become available . nevertheless , we note that the interpretation proposed above is in qualitative agreement with previous linear studies in a full - kinetic framework @xcite and with observations about the relevance of cyclotron - resonant dissipation mechanisms in some regimes of sw turbulence @xcite . we presented the first high - resolution simulations of 2d3v forced hybrid - kinetic turbulence ranging from magnetohydrodynamic scales to scales well below the ion gyroradius . the spectral properties of the simulated turbulence , such as power - law exponents and spectral breaks at ion scales , are in agreement with the existing theory of subproton - scale turbulence and close to the observed sw spectra . moreover , we find that small - scale turbulence in this driven 2d3v setup mainly involves magnetosonic / whistler fluctuations at low @xmath1 , and kaws at somewhat higher @xmath1 . we found that this transition correlates with a change in the relative strength of the damping of the underlying wave modes , suggesting that cyclotron - resonant damping may be relevant in this context . we point out that this scenario is not mutually exclusive of other important effects involving nonlinearities , such as the presence of coherent structures also spotted in the simulations , and they can in fact be coupled with each other . while the model used in this paper presents some limitations and does not accommodate all the dynamical complexity of the sw , the results suggest a possible dependence of subproton - scale kinetic turbulence on the plasma @xmath1 parameter that may be relevant to the time and space variability of the sw . high - resolution simulations in three spatial dimensions , also including electron kinetic effects and different forms of driving , appear necessary to further our understanding of this problem , but will have to wait until computational capabilities become available . the authors acknowledge useful discussions with j. m. tenbarge , a. a. schekochihin , w. dorland , m. kunz , r. bruno and f. pegoraro . we gratefully acknowledge the anonymous referee , whose in - depth comments helped to significantly improve the presentation and discussion of the results . the research leading to these results has received funding from the european research council under the european union s seventh framework programme ( fp7/2007 - 2013)/erc grant agreement no . this project has received funding from the euratom research and training programme 2014 - 2018 . this work was facilitated by the max - planck / princeton center for plasma physics . the simulations were performed on fermi ( cineca , italy ) and on hydra ( rechenzentrum garching , germany ) .	a long - lasting debate in space plasma physics concerns the nature of subproton - scale fluctuations in solar wind ( sw ) turbulence . over the past decade , a series of theoretical and observational studies were presented in favor of either kinetic alfvn wave ( kaw ) or whistler turbulence . here , we investigate numerically the nature of the subproton - scale turbulent cascade for typical sw parameters by means of unprecedented high - resolution simulations of forced hybrid - kinetic turbulence in two real - space and three velocity - space dimensions . our analysis suggests that small - scale turbulence in this model is dominated by kaws at @xmath0 and by magnetosonic / whistler fluctuations at lower @xmath1 . the spectral properties of the turbulence appear to be in good agreement with theoretical predictions . a tentative interpretation of this result in terms of relative changes in the damping rates of the different waves is also presented . overall , the results raise interesting new questions about the properties and variability of subproton - scale turbulence in the sw , including its possible dependence on the plasma @xmath1 , and call for detailed and extensive parametric explorations of driven kinetic turbulence in three dimensions .
the inverse compton scattering of cosmic microwave background ( cmb ) photons by hot electrons in galaxy clusters leads to a distortion of the cmb spectrum along the line of sight , known as the sunyaev - zeldovich ( sz , @xcite ) effect . the magnitude of the sz effect is determined by the comptonization parameter of the cluster gas , @xmath3 , which is proportional to the line of sight integral of the gas pressure . it was recognized swiftly @xcite that for an assumed , simple geometry , the combination of x - ray and sz observations can be used to measure the angular diameter distance to a cluster . the ratio of the observed ( based on radio / sub - mm observations ) and predicted ( based on x - ray observations ) sz signals is proportional to the square of the angular diameter distance , making this , potentially , an exceptionally powerful technique for probing the cosmic distance scale . although the sz effect is now employed frequently to determine extragalactic distances ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , most studies to date have relied on the application of the @xmath4-model @xcite in their x - ray analyses . the @xmath4-model provides a simple analytical approximation to the spatial distribution of the cluster gas . these studies have also generally assumed isothermality and relied on broad beam measurements of the mean , emission - weighted temperature when calculating the gas pressure . with the chandra x - ray observatory and xmm - newton it is now possible to measure the temperature , density and pressure profiles of the cluster gas precisely . the surface brightness distribution , in particular , can be resolved on sub - arcsecond scales . chandra observations have shown that @xmath4-models do not , generally , provide good descriptions of the x - ray gas distributions in the cores of dynamically - relaxed clusters , and that the gas temperature is not isothermal but drops sharply within cluster cores ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? the main limitation of the new x - ray observations is that ( due to background levels and restricted fields of view ; @xmath5 arcmin@xmath6 for the chandra acis - s detector ) the gas temperature can only be measured directly out to radii @xmath7 one third of the virial radius for most clusters . since predictions of the sz effect require the line - of - sight pressure integral through the cluster , extrapolation of the x - ray results is therefore required for combined x - ray / sz work . in this paper we present a recipe for calculating the predicted sz effect in clusters which makes use of the spatially - resolved spectroscopic techniques described in our earlier work , and which includes a new method for extrapolating the x - ray pressure profiles beyond the directly observed region . analytical formulae for the electron density and temperature profiles of the x - ray gas at large radii are obtained under the assumption of hydrostatic equilibrium . these can be computed easily and used for sz analysis . we show that our method , which takes full account of the observed density and temperature profiles of the x - ray emitting gas , provides a significant improvement in accuracy with respect to the standard isothermal @xmath4-model . the outline of this paper is as follows : in section 2 we review the standard @xmath4-model approach and describe our improved method . in section 3 we apply our method to chandra x - ray and sz observations of three highly x - ray luminous , dynamically relaxed galaxy clusters : rx j1347.5 - 1145 ( @xmath8 ) , abell 1835 ( @xmath9 ) and abell 478 ( @xmath10 ) . in section 4 we use the combined data to measure the hubble constant . in section 5 we compare our results to those obtained using the standard isothermal @xmath4-model approach . a summary of the results can be found in sect . [ conclusions ] . unless stated otherwise , all quantities are are quoted for a hubble constant @xmath11 , matter density @xmath1 and vacuum energy density @xmath2 . error bars correspond to 1@xmath12 ( 68.3% ) confidence . within the context of the isothermal @xmath4-model @xcite , the surface brightness profile of a galaxy cluster can be written as @xmath13^{-3\,\beta+\frac{1}{2 } } , \label{betasurface}\ ] ] where @xmath14 is the angular core radius and @xmath4 is the slope parameter . by calculating the x - ray emission due to hot cluster gas at a temperature @xmath15 ( e.g. , * ? ? ? * ; * ? ? ? * ) , one can invert this to yield the central electron density @xmath16 of the corresponding intrinsic ( 3-dimensional ) electron density profile @xmath17^{-\frac{3\,\beta}{2 } } , \label{betamodel}\ ] ] where the core radius @xmath18 follows from the angular diameter distance @xmath19 of the cluster , @xmath20 . the comptonization parameter along a line of sight at an impact parameter @xmath21 is defined by @xmath22 ( e.g. , * ? ? ? * where we have converted the equation into an integral along radius @xmath23 ) . here @xmath24 is the thomson cross - section , @xmath25 the boltzmann constant , @xmath26 the speed of light , and @xmath27 the electron mass . eq . ( [ y - integral ] ) can be integrated analytically for the isothermal @xmath4-model ( e.g. , * ? ? ? * ) to give @xmath28^{-\frac{3}{2}\beta+\frac{1}{2 } } , \label{betacomp}\ ] ] with the normalization constant @xmath29 @xcite , where @xmath30 is the beta function . in previous work , the temperature @xmath15 has usually been taken to be the mean emission - weighted temperature of the x - ray gas , determined with broad - beam instruments . however , as discussed in section 1 , observations with the chandra and xmm - newton satellites have shown that regular , dynamically relaxed galaxy clusters are not isothermal . this temperature variation should be accounted for in the analysis . finally , clusters are not infinitely large , as is usually assumed when calculating the predicted comptonization parameter profile using the @xmath4-model . in what follows we show how modern x - ray data can be used to integrate eq . ( [ y - integral ] ) without the assumption of isothermality , taking account of the finite size of galaxy clusters . in particular , where the region of the cluster directly observed in x - rays is small , the use of an extrapolation procedure like the one described here can become important . in order to calculate the expected y - parameter profile ( eq . [ y - integral ] ) for a particular galaxy cluster , the temperature profile @xmath32 and the electron density profile @xmath33 need to be known . in practise , deprojection analyses of chandra or xmm - newton data ( e.g. , @xcite and references therein ) provide the electron density and x - ray gas temperature in @xmath34 discrete shells , with inner and outer radii @xmath35 and @xmath36 , respectively . we can thus recast eq . ( [ y - integral ] ) into a sum starting with shell j at impact parameter @xmath21 : @xmath37 for clusters with temperatures above about 8 kev , relativistic corrections also become significant . analytic and fitting relations for the relativistic corrections to the sz effect have been worked out by several groups : @xcite . using the analytical formulae of @xcite up to second order , we determine the relativistic correction factor @xmath38 that needs to be applied to each summand of ( [ y - sum ] ) , so that @xmath39 . in the case of the chandra x - ray observatory , the detector size and the particle background limit the regions of clusters for which direct temperature measurements can be made ( @xmath40 ) to a fraction of the virial radius . in order to calculate the predicted comptonization parameter precisely , we need to extrapolate the temperature and electron density profiles out to the edge of the cluster , @xmath41 - or at least past the point where significant contributions to @xmath3 are made . here , for convenience , we will set @xmath42 and @xmath43 ( corresponding to the radii within which the mean enclosed mass density is @xmath44 and @xmath45 times the critical density of the universe @xmath46 at redshift @xmath47 , respectively ) . @xmath48 is typical of the outer radii for which useful information on the gas temperature profile can be obtained from chandra observations . @xmath49 is used to mark the outer edge of the cluster . ( in sect . [ conclusions ] we shall show that the precise choice of @xmath41 does not affect the results significantly . ) we parametrize the total mass distributions in the clusters using an nfw model @xmath50 where @xmath51 is the mass density , @xmath52 is the scale radius , @xmath53 , @xmath54 is the hubble constant , g is the gravitational constant , @xmath55 and @xmath26 is the concentration parameter with @xmath56 . the temperature @xmath57 , electron density @xmath58 , and gas mass - weighted temperature @xmath59 within @xmath60 are assumed to be known from direct spatially - resolved spectroscopy . we extrapolate the electron density from @xmath60 to @xmath41 using a power - law model @xmath61 . the hydrostatic equation is then used to determine the temperature solution @xmath32 for that power - law electron density profile in the given nfw potential ( see appendix [ extrap ] ) . we fit the exponent @xmath62 of the electron density profile so that the model complies with two fixed quantities : 1 . the observed temperature @xmath57 at @xmath60 2 . the _ gas - mass _ weighted temperature @xmath63 at @xmath41 . the power - law model approximates the true electron density profile between @xmath60 and @xmath41 . ( it is assumed that the galaxy cluster ends at @xmath41 . ) to determine the gas - mass weighted temperature @xmath63 within @xmath41 , we use the @xcite solution for the radial dependence of the ratio of kinetic energy @xmath64 and potential energy @xmath65 in an nfw potential with an isotropic velocity dispersion . the key assumption here is that the gas - mass weighted temperature @xmath66 inside a radius @xmath23 is proportional to the kinetic energy of the mass distribution inside this radius ( e.g. , @xcite ) . we then calculate the change , @xmath67 , of the ratio @xmath68 between @xmath60 and @xmath41 as @xmath69}{t_{{\rm m},r_2}/\left[m_{\rm tot}(r_2)/r_2 \right ] } , \label{lokas}\ ] ] where @xmath70 is the total enclosed mass within radius @xmath23 . the quantity @xmath67 can be computed as a function of the scale radius @xmath52 and the concentration parameter @xmath26 of the nfw profile using the analytical formulae included in appendix [ lokmam ] . thus , given both the gas - mass weighted temperature @xmath71 at @xmath60 and a specific nfw mass model from the chandra data , the gas - mass weighted temperature @xmath66 at @xmath41 can be obtained from eq . ( [ lokas ] ) . for the correct electron density slope @xmath62 , this is equal to the temperature that follows from the temperature solution ( eq . [ tprof_analyt ] ) and gas density profile @xmath72 ( for which we assume @xmath73 ) via the integral @xmath74 where @xmath75 is the gas mass inside the radius @xmath23 . we note that in cases where the x - ray data only extend to radii @xmath76 ( within which cooling and/or heating effects may have modified @xmath77 significantly ) , one should extrapolate the gas density ( appendix [ extrap ] ) from @xmath78 to @xmath49 , use this extrapolation to estimate the gas temperature and density at @xmath42 , and then apply eq . ( [ lokas ] ) between @xmath48 and @xmath49 as usual . ( this is possible because the extrapolation recipe is attached continuously to the chandra data so that @xmath66 can be calculated for any exponent @xmath62 of the electron density extrapolation , regardless of where we attach the extrapolation . ) we show below that this approach leads to robust answers in the case of chandra observations of abell 478 , where the data cover only a relatively small radial range . .details of the chandra nfw mass models [ cols="<,<,^,^",options="header " , ] applying the procedure described in sect . [ hubble ] to the measured @xmath4-profile parameters yields hubble constant estimates of @xmath79 for rx j1347.5 - 1145 , @xmath80 for abell 1835 and @xmath81 for abell 478 . one can immediately see that the application of the isothermal @xmath4-model does not yield a consistent estimate when applied naively to chandra data for the present sample of clusters . it is also interesting to compare these estimates with previously published @xmath4-model estimates for the clusters discussed here . we have translated the electron densities from other studies into our assumed cosmology ( @xmath1 , @xmath2 and @xmath11 ) . 1 . p01 determined a hubble constant @xmath82 using a @xmath83 kev isothermal model with the @xmath4-model parameters @xmath84 arcsec , @xmath85 and @xmath86@xmath87 from rosat @xcite . using the same temperature and a slightly different @xmath4-model , r02 found an angular diameter distance that implies @xmath88 . this estimate is consistent with the value we obtain using our fully deprojected and extrapolated cluster model ( tab . [ results ] ) . however , the gas temperature and the other @xmath4-model parameters used by r02 are not consistent with the @xmath4-model parameters from chandra ( tab . [ beta - fits ] ) . 2 . for this cluster , m00 measured a hubble constant @xmath89 for a @xmath90 kev isothermal model with the @xmath4-model parameters @xmath91 arcsec , @xmath92 and @xmath93@xmath87 from a fit to rosat data . this temperature was obtained by allowing for a cooling flow component in the core of the cluster . r02 obtained an angular diameter distance to the cluster which implies @xmath94 , based on a @xmath95 kev isothermal model ( no cooling flow correction ) . the difference between these two measurements can be attributed to the different isothermal temperatures and small differences in the assumed @xmath4-model parameters . note also that the @xmath4-model parameters used in these studies are significantly different from those determined from chandra ( tab . [ beta - fits ] ) over a smaller range of radii . this explains the difference in the predicted hubble constant . however , the m00 result is consistent with the result we obtain from the deprojected and extrapolated x - ray data , which shows that it is possible to partially correct the @xmath4-model in the central region of the cluster when the surface brightness profile observed with the rosat field of view is used . for larger radii this approximation will fail , however , because of the fixed @xmath4-model slope . mmr measured a hubble constant @xmath96 assuming an isothermal @xmath97 kev model with the @xmath4-model parameters @xmath98 arcmin , @xmath99 and @xmath100@xmath87 from rosat . this temperature also accounts for the presence of a cooling flow in the cluster core . @xcite improved upon the result by mmr by determining the comptonization parameter as a function of radius from a deprojection of the chandra data , and by extrapolating the chandra data with a @xmath4-model based on a fit to the combined surface brightness profile from chandra and rosat . they used @xmath101 and obtained a hubble constant estimate @xmath102 , which lies between our value and the one obtained by mmr . we have presented a new method for extrapolating the pressure profiles of galaxy clusters beyond the region accessible to current x - ray satellites such as chandra and xmm - newton , for use in combined x - ray and sz studies . our method assumes hydrostatic equilibrium , an nfw mass profile for the clusters and predicts the temperature profile and an effective power - law slope of the electron density profile . the method does not assume isothermality and provides a simple extrapolation recipe that can be easily implemented . we have applied our method to the chandra data for three x - ray bright , dynamically relaxed galaxy clusters , rx j1347.5 - 1145 , abell 1835 and abell 478 , and have obtained detailed temperature , electron density and comptonization parameter profiles for these clusters out to the virial radius . since all three clusters have published sz detections , we were also able to measure the hubble constant for each , as shown in table [ results ] . the individual measurements are consistent with each other and yield a weighted mean of @xmath0 . this result is consistent with findings from the hubble key project ( @xmath103 ; @xcite ) and combined studies of the cosmic microwave background and x - ray galaxy cluster data ( @xmath104 ; @xcite ) . note that our weighted mean value does not include additional systematical errors , due to , e.g. , clumping or asphericity , although for such regular , relaxed clusters these should be at a minimum . the contribution to the predicted comptonization parameter from material beyond the regions directly observed with chandra , modelled by our extrapolation procedure , varies from system to system . for rx j1347.5 - 1145 the correction to the comptonization parameter in the region used in table [ results ] amounts only to @xmath105 per cent and is negligible . for abell 1835 , however , the contribution from material in the extrapolated region amounts to @xmath106 per cent , and for abell 478 the contribution rises to @xmath107 per cent . this shows that a robust extrapolation method is necessary when combining x - ray and sz observations to infer cosmological information . we have estimated the effect of changing the outer radius @xmath49 by repeating the calculation for abell 478 , but stopping at @xmath108 . this leads to a 4.2 per cent smaller average comptonization parameter , well within the observational error bar . we have compared the comptonization parameter profiles predicted by our method with the profiles obtained by fitting an isothermal @xmath4-model to the chandra data . we find that a naive application of the @xmath4-model approach leads to profiles with significantly different central comptonization parameters and predicts different slopes for the comptonization parameter profiles at large radii . the isothermal @xmath4-model also leads to inconsistent hubble constant estimates when applied to the chandra data sets considered here . it is clear that for galaxy clusters like abell 478 , a field of view larger than the one afforded by the chandra acis - s3 detector would be beneficial . in this respect the combination of chandra and xmm - newton observations will help ( pointecouteau 2004 ) , although the detector background will still prohibit precise measurements of the temperature profile beyond @xmath109 . ( we also note that the high spatial resolution of chandra is important in resolving the central temperature profile and constraining the best - fit nfw mass models . ) observations for a larger sample of clusters , as well as deeper x - ray and sz observations , should make this a powerful method to explore the cosmological distance scale . in the first case , it will be important to concentrate such studies on the largest , dynamically relaxed clusters for which systematic uncertainties associated with the deprojection method and assumption of hydrostatic equilibrium are at a minimum . we thank y. suto , n. itoh and the anonymous referee for comments . we thank the developers of the gnu octave numerical computation language and the gnu scientific library for these tools . swa and acf acknowledge the support of the royal society . 99 allen s.w . , 2000 , mnras , 315 , 269 allen s.w . , ettori s. , fabian a.c . , 2001 , mnras , 324 , 877 allen s.w . , schmidt r.w . , fabian a.c , 2001 , mnras , 328 , l37 allen s.w . , schmidt r.w . , fabian a.c . , 2002 , mnras , 335 , 256 allen s.w . , schmidt r.w , bridle s.l . , 2003 , mnras , in press birkinshaw m. , 1999 , phys . , 310 , 97 carlstrom j.e . , joy m. , grego l. , holder g. , holzapfel w.l . , laroque s. , mohr j.j . , reese e.d . , to appear in : `` constructing the universe with clusters of galaxies '' , iap conference , july 2000 , eds . f. durret and g. gerbal , preprint : astro - ph/0103480 cavaliere a. , fusco - femiano r. , 1978 , a&a , 70 , 677 cavaliere a. , danese l. , de zotti g. , 1979 , a&a , 75 , 322 challinor a. , lasenby a. , 1998 , apj , 499 , 1 eke v.r . , navarro j.f . , frenk c.s . , 1998 , apj , 503 , 569 freedman w.l . , et al . , 2001 , apj , 553 , 47 gradshteyn i.s . , ryzhik i.m . , 2000 , `` table of integrals , series and products '' , sixth edition , academic press , san diego inagaki y. , suginohara t. , suto , y. , 1995 , pasj , 47 , 411 kaastra j.s . , et al . , 2004 , a&a , 413 , 415 itoh n. , kohyama y. , nozawa s. , 1998 , apj , 502 , 7 jones m.e . , et al . , 2001 , submitted to mnras , preprint : astro - ph/0103046 kaastra j.s . , mewe r. , 1993 , legacy , 3 , 16 kitayama t. , et al . , 2004 , pasj , 56 , 17 komatsu e. , et al . , 2001 , pasj , 53 , 57 liedahl d.a . , osterheld a.l . , goldstein w.h . , 1995 , apj , 438 , l115 lokas e.l . , mamon g.a . , 2001 , mnras , 321 , 155 mason b.s . , myers s.t . , readhead a.c.s , 2001 , apjl , 555 , 11 ( mmr ) mauskopf p.d . , et al . , 2000 , apj , 538 , 505 ( m00 ) myers s.t . , baker j.e . , readhead a.c.s . , leitch e.m . , herbig t. , 1997 , apj , 485 , 1 navarro j.f . , frenk c.s . , white s.d.m . , 1995 , mnras , 275 , 720 ( nfw ) pointecouteau e. , giard m. , benoit a. , dsert f.x . , bernard j.p . , coron n. , lamarre j.m . , 2001 , apj , 552 , 42 ( p01 ) pointecouteau e. , arnaud m. , kaastra j. , de plaa j. , 2004 , a&a , submitted ( p04 ) reese e.d . , carlstrom j.e . , joy m. , mohr j.j . , grego l. , holzapfel w.l . , 2002 , apj 581 , 53 ( r02 ) sarazin c.l . , 1988 , `` x - ray emission from clusters of galaxies '' , cambridge university press , cambridge sazonov s.y . , sunyaev r.a . , 1998 , apj , 508 , 1 schindler s. , hattori m. , neumann d.m . , boehringer h. , 1997 , a&a , 317 , 646 schmidt r.w , allen s.w . , fabian a.c . , 2001 , mnras , 327 , 1057 silk j. , white s.d.m . , 1978 , apjl , 226 , 103 sun m. , jones c. , murray s.s . , allen s.w . , fabian a.c . , edge a.c . , 2003 , apj , 587 , 619 sunyaev r. , zeldovich y. , 1972 , comments astrophys . space phys . , 4 , 173 yoshikawa k. , itoh m. , suto y. , 1998 , pasj , 50 , 203 let the mass distribution of a galaxy cluster be parametrized by a nfw mass model ( eq . [ nfw ] ) . also , let the electron density distribution for a certain range of radii be parametrized by a power - law model @xmath110 the temperature as a function of radius can then be determined under the assumption of hydrostatic equilibrium by solving the hydrostatic equation @xmath111 where @xmath112 is the boltzmann constant , @xmath113 is the molecular weight of the gas and @xmath114 is the proton mass ( note that the normalization of ( [ nepow ] ) drops out of this equation ) . using the method of the variation of the constant one finds that is possible to write the solution of ( [ hydro ] ) in the form @xmath115 with the boundary condition @xmath116 . @xmath117 is the hypergeometric function ( e.g. , * ? ? ? * ) which is well - defined on the whole negative real axis and is available in standard mathematical packages . in this appendix we list the @xcite expressions for the potential energy @xmath65 and the kinetic energy @xmath64 in an nfw potential with an isotropic velocity dispersion , as a function of radius @xmath23 . the ratio of these quantities is used in sect . [ extrap_proc ] to determine the cluster model extrapolation . it can be calculated from the concentration parameter @xmath26 and the scale radius @xmath52 . the radial dependence of the potential energy associated with an nfw mass distribution with virial radius @xmath118 as a function of ( scaled ) radius @xmath119 is @xmath120\end{aligned}\ ] ] ( eq . 21 in @xcite ) , where @xmath121 is the asymptotic value @xmath122 ^ 2 } .\ ] ] the radial dependence of the kinetic energy , assuming isotropic velocity dispersion , is @xmath123 - c^2 s^2 [ 7 + 6 \ln ( 1 + c s ) ] \nonumber \\ & & + c^3 s^3 [ \pi^2 - \ln c - \ln s + \ln ( 1+c s ) \nonumber \\ & & + 3 \ln^2 ( 1+c s ) + 6 { \rm li}_2 ( -c s ) ] \}\end{aligned}\ ] ] ( eq . 24 in @xcite ) , where li@xmath124 is the dilogarithm .	we present an improved method for predicting the sunyaev - zeldovich ( sz ) effect in galaxy clusters from spatially - resolved , spectroscopic x - ray data . using the deprojected electron density and temperature profiles measured within a fraction of the virial radius , and assuming a @xcite mass model , we show how the pressure profile of the x - ray gas can be extrapolated to large radii , allowing the comptonization parameter profile for the cluster to be predicted precisely . we apply our method to chandra observations of three x - ray luminous , dynamically relaxed clusters with published sz data : rx j1347.5 - 1145 , abell 1835 and abell 478 . combining the predicted and observed sz signals , we determine improved estimates for the hubble constant from each cluster and obtain a weighted mean of @xmath0 for a cosmology with @xmath1 and @xmath2 . this result is in good agreement with independent findings from the hubble key project and the combination of cosmic microwave background and galaxy cluster data . [ firstpage ] x - rays : galaxies : clusters cosmic microwave background cosmology : observations distance scale galaxies : clusters : individual ( rx j1347.5 - 1145 , abell 1835 , abell 478 )
in the present paper , we study the following third - order integrable nonlinear evolution equation : @xmath0 this equation was introduced recently by qiao @xcite as the second positive member in a new completely integrable hierarchy . the qiao equation we will refer to in this way possesses a lax representation and a bi - hamiltonian structure @xcite . moreover , the qiao equation may be reduced from the two - dimensional euler equation by an approximation procedure , and its solutions may be useful to construct new solvable potentials in newtonian dynamics and to model electrophysiological phenomena in neuroscience @xcite . there is an interesting problem concerning the soliton solution of the qiao equation . in @xcite , where the traveling wave solutions of were studied , it was stated that this equation has no smooth solitons , and only piecewise smooth , but not smooth , continuous traveling wave solutions were found in those papers . on the other hand , in @xcite , where the bifurcations of traveling wave solutions of were studied , it was stated that this equation does possess smooth soliton solutions , and some of those solitons were shown in figure 5 of that paper . we believe that the complexity of investigation of the qiao equation s solitons is caused mainly by the fact that they can be expressed only in a parametric form . in the present paper , we also address the problem on smooth soliton solutions of , but we do not use the traveling wave reduction to an ordinary differential equation . instead , we follow the way used in @xcite for the short pulse equation : we transform the studied new equation into a known old one , and then transform known explicit expressions for solutions of the old equation into sought parametric expressions for solutions of the new one . the paper is organized as follows . in section [ s2 ] , we find a transformation which relates the qiao equation with the well - known modified korteweg de vries ( mkdv ) equation . then , in sections [ s3 ] and [ s4 ] , we apply this transformation , respectively , to the rational solution and to the soliton solution of the mkdv equation , and in this way we obtain three types of smooth solitons of the qiao equation as well as three types of its smooth antisolitons . section [ s5 ] contains some concluding remarks . let us try to transform the new qiao equation into one of well - studied old equations . we do this in the way used in @xcite for the fujimoto watanabe equations ; see also @xcite for more theory and applications of the transformations involved . first of all , we transform the nonlinear evolution equation , which has the form @xmath1 , into an equation of the form @xmath2 . to do this , we change in the dependent variable @xmath3 as @xmath4 the result being the polynomial evolution equation @xmath5 this equation is just a special case of the item ( ii.2 ) in the fujimoto watanabe classification @xcite , @xmath6 with the arbitrary constant @xmath7 fixed to be @xmath8 . therefore , according to @xcite , it is possible to relate the qiao equation , through its previously known form , with the korteweg de vries equation . we will see soon , however , that it is more convenient to transform the qiao equation into the mkdv equation , in order to easily derive solutions of the former from solutions of the latter . next , we take into account that the evolution equation can be written as @xmath9 which means , according to @xcite , that the transformation @xmath10 relates with the third - order evolution equation @xmath11 this point should be explained in more detail . it is easy to see that any function @xmath12 satisfying determines via a function @xmath13 satisfying . in the opposite direction , when we substitute the expressions to , we obtain a fourth - order equation for @xmath12 , which , owing to the form of , can be written as @xmath14 and then can be integrated over @xmath15 , the result being the third - order equation @xmath16 where @xmath17 is an arbitrary function . this means that any solution @xmath13 of is represented parametrically , via with @xmath15 being the parameter , by a solution @xmath12 of with some function @xmath17 . however , we can always fix this function @xmath17 to be zero , by the corresponding change of the parameter @xmath15 , @xmath18 , which has no effect on the represented solution @xmath13 . consequently , all solutions of are represented by all solutions of via the relations . finally , we introduce the new dependent variable @xmath19 , @xmath20 which is convenient because the relation @xmath21 holds due to . according to and , the function @xmath22 is any solution of the mkdv equation @xmath23 whereas the function @xmath12 is determined by @xmath22 via the relations @xmath24 and @xmath25 up to an arbitrary additive constant of integration . of course , there is no need to further transform into the korteweg de vries equation , because it would only complicate the already found way to solve the qiao equation . we can summarize the obtained transformation , which relates the equations and to each other , as follows . all solutions of the qiao equation are represented parametrically by all solutions of the mkdv equation via the relations @xmath26 where the independent variable @xmath15 serves as the parameter . of course , this transformation is local , in the sense that it maps a local solution of to a local solution of , and for every particular @xmath22 some extra analysis is required to see whether the corresponding @xmath27 is real - valued , smooth , and global . in the next two sections , in order to show how the transformation works , we apply it to two simplest nontrivial solutions of the mkdv equation : the rational solution @xmath28 where @xmath29 is an arbitrary nonzero constant , and the soliton solution @xmath30 where @xmath29 and @xmath31 are arbitrary constants , @xmath32 and @xmath33 . one can notice that the expressions and are somewhat different from the ones well known in the literature @xcite . the reason is that the mkdv equation is usually studied in its focusing form , with a positive coefficient at the nonlinear term , whereas we have the defocusing case in , and the expressions and can be obtained from the well - known ones through a complex - valued scale transformation of the dependent variable and a change of notations for the constants . let us derive a solution of the qiao equation from the rational solution of the mkdv equation by means of the transformation . we substitute the function @xmath22 given by to the relations and , make the integration required in , and fix the arbitrary additive constant of integration so that to have @xmath34 for @xmath35 . in this way , we obtain the following expressions : @xmath36 where @xmath37 stands for the hyperbolic arctangent . these expressions contain @xmath15 via the variable @xmath38 only , therefore we can use @xmath39 instead of @xmath15 as the parameter , for simplicity . the function @xmath27 , determined parametrically by with @xmath39 being the parameter , does satisfy , as one can prove by direct computations . until now , we worked with the solution in a purely algebraic way , and the quantities @xmath29 , @xmath15 , @xmath40 and @xmath19 could be even complex - valued there . however , in order that the obtained solution @xmath27 be a real - valued function of real - valued variables , the quantities @xmath29 , @xmath15 and @xmath40 must be real - valued , and the condition @xmath41 must be satisfied , which is caused by the presence of @xmath37 in . the domain of variation of @xmath42 is @xmath43 due to and . the function @xmath27 , due to and , is everywhere smooth ; it is positive , with one maximum equal to @xmath44 , when @xmath45 ; it is negative , with one minimum equal to @xmath44 , when @xmath46 ; and its asymptotes at @xmath47 are zero . we can see easily that the obtained solution is a traveling wave solution , @xmath48 with the function @xmath49 being determined parametrically . the shape of this solitary wave is close to the shape of a hyperbolic secant curve but not identical to it , as shown in figure [ f1 ] . the speed of this solitary wave is equal to @xmath50 ; solitons ( positive waves with @xmath45 ) move to the right , and antisolitons ( negative waves with @xmath46 ) move to the left , as shown in figures [ f2 ] and [ f3 ] , respectively . this antisymmetry @xmath51 between positive and negative solutions of the qiao equation corresponds , via the transformation , to the symmetry @xmath52 of the mkdv equation . and @xmath53 ( solid ) , compared to the curve @xmath54 ( dashed).,scaledwidth=72.0% ] : @xmath53 ( solid ) and @xmath55 ( dashed).,scaledwidth=72.0% ] : @xmath53 ( solid ) and @xmath55 ( dashed).,scaledwidth=72.0% ] to conclude the section , we point out the following two discovered phenomena . first , a rational solution of one nonlinear equation is transformed into a smooth soliton solution of another nonlinear equation . second , only a part of the rational solution of the mkdv equation the part bounded by the condition is used to produce the entire smooth real solution of the qiao equation . now , let us proceed to transformation of the soliton solution of the mkdv equation . the cases of `` @xmath56 '' and `` @xmath57 '' signs in will be considered separately . we substitute the function @xmath22 , with the choice of `` @xmath56 '' sign in it , to the relations and , make the integration required in , fix the constant of integration so that to have @xmath34 for @xmath35 , and obtain in this way the following expressions : @xmath58 these expressions contain @xmath15 via the variable @xmath59 only . therefore we can use @xmath60 instead of @xmath15 as the parameter , for simplicity , when we consider as a parametrically determined solution @xmath27 of the qiao equation . this is a traveling wave solution , @xmath61 with the function @xmath62 being determined parametrically , and the speed of the wave is equal to @xmath63 . of course , in order that the obtained solution be a real solution , the quantities @xmath29 , @xmath31 , @xmath60 and @xmath40 must be real - valued in the right - hand sides of and ( we take @xmath64 without loss of generality ) , and the condition @xmath65 must be satisfied due to the presence of the square root . further constraints on @xmath60 and @xmath31 , caused by the presence of @xmath37 in and the denominator in , depend on the sign of @xmath29 . when @xmath45 , the parameter @xmath60 must lie in the interval latexmath:[\[\label{e21 } in order that @xmath42 be real - valued . with this constraint on @xmath60 , the domain of variation of @xmath42 is @xmath43 , and the function @xmath27 is everywhere smooth and positive , with one maximum equal to @xmath67 and zero asymptotes at @xmath47 . the soliton , determined by , moves with the speed @xmath63 to the right , as shown in figure [ f4 ] . when @xmath68 , the height and the speed of this soliton tend to @xmath44 and @xmath50 , respectively . in fact , the soliton with @xmath45 , derived from the rational solution of the mkdv equation in section [ s3 ] , is a limiting case of the soliton with @xmath68 , as one can prove easily . on the other hand , when @xmath69 , the height and the speed of the soliton tend to @xmath70 and @xmath71 , respectively . figure [ f5 ] shows how the shape of the obtained soliton depends on the ratio @xmath72 . we see that the soliton s top widens and flattens , when @xmath31 is very close to @xmath29 , that is when the soliton s motion is extremely slow . however , the soliton s shape changes very little , when @xmath31 runs over the wide interval @xmath73 . this is shown in figure [ f6 ] , where we have adjusted the values of @xmath29 so that to have solitary waves of equal heights . it is interesting that , of the two solitary waves shown there , the one which is only a little bit wider moves more than two times slower than the other one . and @xmath74 : @xmath53 ( solid ) and @xmath75 ( dashed).,scaledwidth=72.0% ] and @xmath53 : @xmath76 ( solid ) , @xmath77 ( dot - dashed ) , @xmath78 ( dashed ) , and @xmath79 ( dotted).,scaledwidth=72.0% ] : @xmath80 and @xmath81 ( solid ) ; @xmath82 and @xmath83 ( dashed).,scaledwidth=72.0% ] when @xmath46 in and , the variable @xmath42 is real - valued on the whole interval @xmath84 for any @xmath31 , @xmath85 , and the domain of variation of @xmath42 is @xmath43 , but the function @xmath27 has one or two singularities if @xmath86 or @xmath87 , respectively . the expressions and with @xmath88 do determine a smooth soliton . this solitary wave is positive , has one maximum equal to @xmath67 and nonzero asymptotes @xmath89 at @xmath47 , and moves with the speed @xmath63 , that is to the left , as shown in figures [ f7 ] and [ f8 ] . when @xmath68 , the height of the soliton tends to @xmath89 , which is the level of the asymptotes , and the speed of the soliton tends to @xmath50 . when @xmath90 , the soliton s height and speed tend to @xmath91 and @xmath92 , respectively . and @xmath53 : @xmath93 ( solid ) , @xmath94 ( dashed ) , and @xmath95 ( dotted).,scaledwidth=72.0% ] and @xmath96 : @xmath53 ( solid ) and @xmath97 ( dashed).,scaledwidth=72.0% ] the case of `` @xmath57 '' sign in the mkdv equation s solution is processed in the same way as the case of `` @xmath56 '' sign . as the result , we obtain the following expressions : @xmath98 in order to get real , smooth and global solutions of the qiao equation , we must supplement these expressions either with @xmath99 for positive values of @xmath29 , or with latexmath:[\[\label{e26 } for negative values of @xmath29 . the obtained solitary waves are negative . the antisoliton moves to the right and has nonzero asymptotes , as shown in figure [ f9 ] . it corresponds to the soliton , determined by , and , in the sense of the antisymmetry between positive and negative solutions , discussed in section [ s3 ] . the antisoliton , determined by , and , moves to the left and has zero asymptotes , as shown in figure [ f10 ] . it corresponds to the soliton in the sense of that antisymmetry . owing to this correspondence , there is no need to discuss these antisolitons in more detail . and @xmath96 : @xmath53 ( solid ) and @xmath97 ( dashed).,scaledwidth=72.0% ] and @xmath74 : @xmath53 ( solid ) and @xmath75 ( dashed).,scaledwidth=72.0% ] to conclude the section , we remind the following important observation . when we used the soliton solution of the mkdv equation to obtain smooth real solutions of the qiao equation , we had to impose certain constraints on the quantities @xmath15 , @xmath40 , @xmath29 and @xmath31 in the original solution , otherwise the obtained solutions would be singular or complex . in this paper , we found a transformation which relates a new integrable nonlinear evolution equation of qiao with the well - known mkdv equation . by means of that transformation , we derived smooth soliton solutions of the qiao equation from the known rational and soliton solutions of the mkdv equation . we obtained three types of positive solitary wave solutions , or solitons . solitons of two types have zero asymptotes and move to the right , whereas solitons of the third type have nonzero asymptotes and move to the left . also we obtained three types of negative solitary wave solutions , or antisolitons , which have the same shape as the corresponding solitons but differ from them in their sign and direction of motion . it is remarkable that the rational solution of the mkdv equation is transformed into a smooth soliton solution of the qiao equation . an important detail of this is that only a part of the rational solution of the mkdv equation produces the entire smooth soliton solution of the qiao equation . when the soliton solution of the mkdv equation is transformed , it is also necessary to impose certain constraints on its variables and constants , in order that the obtained solution of the qiao equation be real and smooth . these observations may be important for further development of the method to obtain parametric representations for solutions of nonlinear wave equations via transformations of the equations . last but not least , we believe that the discovered transformation will be useful to obtain and study multisoliton solutions of the qiao equation with the help of the hirota bilinear representation of the mkdv equation @xcite . this work was partially supported by the brffr grant @xmath10110 - 117 . the author also thanks the max planck institute for mathematics for hospitality and support .	we find a transformation which relates a new third - order integrable nonlinear evolution equation , introduced recently by qiao , with the well - known modified korteweg de vries equation . then we use this transformation to derive smooth soliton solutions of the new equation from the known rational and soliton solutions of the old one .
the structure of bound states in linear systems follows the symmetry of the underlying potential , a commonly known example being wave functions of eigenstates in symmetric double - well potentials @xcite . the addition of the self - attractive nonlinearity leads to a qualitative change of the situation , causing the transition from the symmetric ground state to an asymmetric one , if the strength of the nonlinearity exceeds a critical value @xcite . this transition was studied in detail for bose - einstein condensates ( becs ) loaded into double - well potentials @xcite . experimentally , the transition was realized in bec @xcite and in nonlinear optics , where a double - well structure was induced in a photorefractive material @xcite . the limit case of the double - well setting with a tall potential barrier between the wells corresponds to the _ dual - core _ system , such as optical fibers @xcite and bragg gratings @xcite with the twin - core structure , and pairs of linearly coupled planar waveguides with the @xmath0 ( second - harmonic - generating ) intrinsic nonlinearity mak - chi2 . the _ symmetry - breaking bifurcation _ ( sbb ) , which destabilizes the symmetric ground state in nonlinear dual - core systems and gives rise to an asymmetric one , was first discovered in the discrete self - trapping model @xcite . in nonlinear optics , the sbb was predicted for continuous - wave ( spatially uniform ) states @xcite , and for solitons @xcite , in the models of dual - core fibers and bragg gratings . the sbb was also analyzed for matter - wave solitons in the bec loaded into double - channel potential traps @xcite-@xcite . the self - focusing cubic nonlinearity ( the kerr term in optics ) gives rise to the symmetry - breaking phase transition of solitons of the first kind ( alias the subcritical sbb ) in the twin - core system . in that case , the branches of asymmetric modes emerge as unstable ones , going backward and then stabilizing themselves at turning points , from which they continue the evolution in the forward direction @xcite . the character of the phase transition alters to the second kind ( i.e. , the sbb type changes from sub- to supercritical ) under the combined action of the self - focusing nonlinearity and a periodic potential acting along the free axis ( transverse to the double - well potential ) @xcite . the supercritical sbb gives rise to stable asymmetric branches which immediately go in the forward direction @xcite . the sbbs in the twin - core bragg grating and @xmath6 waveguides are of the forward type too @xcite . a noteworthy counterpart of the linear double - well potential is a _ pseudopotential _ @xcite induced by a spatial modulation of the nonlinearity coefficient in the form of a symmetric pair of sharp peaks . this configuration may be implemented in optics and bec alike barcelona , pu . the ultimate form of the double - peak pseudopotential features the nonlinearity concentrated at two points , in the form of a symmetric pair of delta - functions ( which may be approximated by narrow gaussians ) @xcite . the sbb of solitons in the dual - core pseudopotentials belongs to the subcritical type @xcite . the sbb of localized modes was also recently investigated in the model of a linear waveguide with two narrow @xmath7 stripes embedded into it @xcite ( the solution for the mode pinned to a single stripe was found in ref . [ 23 ] ) . for discrete solitons in dual - core lattices , the sbb was analyzed too , assuming the uniform transverse coupling between parallel chains herring , or the coupling applied at a single pair of sites @xcite . the bifurcation is subcritical in the former case , and supercritical in the latter one . the simplest realizations of the sbb in discrete media are provided by linear lattices with a symmetric pair of embedded nonlinear sites ( this setting was introduced in refs . @xcite ) , or a symmetric pair of nonlinear elements side - coupled to the linear chain almas . in particular , the sbb for localized modes in the linear lattice with a pair of sites carrying the cubic nonlinearity was recently studied in ref . @xcite , where it was concluded that the bifurcation is , chiefly , of the subcritical type ( except for the case when the nonlinear sites are separated by a single lattice spacing , see below ) . the objective of the present work is to investigate localized symmetric , asymmetric , and antisymmetric solitary modes in the bichromatic one - dimensional linear lattice with a pair of inserted @xmath0-nonlinear sites . the lattice is described by the following equations:@xmath8where @xmath9 and @xmath10 are complex amplitudes of the fundamental - frequency ( ff ) and second - harmonic ( sh ) fields at the @xmath11-th site , the constant of the linear coupling between adjacent sites is scaled to be @xmath12 , as well as the strength of the @xmath0 nonlinearity inserted at sites @xmath13 and @xmath14 , which are represented by the kronecker s symbols @xmath15 and @xmath16 ( i.e. , integer @xmath1 is the distance between the nonlinear sites ) , and @xmath3 is the real mismatch parameter . the system can be implemented as an array of parallel optical waveguides , with @xmath17 being the propagation distance along the waveguides , as previously done for a great variety of models supported by such quasi - discrete settings in nonlinear optics @xcite . arrayed @xmath6 structures may be also built by means of the quasi - phase - matched technique @xcite . in these contexts , discrete @xmath0 systems were introduced in ref . @xcite . to realize the present model , based on eqs . ( [ u1 ] ) , ( [ v1 ] ) , two selected cores in the waveguiding array can be made quadratically nonlinear by fabricating them of an appropriate material , or applying the @xmath0-inducing poling to this pair @xcite . in fact , the model with two kerr - nonlinear sites , studied in ref . @xcite , corresponds to the _ cascading limit _ @xcite of eqs . ( [ u1 ] ) , ( [ v1 ] ) for @xmath18 . in this work , we demonstrate that the system with negative , zero , and positive values of the mismatch ( @xmath3 ) opens new possibilities for the creation and manipulations of discrete solitary modes , in comparison with the cascading limit . in particular , we demonstrate that the character of the sbb can be switched from super- to subcritical by varying the mismatch , which also alters the stability of the modes and their existence thresholds . the paper is organized as follows . analytical solutions for symmetric , asymmetric , and antisymmetric localized modes generated by eqs . ( [ u1 ] ) , ( [ v1 ] ) in the infinite lattice are produced in section 2 . in the general case , the solution is implicit , while explicit solutions are obtained for the smallest separation between the nonlinear sites , @xmath2 . numerical results , obtained for finite lattices , are reported in section 3 . the stability of the discrete modes is considered in that section too . the paper is concluded by section 4 . analytical solutions to eqs . ( [ u1 ] ) , ( [ v1 ] ) for stationary modes with propagation constant @xmath19 are sought for in the usual form,@xmath20which reduces eqs . ( [ u1 ] ) and ( [ v1 ] ) into the stationary equations for real discrete fields @xmath21:@xmath22at @xmath23 , an exact solution to eqs . ( [ u ] ) and ( [ v ] ) is obvious:@xmath24where @xmath25 and @xmath26 are arbitrary amplitudes , and @xmath27 are determined by the following relations:@xmath28due to condition @xmath27 , the limit value of @xmath19 which corresponds to @xmath29 or @xmath30 can be found from eq . ( [ kappa ] ) : @xmath31the localized modes existing at @xmath32 . similarly , at @xmath33 , the exact solution to eqs . ( [ u ] ) and ( [ v ] ) is@xmath34and in the inner region , @xmath35 , the solution is constructed as@xmath36 the conditions of the continuity of the discrete fields at points @xmath13 and @xmath37 lead to linear relations between the amplitudes of the solutions in the outer and inner regions:@xmath38finally , the nonlinear equations at sites @xmath13 and @xmath14 take the following form:@xmath39 a_{1 } \notag \\ + \frac{1}{2}\left [ b_{11}\exp \left ( -\kappa _ { 1}\right ) + b_{12}\exp \left ( -\left ( l-1\right ) \kappa _ { 1}\right ) \right ] + a_{1}a_{2}=0 , \notag \\ -\left [ 4k+q+1-\frac{1}{2}\exp \left ( -\kappa _ { 2}\right ) \right ] a_{2 } \notag \\ + \frac{1}{2}\left [ b_{21}\exp \left ( -\kappa _ { 2}\right ) + b_{22}\exp \left ( -\left ( l-1\right ) \kappa _ { 2}\right ) \right ] + \frac{1}{2}a_{1}^{2}=0 , \label{m=0}\end{gathered}\]]@xmath39 c_{1 } \notag \\ + \frac{1}{2}\left [ b_{11}\exp \left ( -\kappa _ { 1}\left ( l-1\right ) \right ) + b_{12}\exp \left ( -\kappa _ { 1}\right ) \right ] + c_{1}c_{2}=0 , \notag \\ -\left [ 4k+q+1-\frac{1}{2}\exp \left ( -\kappa _ { 2}\right ) \right ] c_{2 } \notag \\ + \frac{1}{2}\left [ b_{21}\exp \left ( -\kappa _ { 2}\left ( l-1\right ) \right ) + b_{22}\exp \left ( -\kappa _ { 2}\right ) \right ] + \frac{1}{2}c_{1}^{2}=0 . \label{1}\end{gathered}\]]amplitudes @xmath40 and @xmath41 can be eliminated in favor of @xmath42 and @xmath43 , making use of linear equations ( [ lin ] ) , which leads to the system of four equations , ( [ m=0 ] ) and ( [ 1 ] ) , for four remaining unknown amplitudes , @xmath44 and @xmath43 . families of solutions for localized modes are characterized below by dependences of their powers ( norms ) , @xmath45on the propagation constant , @xmath19 , a dynamical invariant of eqs . ( [ u1 ] ) , ( [ v1 ] ) being the total power , @xmath46 the solutions for the symmetric and antisymmetric modes are defined , respectively , by constraints @xmath47 , @xmath48 , or @xmath49 , @xmath50 with @xmath51 . in the latter case , only the ff field is antisymmetric , while its sh counterpart remains symmetric ; note also that the antisymmetric modes with even and odd @xmath1 may be categorized , respectively , as ones of the on - site and inter - site types @xcite . explicit solutions for the symmetric and antisymmetric modes can be obtained , in a simple approximate form , for @xmath52 , when eq . ( [ kappa ] ) yields @xmath53hence the modes are broad in this limit , according to eq . ( [ a12 ] ) . further , the amplitudes and powers ( [ n ] ) of the symmetric and antisymmetric modes , found in the same limit , are@xmath54a corollary of this result is that , except for the case of @xmath5 , when @xmath55 the total power of the modes never vanishes at @xmath56 , hence there is a finite power threshold ( a minimum value of the total power ) necessary for the existence of the localized states at @xmath57 . in particular , for the symmetric modes with @xmath58 , the threshold is @xmath59 , which , indeed , vanishes solely at @xmath5 . on the other hand , in lattices of a finite size , the threshold power of the symmetric modes vanishes in the limit of @xmath60 [ i.e. , when @xmath61 remains finite , provided that @xmath3 is positive , see eq . ( [ kappakappa ] ) ] . indeed , in this limit case the top line in eq . ( [ sas ] ) implies the vanishing of both @xmath25 and @xmath26 , while the nonvanishing of the threshold power in the respective infinite lattice is accounted for by the simultaneous divergence of the spatial width of the ff component , @xmath62 . obviously , the latter factor can not compensate the vanishing of the amplitude in finite - size lattices . the above analysis yields results in the implicit form , given by eqs . ( lin)-([1 ] ) . explicit solutions can be obtained for the smallest distance between the nonlinear sites , @xmath2 . in this case , one does not need to introduce solutions ( [ inner ] ) for the inner layer , while the remaining equations for amplitudes @xmath44 and @xmath43 take the following form:@xmath63 a_{1}+\frac{1% } { 2}c_{1}+a_{1}a_{2 } & = & 0 , \label{11 - 1 } \\ -\left [ 4k+q+1-\frac{1}{2}\exp \left ( -\kappa _ { 2}\right ) \right ] a_{2}+% \frac{1}{2}c_{2}+\frac{1}{2}a_{1}^{2 } & = & 0 , \label{12 - 1}\end{aligned}\]]@xmath63 c_{1}+\frac{1% } { 2}a_{1}+c_{1}c_{2 } & = & 0 , \label{21 - 1 } \\ -\left [ 4k+q+1-\frac{1}{2}\exp \left ( -\kappa _ { 2}\right ) \right ] c_{2}+% \frac{1}{2}a_{2}+\frac{1}{2}c_{1}^{2 } & = & 0 . \label{22 - 1}\end{aligned}\]]the symmetric solution , with @xmath64 and @xmath65 , can be easily found from here:@xmath66 a_{2}. \label{symm-1}\end{aligned}\]]the search for the sbb , i.e. , solutions with infinitely small @xmath67 and @xmath68 , yields an equation for the bifurcation point : @xmath69 . substituting @xmath70 from eq . ( [ symm-1 ] ) , it takes the following form:@xmath71 % \left [ 2k-\exp \left ( -\kappa _ { 1}\right ) \right ] = 1 . \label{bif-1}\ ] ] further analysis demonstrates that , at @xmath2 , the sbb is supercritical at all values of @xmath3 , including the cascading limit , @xmath72 . the latter fact implies that the sbb is also supercritical in the model with the cubic nonlinearity at two adjacent sites . indeed , this can be checked to be correct in the cubic model , while for all @xmath4 the corresponding sbb is subcritical @xcite ( the character of the sbb for @xmath2 , super- or subcritical , was not considered in ref . @xcite ) . it is also possible to find antisymmetric solutions , with @xmath73 , @xmath74 : @xmath75 a_{2}. \label{anti-1}\end{aligned}\]]a point where the antisymmetry could be spontaneously broken , similar to the sbb of the symmetric modes , is determined by the condition that solutions with infinitely small @xmath76 and @xmath68 emerge . after a simple analysis , this condition leads to the following equation:@xmath71 % \left [ 2k+2-\exp \left ( -\kappa _ { 1}\right ) \right ] = 1 , \label{antibif}\]]cf . ( [ bif-1 ] ) . replacing here @xmath77 and @xmath19 in the square - bracket combinations by their expressions in terms of @xmath61 and @xmath78 given by eqs . ( [ kappa ] ) , it is easy to check that eq . ( [ antibif ] ) can _ never _ be satisfied , unlike eq . ( [ bif-1 ] ) ( the left - hand side always takes values @xmath79 ) , hence the antisymmetric modes _ do not _ undergo the bifurcation . first , we present families of numerical solutions of eqs . ( [ u ] ) and ( v ) with a finite number of sites , @xmath80 , obtained for @xmath5 and @xmath81 . in fig . [ fig1 ] , the families of symmetric , asymmetric , and antisymmetric modes are represented by the @xmath82 curves for the ff and sh components . these curves completely overlap with those predicted by the implicit analytical solutions , in the form of eqs . ( [ lin])-([1 ] ) . in fig . [ fig1](a ) , the family of symmetric solutions features the sbb at a finite value of @xmath19 , and an asymptotic behavior with @xmath83 at @xmath84 ( i.e. , the vanishing of the threshold power ) , in accordance with the analytical approximation based on eqs . ( [ kappakappa ] ) and ( [ sas ] ) . this result is of obvious interest to the experiment , suggesting the possibility to create the nonlinear guided modes , using low - power input beams . on the other hand , fig . [ fig1](b ) demonstrates a completely different behavior of the norms of the antisymmetric solutions at @xmath84 : while @xmath85 approaches a finite value in this limit , @xmath86 diverges at @xmath84 , also in accordance with eqs . ( [ sas ] ) and ( [ kappakappa ] ) ( in terms of the numerical results , both the divergence of @xmath87 and maintaining the finite value of @xmath85 are limited by the finite size of the lattice ) . the stability of the localized modes of the different types , which is also indicated in fig . [ fig1 ] , was established by solving the linear eigenvalue problem for small perturbations , and verified through direct simulations of perturbed evolution in the framework of eqs . ( [ u1 ] ) , ( v1 ) . typical profiles of modes featuring the different types of the symmetry are shown in fig . [ fig2 ] , and the ( in)stability of these modes is illustrated in the first three columns of figs . [ fig2_dyn ] by means of direct simulations . as might be expected , the instability of the symmetric localized modes past the sbb point ( @xmath88 ) transforms it into an excited state close to a stable asymmetric mode . the simulations also corroborate the change of the stability of the antisymmetric modes [ at @xmath89 in fig . [ fig1](b ) ] , which is predicted by the computation of eigenvalues for small perturbations . in particular , the fourth column of fig . [ fig2_dyn ] demonstrates an example of stable antisymmetric modes . to highlight the character of the sbb , we use natural measures of the asymmetry of the solutions taken in the form of eqs . ( [ a12])-([inner ] ) , namely , @xmath90for @xmath5 , the asymmetries are plotted versus @xmath19 and the total power ( norm ) , defined as per eq . ( [ total ] ) , in fig . [ fig5 ] , which demonstrates that the bifurcation is supercritical in this case . a behavior different from that at @xmath5 is observed at @xmath91 . as seen in fig . [ fig3 ] , not only the antisymmetric solutions but also symmetric ones feature the divergence of the norm of the @xmath92-field as @xmath19 approaches the corresponding limit value @xmath93 see eq . ( [ klim ] ) . this feature can also be easily explained by the above analysis : in the limit of @xmath94 , eq . ( [ kappa ] ) yields @xmath95 , hence the respective width @xmath96 diverges , while @xmath78 remains finite [ cf . ( [ kappakappa ] ) ] . further , eqs . ( [ lin])-([1 ] ) yield , in this limit , finite @xmath26 for both symmetric and antisymmetric branches , while the respective value of @xmath97 is vanishing as per @xmath98 cf . eq . ( sas ) ] , hence the corresponding norm @xmath85 vanishes too , while @xmath99 diverges . the linear - stability analysis demonstrates that the symmetric branch is stable up to the sbb point ( @xmath100 ) . past the bifurcation point , the asymmetric branch appears in an unstable form , and with the further increase of @xmath19 it becomes stable at @xmath101 . direct numerical simulations displayed in fig . [ fig3_dyn ] confirm the predictions of the linear - stability analysis . in particular , the unstable asymmetric and antisymmetric modes are transformed by the perturbed evolution into breathers . finally , the sbb in the case of @xmath91 is of the supercritical type , similar to that displayed in fig . [ fig5 ] for @xmath5 . the increase of mismatch @xmath3 to large positive values leads to a transition between the supercritical and subcritical types of the sbb . in particular , fig . [ fig4 ] shows that the @xmath103 branch of the asymmetric solutions passes a shallow minimum right after the bifurcation point , which is a signature of a subcritical bifurcation , while the bifurcation in fig . fig1(a ) was supercritical . the change of the type of the sbb at @xmath104 from super- to subcritical is further illustrated by fig . it is worth noting that the bifurcation is subcritical only in terms of the ff field , @xmath105 . for all @xmath4 it is possible to find a boundary value @xmath106 , such that the sbb is of the super- and subcritical types at @xmath107 and @xmath104 , respectively . numerical analysis of eqs . ( [ m=0 ] ) and ( [ 1 ] ) demonstrates that @xmath106 rapidly grows with the decrease of @xmath1 : @xmath108 , @xmath109 , and @xmath110 . as said above , the sbb keeps its supercritical character at all values of @xmath3 for @xmath2 . the possibility to switch between the different types of the sbb , i.e. , between the phase transitions of the second and first kinds by means of the mismatch , may be readily implemented in the experiment , as the mismatch may be varied by means of the wavelength of the input beam . the trend of the transition to the subcritical bifurcation with the increase of @xmath3 complies with the fact that , as mentioned above , at large @xmath3 the cascading approximation applies to eqs . ( [ u ] ) , ( [ v ] ) , making them asymptotically tantamount to the equation for the single - component ( ff ) lattice with the two sites carrying the cubic self - focusing nonlinearity . in the latter case , the sbb for the localized modes is subcritical for all @xmath111 @xcite . the results of the analysis performed at all values of mismatch @xmath3 , negative , zero , and positive , are summarized in fig . [ fig_ksbb_q ] , where the critical value of the propagation constant at which the sbb happens , @xmath112 , is plotted as a function of @xmath3 at different fixed values of distance @xmath1 between the nonlinear sites . note that , at large @xmath1 , the bifurcation point @xmath113 is very close to the limit value @xmath114 given by eq . ( [ klim ] ) , which is simply explained by the fact that large @xmath1 implies an exponentially weak coupling between the two nonlinear sites , hence the stability margin of the symmetric state is vanishingly small . on the other hand , at large @xmath102 the curves approach constant values , which correspond , in the cascading limit , to the system with the cubic nonlinearity . the objective of this work is to develop the analysis of the stability and spontaneous symmetry breaking of discrete solitary modes in one - dimensional lattices with the quadratic nonlinearity , in the most basic setting with the nonlinearity applied at a two sites embedded into the linear host lattice , with distance @xmath1 between them . the system can be readily implemented as an optical waveguiding array . the solutions for symmetric , asymmetric and antisymmetric modes are obtained in the implicit analytical form , which becomes explicit for @xmath2 . the stability of the modes was investigated through the computation of eigenvalues for small perturbations , and checked against direct simulations of the evolution of perturbed modes . the analysis has demonstrated that properties of the modes undergo essential evolution with the change of the mismatch , @xmath3 : the sbb ( symmetry - breaking bifurcation ) changes from super- into subcritical ( except for the case of @xmath115 ) , the existence threshold for the symmetric localized modes vanishes exactly at the point of @xmath5 , and the stability of the branches changes too . these results suggest a straightforward implementation in the experiment . the analysis reported in this work can be naturally extended in other directions . in particular , following refs . @xcite and @xcite , it may be interesting to consider a finite ring - shaped lattice with the nonlinear sites set at diametrically opposite points . a challenging problem is to generalize the analysis for two - dimensional lattices . e. a. ostrovskaya , y. s. kivshar , m. lisak , b. hall , f. cattani , and d. anderson , phys . a * 61 * , 031601 ( 2000 ) ; r. dagosta , b. a. malomed , c. presilla , phys . lett . a * 275 * , 424 ( 2000 ) ; r. k. jackson and m. i. weinstein , j. stat . phys . * 116 * , 881 ( 2004 ) ; d. ananikian and t. bergeman , phys . a * 73 * , 013604 ( 2006 ) ; e. w. kirr , p. g. kevrekidis , e. shlizerman , and m. i. weinstein , siam j. math 40 * , 566 ( 2008 ) . g. j. milburn , j. corney , e. m. wright , and d. f. walls , phys . a 55 , 4318 ( 1997 ) ; a. smerzi , s. fantoni , s. giovanazzi , and s. r. shenoy , phys . rev . lett . * 79 * , 4950 ( 1997 ) ; s. raghavan , a. smerzi , s. fantoni , and s. r. shenoy , phys . rev . a * 59 * , 620 ( 1999 ) ; k. w. mahmud , h. perry , and w. p. reinhardt , phys . a * 71 * , 023615 ( 2005 ) ; e. infeld , p. zi , j. gocalek , and m. trippenbach , phys . e * 74 * , 026610 ( 2006 ) ; g. theocharis , p. g. kevrekidis , d. j. frantzeskakis , and p. schmelcher , phys . e * 74 * , 056608 ( 2006 ) ; g. l. alfimov and d. a. zezyulin , nonlinearity * 20 * , 2075 ( 2007 ) ; c. wang , p. g. kevrekidis , n. whitaker , and b. a. malomed , physica d * 237 * , 2922 ( 2008 ) . m. albiez , r. gati , j. flling , s. hunsmann , m. cristiani , and m. k. oberthaler , phys . lett . * 95 * , 010402 ( 2005 ) ; r. gati , m. albiez , j. flling , b. hemmerling , and m. k. oberthaler , appl . b * 82 * , 207 ( 2006 ) . c. par and m. forjaczyk , phys . rev . a * 41 * , 6287 ( 1990 ) ; a. i. maimistov , kvant . elektron . * 18 * , 758 [ sov . j. quantum electron . * 21 * , 687 ( 1991 ) ] ; n. akhmediev and a. ankiewicz , phys . lett . * 70 * , 2395 ( 1993 ) ; p. l. chu , b. a. malomed , and g. d. peng , j. opt . soc . a b * 10 * , 1379 ( 1993 ) ; b. a. malomed , in : progr . optics * 43 * , 71 ( e. wolf , editor : north holland , amsterdam , 2002 ) . m. trippenbach , e. infeld , j. gocalek , m. matuszewski , m. oberthaler , and b. a. malomed , phys . a * 78 * , 013603 ( 2008 ) ; n. v. hung , m. trippenbach , and b. a. malomed , _ ibid_. phys . a * 84 * , 053618 ( 2011 ) . b. maes , m. soljai , j. d. joannopoulos , p. bienstman , r. baets , s .- gorza and m. haelterman , opt . express , * 14 * , 10678 ( 2006 ) ; e. n. bulgakov and a. f. sadreev , phys . rev . b * 81 * , 115128 ( 2010 ) ; e. bulgakov , a. sadreev , and k. n. pichugin , _ ibid_. b * 83 * , 045109 ( 2011 ) . o. bang , p. l. christiansen , and c. b. clausen , phys . e * 56 * , 7257 ( 1997 ) ; t. peschel . u. peschel and f. lederer , _ ibid_. e * 57 * , 1127 ( 1998 ) ; s. darmanyan , a. kobyakov , and f. lederer , _ ibid_. e * 57 * , 2344 ( 1998 ) ; s. darmanyan , a. kamchatnov , and f. lederer , _ ibid_. e * 58 * , r4120 ( 1998 ) .	we construct families of symmetric , antisymmetric , and asymmetric solitary modes in one - dimensional bichromatic lattices with the second - harmonic - generating ( @xmath0 ) nonlinearity concentrated at a pair of sites placed at distance @xmath1 . the lattice can be built as an array of optical waveguides . solutions are obtained in an implicit analytical form , which is made explicit in the case of adjacent nonlinear sites , @xmath2 . the stability is analyzed through the computation of eigenvalues for small perturbations , and verified by direct simulations . in the cascading limit , which corresponds to large mismatch @xmath3 , the system becomes tantamount to the recently studied single - component lattice with two embedded sites carrying the cubic nonlinearity . the modes undergo qualitative changes with the variation of @xmath3 . in particular , at @xmath4 , the symmetry - breaking bifurcation ( sbb ) , which creates asymmetric states from symmetric ones , is supercritical and subcritical for small and large values of @xmath3 , respectively , while the bifurcation is always supercritical at @xmath2 . in the experiment , the corresponding change of the phase transition between the second and first kinds may be implemented by varying the mismatch , via the wavelength of the input beam . the existence threshold ( minimum total power ) for the symmetric modes vanishes exactly at @xmath5 , which suggests a possibility to create the solitary mode using low - power beams . the stability of solution families also changes with @xmath3 .
the brightest cluster galaxies ( bcgs ) are the most luminous and most massive galaxies in the universe . bcgs are located close to the centre of the clusters of galaxies based on the x - ray observations or gravitational lensing observations ( e.g. , jones & forman 1984 ; smith et al . it was noted very early on that some bcgs show an excess of light ( ` envelopes ' ) over the de vaucouleurs ( @xmath7 ) profile at large radii ( matthews et al . 1964 ; oemler 1973 , 1976 ; schombert 1986 , 1987 , 1988 ; graham et al . 1996 ) , and a large fraction of bcgs are termed as cd galaxies ( e.g. , patel et al . although the origin of such extended envelopes is still not completely clear ( e.g. , patel et al . 2006 ) , the extended stellar halos of bcgs to surface brightness @xmath8 are likely from bcgs themselves : the intra - cluster light has much lower surface brightness and only dominates at large radius ( @xmath9 ; for detailed discussions , see zibetti et al . 2005 ; bernardi et al . 2007 ; lauer et al . 2007 ) . there has been debate on the formation mechanisms for bcgs and cd galaxies . several mechanisms have been proposed , such as galactic cannibalism ( the merging or capture of cluster satellites due to dynamical friction , ostriker & tremaine 1975 ; white 1976 ; ostriker & hausman 1977 ) , tidal stripping from cluster galaxies ( gallagher & ostriker 1972 ; richstone 1975 , 1976 ; merritt 1985 ) , and star formation on bcgs by cooling flows ( e.g. , fabian 1994 ) . therefore , it is important to study the photometric properties of bcgs using large and homogeneous bcg samples , which has now become available due to large surveys such as the sloan digital sky survey ( sdss ; for earlier studies , e.g. postman & lauer 1995 , see references below ) . moreover , the statistical properties of the extended envelopes of bcgs may also help to understand how the bcgs form and evolve , especially when combined with high - redshift cluster samples such as ediscs ( white et al . 2005 ) . recently , there has been much progress in this area from both observational and theoretical fronts . observations show that bcgs have distinct properties from the bulk of elliptical galaxies . the size - luminosity relation and faber - jackson relation ( @xmath10 relation ) for bcgs are significantly steeper than those of non - bcg elliptical galaxies ( lauer et al . 2007 ; bernardi et al . 2007 ; desroches et al . 2007 ; von der linden et al . 2007 ) . it implies that bcgs are larger and have lower velocity dispersions than non - bcg elliptical galaxies , which confirms the pioneering work based on a small bcg sample by oegerle & hoessel ( 1991 ) . the differences between bcgs and non - bcg elliptical galaxies demonstrate that the bcgs may form in a qualitatively different way from non - bcg elliptical galaxies ( de lucia & blaizot 2007 ) . furthermore , the steeper @xmath10 relation for bcgs than those of non - bcg elliptical galaxies leads to a contradiction in the prediction of black hole ( bh ) masses using the @xmath11 or @xmath12 relationships . the usual @xmath11 relation can not predict black hole masses larger than @xmath13 @xmath14 in the universe , which is inconsistent with values estimated by other means ( lauer et al . 2007 ) . recent simulations and semi - analytic works in the cold dark matter hierarchical structure formation framework provide a plausible picture for the formation of bcgs : they tend to form at high - density peaks when their inhabited dark matter halos collapse at high redshift , and then x - ray driven cooling flows allow a rapid collapse and the formation of a stellar component ( cowie & binney 1977 ; fabian & nulsen 1977 ; fabian 1994 ) . while more than half of the stellar mass in the bcgs may have formed before redshift three , bcgs still grow substantially through dry ( dissipationless ) mergers when their host massive halos accrete and merge with other halos since redshift one ( gao et al . 2004 ; de lucia & blaizot 2007 ) . this picture is largely consistent with observations , which does not require cooling flows to provide the cold gas if bcgs form late ; it also overcomes a potential problem that the merger rate in clusters may be too low due to the high velocity dispersion in dynamically relaxed clusters . furthermore , dry mergers have been directly observed in cluster environments as well as in the field ( e.g. , lauer 1988 ; van dokkum 2005 ; tran et al . 2005 ; bell et al . 2006 ) . however , there are still no systematic investigations on the formation of cd - like bcgs galaxies , especially how the extended stellar halos of cd galaxies form : whether they appear during the dry merger phase , and whether they are the primary cause of the distinct properties of bcgs . there is also no agreement about the contributions of the intra - cluster light ( e.g. , gonzalez et al . 2005 ; zebetti et al . 2005 ; lauer et al . 2007 ) . the c4 cluster catalogue ( miller et al . 2005 ) with 748 cluster of galaxies was constructed from the spectroscopic data of second data release ( dr2 ) of the sloan digital sky survey ( sdss ) . this is one of the largest published homogeneous database of nearby clusters and groups ( @xmath15 ) . in principle , it is suitable for statistical studies of bcgs ( see also the more recent , much larger maxbcg catalogue of 13,823 clusters of koester et al . 2007 ) . based on bcgs from this published cluster catalogue , or the un - published c4 cluster catalogue selected from sdss dr3 , bernardi et al . ( 2007 ) and von der linden et al . ( 2007 ) find that bcgs have different properties , in particular their scaling laws , from non - bcg early type galaxies . their results are , however , not entirely consistent with each other , which could be due to different sample selections or photometry methods . in fact , the slope indices in the size - luminosity and faber - jackson ( @xmath10 ) relations found by bernardi et al . ( 2007 ) are also significantly different from those of lauer et al . ( 2007 ) , although the trends are the same and these two groups use similar methods to measure luminosities and effective radius @xmath16 ; both are based on fitting the @xmath7 law to the @xmath17band surface brightness profiles within 50 or to some surface brightness limit of the bcgs . however , the samples they used are different . lauer et al . ( 2007 ) use their own all sky , volume limited ( z@xmath180.05 ) survey of 119 bcgs with precise ground - based surface photometry ( postman & lauer 1995 ) . therefore , it is worth investigating whether sample selections and photometry can lead to inconsistencies on scaling laws . in this paper , we construct a different , nearby ( with redshift @xmath19 ) bcg sample based on the c4 cluster catalogue , and perform our own photometry . in particular , we examine how the scaling relations change when different isophotal limits are applied , and compare the results with those of bernardi et al . ( 2007 ) , lauer et al . ( 2007 ) and von der linden et al . ( 2007 ) . the outline of the paper is as follows . in 2 and 3 we describe the selection of our local bcg sample and photometric data reduction . we present our main results in 4 and finish with a discussion & summary in 5 . throughout this paper we adopt a cosmology with a matter density parameter @xmath20 , a cosmological constant @xmath21 and a hubble constant of @xmath22 ( @xmath23 ) . our early type bcg sample is drawn from the early version of the c4 cluster catalogue ( miller et al . 2005 ) . given that bcgs are located near the centre of clusters of galaxies which is often crowded with galaxies and contaminated by intra - cluster light , to ensure the reliable photometric measurement with high enough s@xmath24n ratio for each sample bcg to @xmath25@xmath26 mag / arcsec@xmath27 that is significantly higher than the intra - cluster light ( zibetti et al . 2005 ; lauer et al . 2007 ; bernardi et al . 2007 ) , we restrict our bcg sample to be brighter than 14.5 mag in the sdss @xmath1-band model magnitude . after we exclude 10 duplicated objects ( bernardi et al . 2007 ) , there are 114 galaxies selected out of a total of 748 bcgs in the c4 cluster catalogue . due to fiber collisions in the spectroscopic data , it is necessary to incorporate the sdss photometric catalogue to avoid missing bcgs using the c4 algorithm . the dr2 version of the c4 catalogue tried to correct for this by selecting the brightest cluster galaxy based on the photometric catalog . however , as several works ( e.g. , bernardi et al . 2007 ; von der linden et al . 2007 ) noticed , some stars or spiral galaxies ( some of these are located at the edge of clusters ) have been mis - classified as bcgs . we thus performed visual inspections of two - colour images in the @xmath28-band and @xmath1-band for all the 114 bcg candidates and their clusters in a region of @xmath29 and @xmath30 , respectively . we found that c4 1024 , c4 2092 , c4 2097 , c4 2031 , c4 2109 , c4 3021 and c4 3268 are seriously contaminated by adjacent bright stars or foreground galaxies ; c4 3047 is located at the edge of the corrected frame ( stoughton et al . 2002 ) ; c4 3235 is a star ; c4 2178 and c4 3257 are nearly edge - on galaxies with dust lanes . in addition , there are 5 objects ( c4 1265 , c4 1332 , c4 2130 , c4 3098 , and c4 3205 ) whose parent cluster richness , defined as the number of galaxies within 1 @xmath31 , is less than 10 , and hence they may be wrong identifications . in the remaining 98 objects , there are 10 bcgs that are late type galaxies with obvious spiral arms ( c4 1023 , c4 1053 , c4 1324 , c4 1366 , c4 2001 , c4 3016 , c4 3143 , c4 3246 , c4 3282 , and c4 3286 ) . these are also excluded from our early type bcg sample , leaving us with 88 possible early - type bcgs . we further excluded 3 objects ( c4 1048 , c4 1186 and c4 3285 ) with @xmath32 that are fainter than @xmath33 galaxies ( @xmath34 for @xmath35 , see blanton et al . 2003a ) and they are unlikely bcgs ( bernardi et al . 2007 ) . the detailed information for all 29 rejected objects are listed in table 1 . our final sample consists of 85 bcgs that are in the redshift mange of 0.03@xmath36z@xmath360.09 and @xmath37band ( model ) apparent magnitude in @xmath38 . furthermore , we carefully searched for the second ranked cluster galaxies in the 85 c4 clusters and found that the bcgs are about 1.6 mag brighter on average than the second ranked cluster galaxies for the @xmath37band absolute magnitude at @xmath39 . only 6 bcgs have absolute magnitudes within 0.5 mag of the second ranked cluster galaxies . during this process , we found one mis - identified bcg for c4 2020 ( abell 0119 ) , the brightest galaxy in this cluster is ugc 00579 , instead of ark 021 as listed in the c4 catalog . we caution that such cases may also exist in the unpublished c4 catalogue based on the dr3 of the sdss ( see von der linden et al . 2007 ) . notice that there are 4 bcgs ( c4 2049 , c4 1176 , c4 3311 and c4 1035 ) in our bcg sample described above with obvious merger features . c4 2049 , 1176 and 1035 are merger galaxies with two close nuclei and c4 3311 is with clear asymmetrical shapes . we will discuss these four objects in more detail in [ sec : merger ] . in order to perform direct comparisons with non - bcgs , we construct a control sample of normal early - type galaxies from the sdss dr2 spectroscopic catalogue with galaxy velocity dispersion measurement in the mpa gal@xmath40info catalogue.data.html ] these control galaxies are chosen to be in the same redshift and apparent magnitude ranges as sample bcgs , i.e. , 0.03@xmath36z@xmath360.09 and 13.5@xmath36mag@xmath3614.5 . 244 early - type galaxies are selected as the control sample after excluding nearly edge - on objects ( most of them have dust features ) , contaminated sources , and objects identified as c4 bcgs . the distributions of the corrected redshifts relative to the local group ( see blanton et al . 2005 ) , ( m@xmath41 ) colour , and absolute magnitudes at @xmath39 for these two samples are shown in fig . [ redz_col.eps ] . it can be seen that the control sample has a lower redshift range than those of bcgs . this is because non - bcgs are intrinsically fainter than the bcgs24 in the @xmath1-band are bcgs . lin @xmath42 mohr ( 2004 ) found that bcgs are about one magnitude ( 0.83 ) brighter on average than the second ranked cluster galaxies even within a small metric radius of @xmath43 kpc . this does not contradict with the larger difference ( 1.6 mag on average ) we found for our bcg sample because we measure the magnitudes to @xmath39 , which corresponds to a larger physical size than the radius used by them . ] ( see the bottom panel ) , so for the same magnitude limit , they reside more locally . therefore , it is difficult to construct a control sample in the same luminosity range as that of bcgs . instead we attempt a comparison in a narrow luminosity range , @xmath44 ( indicated by two vertical dashed lines at the bottom panel of fig . [ redz_col.eps ] ) where there are sufficient overlaps between the control sample and bcgs ( see [ sec : scaling ] for comparisons ) . furthermore , both samples have colour @xmath45 , implying that all of them are early type galaxies colour - wise ( see blanton et al . 2003b ; shen et al . 2003 and reference therein ) . also the concentration index @xmath46 for virtually all our galaxies is larger than 2.6 , again consistent with their being early - type galaxies ( shimasaku et al . 2001 ; strateva et al . 2001 ) . all the basic parameters for these 85 bcgs are listed in table 2 . [ cols="^,^,^ , < " , ] given that the sdss image survey in the @xmath47 and @xmath48bands are relatively shallow and the @xmath49band images suffer from the ` red halo ' effect ( e.g. , michard 2002 ; wu et al . 2005 ) , we use the @xmath50 and @xmath37band images in this work . the corrected frames have already been processed by the sdss photometric pipeline ( photo ) , including bias - subtraction , flat - fielding , cosmic ray removal and corrections for pixel defects . we use the astrometry for these two bands obtained by the sdss astrometry pipeline ( astrom ) , which has a typical error of less than @xmath51 ( stoughton et al . 2002 ) . on the other hand , for reasons we will discuss below we perform our own photometry . although the images from the sdss archive were processed with photometric corrections , as wu et al . ( 2005 ) pointed out , some spurious features still exist in some images . the spatial variation of these features was about 1 - 2 adu . without corrections , target galaxies that happen to be located within such features will have inaccurate background subtraction and poor surface brightness profiles , especially in the outskirts of the images . we thus first correct for these structures following wu et al . ( 2005 ) before constructing the sky background model below . as lauer et al . ( 2007 ) pointed out that the sdss photometric reduction systematically under - estimates the luminosities and half - light radius of bcgs . this arises because the photo pipeline often over - estimates the sky background for galaxies with large size and/or in crowded fields . both problems are present for bcgs . as a result , bernardi et al . ( 2007 ) performed their own photometric reductions for their sample . on the other hand , von der linden et al . ( 2007 ) address this problem by adding up to 70% of the difference between the local ( 256 @xmath52 256 pixels ) and the global ( 2048 @xmath52 1498 pixels ) sky background ( both are available from photo ) around the bcgs to the radial surface brightness profiles and the photometry . however , the sky background in the sdss frames often shows spatial gradients and asymmetry ( see the third row of fig . [ sky_estimate.eps ] for an example ) . therefore , adding a constant background value to the radial surface brightness profiles may not yield the most accurate photometry , at least on one - to - one basis ( see fig . [ sbp_comparison.eps ] for comparisons ) . therefore , we performed our own photometry following the method of zheng et al . ( 1999 ) and wu et al . ( 2002 ) , which was developed to perform deep photometry for nearby large spiral galaxies . this method has been applied to ngc 5907 and ngc 4565 , which achieved accurate photometry down to an isophotal limit of @xmath53 in the intermediate bands of the batc system ( fan et al . 1996 ) . to obtain accurate information of the sky background , we first generate a 2048 @xmath52 1498 pixels ( @xmath54 ) background - only image using sextractor ( bertin @xmath42 amounts 1996 ) by masking all the detected objects with counts above 1@xmath55 noise ( of the whole frame ) in a frame smoothed by a circular gaussian with a standard deviation @xmath55 = 3 pixels . as most of sample bcgs have sizes of a few square arcminutes , there are sufficient regions in the masked image to determine the sky background . a median filter with @xmath56 pixels is then convolved with the unmasked pixels , after which second - order legendre polynomials are used to fit both rows and columns respectively ( see zheng et al . 1999 and wu et al . the fitted legendre polynomials are then further smoothed using a circular gaussian filter with @xmath55 = 9 pixels to obtain our final sky background model . for most bcgs , we find that the sky background is tilted with a spatial variation about 1 @xmath57 2 adu across the whole frame . we can subtract this model from the frame to obtain the sky - free image . our sky subtraction procedure is similar to that in gonzalez et al . furthermore , the sky background subtracted not only includes the contributions of the intra - cluster light ( which is at a much lower level ) , but also from other astrophysical sources ( dust emission etc . , see gonzalez et al . 2005 for a discussion ) . our procedure is illustrated in fig . [ sky_estimate.eps ] for c4 3086 . the first , second and third rows of fig . [ sky_estimate.eps ] show the frame , the masked frame and the smoothed sky background in the @xmath37band . notice that the sky background shows a gradient across the frame . the bottom panels show the distributions of counts in the sky - subtracted frame for all unmasked pixels and for the local vicinity around the target bcg . if the sky background model is successful , then we expect the background counts to follow a gaussian distribution with a mean close to zero . this is indeed the case , as can be seen from the bottom panels of fig . [ sky_estimate.eps ] . the distributions in the @xmath37band for the whole image and the local region are well - fit by gaussians with dispersion of 5.52adu and a mean value of about @xmath58 adu . after background subtraction , we geometrically align the two - colour frames , and then trim the corrected frames to 501@xmath52501 pixels centred on the target . we run sextractor again on the trimmed frame to generate a ` segmentation ' image , which identifies all objects with flags in the frames . a mask image with all the detected objects except the galaxy of interest flagged can then be obtained from the ` segmentation ' image . we carefully examined all the mask images in two colours and corrected a few bad images manually to create good mask images for all galaxies . photometry is then performed on the trimmed images with the masked areas excluded from the reduction ; we discuss the details below . to compare the scaling relations , such as the size - luminosity relation and faber - jackson relation , of bcgs with those of the bulk of elliptical galaxies , we need to perform accurate surface photometry . in this work , we measure the isophotal magnitudes within a certain radius , instead of magnitudes based on a model ( e.g. , a de vaucouleurs [ 1948 ] model or a s@xmath59rsic [ 1968 ] model ) as a large fraction of bcgs is not well described by such simple models . furthermore , to see how the scaling relations vary as a function of isophotal limits , we measure the photometric parameters to four different surface brightness limits , 22 , 23 , 24 , and 25 mag / arcsec@xmath27 . we emphasize that the photometry for the control sample has been performed in the same manner as that for bcgs . the surface photometry analysis is performed following wu et al . we briefly outline the procedures below ; the readers are referred to that paper for more details . the isophote / ellipse task in iraf is used to fit each of the trimmed background - subtracted images with a series of elliptical annuli from the centre to the outskirts . the width of annuli is chosen to increase uniformly in logarithmic steps , with the semi - major axis radius increasing by 10% between two adjacent annuli . the size of the annuli in the outer parts is therefore larger , which suppresses the shot noise in the outer regions where the signal - to - noise ratio ( s / n ) is lower . the ellipticity , position angle , and other quantities are also fitted simultaneously . the @xmath37band images are first used to define the isophotal annuli , which are then applied to the @xmath50band images . the coordinates of the photometric peak in both @xmath37band and @xmath50band are obtained by the daophot package and are fixed during fitting . we estimate the seeing of our galaxies in two colours by averaging the star profiles in the corresponding frames respectively . the average values of the fwhms of the seeings in both colours are @xmath60 , which are similar to the measurements by the sdss collaboration . we integrate the observed surface brightness profiles directly to estimate the isophotal apparent magnitudes and half - light radius @xmath61 , within which half of the integrated flux of galaxies is contained , under each isophotal limit . the absolute magnitude ( @xmath62 ) is derived from the obtained isophotal apparent magnitude ( @xmath63 ) by @xmath64 , taking into account the extinction ( @xmath65 ) in each filter by the sdss , and the k - correction ( using the kcorrect algorithm of blanton & roweis 2007 ) . other photometric parameters , such as ellipticities , coefficients of lowest order deviations from perfect ellipse @xmath66 and @xmath67 ( see bender et al . 1988 ) , are derived following hao et al . notice that the @xmath68 parameters are weighted by the intensity between twice the seeing radius and the radius where the surface brightness is @xmath39 in the @xmath37band . we discuss the isophotal shapes in sect . [ sec : isophotalshapes ] . all our radial profiles in this paper use the equivalent radius of an ellipse , @xmath69 , where @xmath70 and @xmath71 are the lengths of semi - major and semi - minor axes of the ellipse . the surface brightness profiles in the two bands are performed accounting for the galactic extinction correction , cosmological dimming , and the k - correction ( blanton & roweis 2007 ) . the photometric errors are estimated as in wu et al . the observational errors in the surface brightness profiles in each band include random errors ( e.g. , readout noise , the shot noise of the sky background and bcgs ) and the error from the sky subtraction . to illustrate the performance of our photometry , we show the comparison of our determined petrosian parameters with those of sdss in fig . [ re_contrast.eps ] . one can see that the sdss pipeline under - estimates the sizes ( and thus luminosity ) of bcgs , which has been pointed out by other workers ( e.g. , lauer et al . 2007 ) . there is a systematic trend : the under - estimation becomes more serious for larger ( brighter ) bcgs . this is in good agreement with the results of desroches et al . ( 2007 , see their fig . 1 . ) . as mentioned before , von der linden et al . ( 2007 ) adopted a simple prescription to correct for the sky subtraction problem . in fig . [ sbp_comparison.eps ] we show a comparison between their surface brightness profile , ours and that from the sdss for three example galaxies . it can be seen that our results are in very good agreement with those of von der linden et al . while there are small differences at faint surface brightness ( @xmath39 ) , such differences have little impact on their results since they adopted a surface brightness of @xmath72 to measure the physical quantities . so for statistical purposes , their method can offer an efficient way of performing photometry , particularly at high surface brightness limits . as discussed in the introduction , the surface brightness profiles of the majority of cd galaxies show strong deviation from a perfect de vaucouleurs or s@xmath59rsic profile due to cd galaxies embedded in an extensive luminous stellar halo . in order to quantitatively measure the extended envelope of bcgs , we use an objective method based on the petrosian @xmath73 profiles ( patel et al . 2006 ) , defined as @xmath74 where @xmath25 is the surface brightness in magnitudes at radius @xmath1 and @xmath75 is the mean surface brightness within @xmath1 ( petrosian 1976 ) . there is a distinct signature of a plateau in the petrosian @xmath73 profiles for cd galaxies with an extended stellar halo ( see below ) . such a plateau is not present for normal elliptical galaxies which are usually well - fit by a de vaucouleurs surface brightness profile ( e.g. , kj@xmath76rgaard et al . 1993 ; brough et al . 2005 ) . furthermore , notice that the plateau is not present for the s@xmath59rsic surface brightness profile that better fits some ellipticals ( see fig . 1a of graham et al . 1996 ) . fig . [ example_sbp_eta.eps ] shows the surface brightness profiles and petrosian @xmath73 profiles for 3 bcgs . for bcg c4 1187 , its profile is well fit by a s@xmath77rsic with @xmath78 ( in the @xmath28-band ) model , and the @xmath73 profile does not show any distinct feature . in contrast , for c4 3034 and c4 3077 , their surface brightness profiles deviate significantly from a single s@xmath59rsic profile ( or the @xmath79 profile , @xmath80 for the de vaucouleur profile ) , and the @xmath73 profile shows a plateau . furthermore , we can see that the more significant the deviation of the surface brightness profile from the a single s@xmath59rsic profile ( i.e. , the more prominent the stellar halo in the outskirt ) , the flatter the plateau in the @xmath73 profile becomes . to quantitatively determine how a surface brightness profile deviates from the @xmath79 profile we calculate the gradient of @xmath73 , @xmath81 . as mentioned before , a single @xmath79 model profile always results in a monotonic @xmath82 profile for any @xmath83 . the bottom panels of fig . [ example_sbp_eta.eps ] show this quantity as a function of radius . it is clear that galaxies with the most significant deviation from the model profile also have deeper valleys in the @xmath82 profile . we label the radius of the minimum in @xmath82 as @xmath84 , and also define a depth , @xmath85 , as the difference between the minimum and maximum of @xmath82 outside twice the seeing radius and @xmath84 . the right histogram in fig . [ rela1.eps ] shows the distribution of @xmath85 for our sample bcgs ( the median value is @xmath86 ) . we can see that the distribution of @xmath85 shows a weak bi - modality , but is still consistent with a continuous distribution within poisson error bars . this shows that the deviation of the surface brightness profiles from the a single s@xmath59rsic profile does not have any sharp transitions . therefore , it is difficult to give an unambiguous criterion to separate cd galaxies from non - cd bcgs . the left panel of fig . [ rela1.eps ] shows the correlations between @xmath85 and @xmath84 . the null - hypothesis that there is no correlation between @xmath85 vs. @xmath84 can be rejected at the @xmath87 level . so statistically the larger the radius @xmath84 , the more significant the extended envelope a bcg appears to have . the middle and right panels in fig . [ rela1.eps ] show the correlations between @xmath85 with the @xmath37band luminosity and the half - light radius for bcgs . the null - hypothesis that there are no correlations between @xmath85 vs. @xmath88 , and @xmath85 vs. @xmath89 can be rejected at the @xmath90 , and @xmath91 confidence levels . these correlations demonstrate that more luminous and larger bcgs tend to have more extended stellar envelopes and therefore the fraction of cd galaxies increases . as mentioned in [ sec : sample ] that there are four galaxies in our bcg sample that have either two close nuclei or other clear merging signatures , such as broad faint fans or asymmetry in the morphology . the top panel of fig . [ merger_spec.eps ] shows the image of c4 1176 , from which one can clearly see two nuclei and a broad surrounding fan to the north of the image . the sdss provides the spectra for both nuclei , which are shown in the bottom panels of fig . [ merger_spec.eps ] . from the spectra the redshifts of the two nuclei are 0.0738 and 0.0736 respectively , corresponding to only 60 @xmath92 difference in the line - of - sight velocity . the angular distance between the two nuclei , @xmath93 , implies a projected separation of just @xmath94 . therefore , c4 1176 is most likely a merger galaxy with two nuclei . furthermore , the spectroscopic properties of these two nuclei are similar ; both are typical of early type galaxies with weak emission lines of liner characteristics , indicating that bcg c4 1176 may be a dissipationless ( dry ) merger system . + for the other three bcgs , we can see from the top panel of fig . [ merger_sbp_eta.eps ] that c4 2049 and c4 1035 also have two possible close nuclei . from the sdss spectra , we find that the projected distance between the two nuclei for c4 1035 is only @xmath95 and the relative line - of - sight velocity is 120 @xmath92 , and thus may be another merging bcg . for c4 2049 , the sdss has only spectrum for one nucleus . the angular distance between these two nuclei is @xmath96 . if they are at the same redshift , then the projected distance is @xmath97 . hence c4 2049 is another possible merger bcg . for c4 3311 , although one can not see double nuclei at its centre ( down to the sdss spatial resolution ) , it shows clear asymmetry in shape , which is indicative of merging activities . the second , third and bottom rows of fig . [ merger_sbp_eta.eps ] show the surface brightness profiles , @xmath73 profiles and @xmath98 profiles for all four merger / possible merger bcgs . as can be seen , there are obvious plateaus in the @xmath73 profiles for 3 out of 4 merger bcgs ( c4 2049 , 1176 , and 3311 ) . the @xmath85 values measured from the @xmath98 profile are 1.34 , 1.34 , and 1.04 , for c4 2049 , 1176 , and 3311 respectively , which are much larger than the median @xmath85 value ( @xmath86 ) for the whole bcg sample . therefore , these three bcgs clearly have extended envelopes in their outskirts and are likely cd galaxies . the high fraction of cd galaxies in merger bcgs and the fans corresponding to cd galaxy envelopes suggest that the extended stellar halos of bcgs are from mergers and they appear shortly after merging takes place . in fact , the role of dry ( dissipationless ) mergers in the formation of ellipticals has recently received much attention . as pointed out by bell et al . ( 2006 ) and van dokkum ( 2005 ) , the typical morphological signatures of dry mergers are broad stellar fans , short tidal tail with similar colour as the galaxy itself , and asymmetries at very faint surface brightness levels . the four examples of merger bcg candidates in our sample shown at fig . [ merger_sbp_eta.eps ] indeed have all these signatures , but they also show even more direct merger signs : in three of these four cases , there are two close nuclei . fig . [ example_merger.eps ] shows ten more examples of possible merger systems from our sample . they have broad low - surface brightness fans or distinct asymmetries ( top panels ) ; many visually appear to inhabit in environment with multiple small satellites ( bottom panels ) . these images provide strong evidence for the connection of mergers with the formation of stellar halos in bcgs . the 59 bcgs with spectroscopy from sdss have spectra similar to the examples shown in fig . [ merger_spec.eps ] , indicating our bcg samples are typical early - type galaxies . their spectra do not show significant star formation at the centre , although one third of our bcgs show weak emission lines with liner characteristics . therefore , the mergers appear to be largely dissipationless ( dry ) . we will return to the role of mergers in the formation of bcgs in a future study . we also investigated the shapes of isophotes for bcgs and compared with those of the control sample . the isophotes of ellipticals are usually well fit by ellipses . however , small but significant deviations exist , which are usually described by the fourier components of the deviations . the most significant one is the @xmath68 ( lauer et al . 1985 ; bender et al . 1989 ) , with a positive ( negative ) @xmath68 indicating a disky ( boxy ) isophotal shape . [ a4_mag.eps ] shows how the @xmath68 parameter varies with the luminosity of galaxies for both the bcgs and the control sample . it is clear from fig . [ a4_mag.eps ] that @xmath68 decreases as the luminosity of galaxies increases , hence the fraction of boxy galaxies increases for both samples ; this trend has been pointed out by many authors ( e.g. , bender et al . 1989 ; hao et al . however , the boxy fraction for bcgs ( 51% ) is not much larger than that of the bulk of ellipticals ( 37% ) , even though bcgs are more luminous than the bulk of ellipticals in the control sample . in particular , we can see from table 2 and fig . [ a4_mag.eps ] that the boxy fractions of bcgs with luminosities in the range of @xmath99 and mag@xmath100 are @xmath101 and @xmath102 respectively . so there are still a large fraction of very luminous early type galaxies with disky isophotes ( @xmath103 ) , which is consistent with the hst image results on bcgs by laine et al . faber et al . ( 1997 ) and lauer et al . ( 2005 ) pointed out that galaxies brighter than @xmath104 mag ( corresponding to @xmath105 mag , after correcting the difference in @xmath106 , and taking @xmath107 , krick et al . 2006 ) tend to have ` cored ' luminosity profile , while galaxies fainter than @xmath108 mag ( corresponding to @xmath109 mag ) tend to have ` power - law ' luminosity profiles ( see also rest et al . 2001 ; lauer et al . other works ( e.g. , rest et al . 2001 ) show that bright galaxies are slowly rotating , boxy systems ( davies et al . 1983 ) while faint elliptical galaxies are fast rotating , disky systems ( bender 1988 ; bender et al . 1989 ; nieto et al . 1988 ) . however , bcgs appear to somewhat contradict with this trend : some very luminous bcgs still have disky isophotes . the flattening in elliptical galaxies can arise either from an - isotropic motions or from rotation . faint ellipticals appear to be flattened by rotation ( e.g. , van den bosch et al . 1994 ; lauer et al . 2005 ) , while the luminous ones may be supported by an - isotropic motions . as we have shown in the last subsection , bcgs may have experienced fairly recent merging events , and thus may still retain some of the orbital angular momentum from merging , and their isophotes may be disky . there have already been many works on the scaling relation of bcgs based on homogeneous bcg samples ( bernardi et al . 2007 ; desroches et al . 2007 ; von der linden et al . 2007 ; lauer et al . 2007 ) . however , so far no work has investigated how the slopes change with the isophotal limit , which can in principle provide more knowledge of the dynamical structure of bcgs . the left panels of fig . [ para_contrast.eps ] show how the size ( @xmath61 ) and luminosity ( @xmath110 ) vary with the isophotal limits . it is clear that the size ( @xmath61 ) and luminosity of bcgs increase more sharply than that of the control sample as the isophotal limit becomes deeper . the right panel shows how the mean surface brightness within @xmath61 ( @xmath111 ) and the concentration factor ( @xmath112 ) vary with the isophotal limit . we can see that @xmath111 of bcgs decreases more sharply than that of control sample , but @xmath112 of bcgs has the same trend as that of the control sample as the isophotal limit becomes deeper . therefore , bcgs appear to have significantly larger average sizes and are more diffuse in surface brightness . these trends agree with those found in previous studies ( e.g. , lugger 1984 ; von der linden et al . 2007 ) . fig . [ size_l.eps ] shows the size - luminosity relations for our bcgs and the control sample , measured to four isophotal limits , @xmath113 , @xmath114 , @xmath115 and @xmath116 mag / arcsec@xmath27 respectively . the solid and dotted lines in each panel show the best power - law ( @xmath117 ) fits ; the @xmath4 value is indicated at the top left corner of each panel . per degree of freedom is large , indicating the presence of systematics . to approximately account for these , we renormalised the statistical errors by a constant factor until the @xmath118 per degree of freedom is unity . the errors on the power - law indices are obtained using these renormalised errors . a similar procedure was adopted in von der linden et al . ( 2007 ) . ] clearly the size - luminosity relation becomes steeper as the isophotal limit becomes fainter for both the bcg and the control sample . however , we can also see that the size - luminosity relation for bcgs is steeper than that of the control sample at the same isophotal limit . in particular , the size - luminosity relations for bcgs is already slightly steeper ( but not very statistically significant ) than those of the bulk of ellipticals when the isophotal limit is @xmath119 , corresponding to a physical half - light radius of about 5kpc ( see fig . [ para_contrast.eps ] ) . it indicates that the deviation of size - luminosity relation for bcgs from that of bulk of ellipticals may already occur for the inner regions of ellipticals , and thus is not because most bcgs have extended stellar halos . rather , this suggests the deviations arise from different dynamical structures of bcgs from the bulk of ellipticals . bernardi et al . ( 2007 ) arrived at a similar conclusion by comparing the size - luminosity relations for bcgs with and without extended stellar envelopes ( see also von der linden 2007 ) . the power - law index ( @xmath120 ) for the size - luminosity relation we derived for bcgs at an isophotal limits of @xmath116 mag / arcsec@xmath27 is slightly smaller than the value ( @xmath121 ) obtained by bernardi et al . ( 2007 ) , but the power - law index ( @xmath122 ) for the control sample at the same isophotal limit is a bit larger than their value ( @xmath123 ) . the difference in the slope measured by us and by bernardi et al . ( 2007 ) may be because we measure the isophotal magnitudes by integrating the observed surface brightness profiles directly , while bernardi et al . ( 2007 ) measured the magnitude using a de vaucouleurs or s@xmath77rsic model for the surface brightness profile . in fact , when we use the same method as bernardi et al . ( 2007 ) to perform photometry , the power - law index for the size - luminosity relation we obtain is almost the same ( @xmath124 ) as that of bernardi et al . ( 2007 ) , indicating that different ways of photometry can give different power - law indices in the size - luminosity relation . von der linden ( 2007 ) also derived the size - luminosity relation down to 23 @xmath0 for their bcgs . the slope they derived is @xmath125 , similar to that for their control sample ( @xmath126 ) . our slope for the bcgs ( @xmath127 ) is steeper than their value at the same isophotal limit , while the slope for our control sample ( @xmath128 ) agrees well with theirs . we have 66 overlapping galaxies with the bcg samples used by von der linden et al . if we use only these galaxies , we derive a slope similar to theirs . our steeper slope appears to arise from the fact that our bcgs are brighter since their sample includes more fainter bcgs which makes their slope shallower ( von der linden 2007 , private communication ) . therefore , the sample construction also influences the power - law index in the size - luminosity relation . the power - law index @xmath129 in the size - luminosity relation by lauer et al . ( 2007 ) for their core galaxies with @xmath130 ( most core galaxies with @xmath131 are bcgs ) measured in a similar way as bernardi et al . ( 2007 ) is significantly steeper than those of bernardi et al . ( 2007 ) , von der linden ( 2007 ) and our results . it again illustrates that how samples are constructed and the methods used to do the photometry lead to different power - law indices in the size - luminosity relation . comparisons between different studies should be aware of this . it is well known that the luminosities and velocity dispersions of many early type galaxies satisfy the faber - jackson relation , @xmath132 , where @xmath133 ( faber & jackson 1976 ) . given that the sdss spectra were taken within a 3@xmath134 fiber aperture ( j@xmath135rgensen et al . 1995 ; bernardi et al . 2003a ) , we correct the velocity dispersions for bcgs and the control sample following von der linden et al . ( 2007 , see their equation 4 ) . the uncorrected value of @xmath55 for each bcg is shown in column 5 of table 2 . [ l_sigma.eps ] shows the @xmath10 relation for 59 out of 85 bcgs and the control sample with sdss spectral information , measured to four different isophotal limits , 22 , 23 , 24 , and 25 @xmath0 . clearly at a given surface brightness limit , the slope for bcgs in the fabor - jackson relation is steeper than that of the bulk of early type galaxies . the index @xmath5 for both the bcg and control samples increases as the isophotal limits become deeper . the @xmath5 value for the control sample is around 4 , consistent with the canonical value for the faber - jackson relation . however , this is not the case for bcgs : only when the isophotal limit is 22 @xmath0 , corresponding to the inner region of bcgs , @xmath5 is about 4 . for deeper isophotal limits , the @xmath5 values for bcgs are significantly larger , approaching 5.6 for a surface brightness limit of @xmath136 . von der linden ( 2007 ) found @xmath137 for their bcgs at an isophotal limit of @xmath72 , which is within @xmath138 of our value ( @xmath139 ) . for their control samples , they found @xmath140 , within @xmath141 of our result ( @xmath142 ) . lauer found @xmath143 for their core galaxies with @xmath144 , which is significantly larger than those of values mentioned above . note that the sample lauer et al . ( 2007 ) used is not a pure bcg sample , although most of their core galaxies with @xmath145 are bcgs . the steeper power - law indices in the size - luminosity and ( to a less extent ) in the faber - jackson relations may arise because bcgs are brighter than the control sample or because there are intrinsic differences between bcgs and the bulk of elliptical galaxies . in order to differentiate these two possibilities , we repeat the above analysis for our bcg and control samples in the same magnitude range , @xmath44 ( see [ sec : sample ] ) . fig . [ scaling_part.eps ] shows the size - luminosity ( left ) and faber - jackson ( right ) relations . the top panels use all the 94 control - sample galaxies and 45 bcgs within the luminosity range . since fig . [ redz_col.eps ] shows the control sample is still fainter on average than the bcg sample within this narrow range , we have randomly selected 45 out of the 94 control - sample galaxies so they match well with the luminosity distribution of the 45 bcgs . the statistical results for different realisations are similar ; the bottom panels shows the results for one realisation . from this figure , we clearly see that the power - law indices for the size - luminosity relations for bcgs are steeper than those for bright non - bcg elliptical galaxies ; the power - law indices for the faber - jackson relations are also steeper , but only at @xmath146 levels , and so are not statistically significant . to summarise , the data are consistent with a scenario where the cluster environment plays an important role on the bcg formation . however , the numbers of galaxies in our control and bcg samples are still too small to differentiate the two possibilities mentioned above . [ fp.eps ] shows the fundamental plane for our bcgs and the control sample . while they appear to have different power - law slopes in the faber - jackson relation , both bcgs and the control sample follow roughly the same fundamental plane , although the scatters in the bcgs appear smaller . at bright isophotal limits , @xmath147 , and @xmath72 , the fundamental planes for bcgs and the control samples appear to differ somewhat . these results are in good agreement with bernardi et al . in this paper we have studied the properties of 85 bcgs carefully selected from the c4 cluster catalog ( miller et al . 2005 ) . to address the problems of the sdss pipeline photo in crowded fields and for large galaxies , we have performed our own photometry for these bcgs , as well as for a control sample of galaxies in similar redshift and apparent magnitude ranges . particular attention was paid to accurate extraction of the sky background in clusters using a method developed for high - precision photometry for large , nearby spiral galaxies ( zheng et al . 1999 ; wu et al . 2002 , see [ sec : data ] ) . it came as somewhat a surprise to us that the simple recipe of von der linden et al . ( 2007 ) gives surface brightness profiles similar to ours , and hence for many statistical purposes their simple recipe can be adopted as a more efficient alternative . we analysed the surface brightness properties of our bcgs using the @xmath73 profiles ( see eq . [ eq : eta ] ) . we demonstrate that its gradient , @xmath148 , can be used as an indicator of extended envelopes of cd galaxies ( patel et al . furthermore , we find that the depth of valley , @xmath85 in the @xmath82 profiles can be used to quantitatively measure the properties of the plateaus in the surface brightness profiles . generally , the deeper the valleys are ( corresponding to a larger value of @xmath85 ) , the flatter and wider the plateaus are in the @xmath73 profiles . however , the @xmath85 parameter appears to vary in a continuous way with the galaxy parameters , such as their size and luminosity . while there is a clear trend that more luminous and larger bcgs have larger stellar halos . the continuous variation makes it difficult to un - ambiguously classify cd or non - cd galaxies based on the surface brightness profiles alone . we measured the photometric properties of galaxies to four different isophotal limits , @xmath149 , and @xmath39 , and used these to investigate the scaling laws and their dependence on these limits . the latter is of particular relevance as some previous investigations derived scaling relations for bcgs using different samples and different isophotal limits ( e.g. , the study of von der linden et al . 2007 used an isophotal limit of 23@xmath0 ) . we find that the size - luminosity relation and fabor - jackson ( @xmath10 ) relations of bcgs are consistent with being steeper than those of the bulk of early type galaxies for a control sample selected within the same absolute magnitude range ( @xmath150 ) , although the latter is not statistically significant due to the limited sample sizes . furthermore , as the photometric limit becomes deeper , the size - luminosity relation for bcgs and the bulk of the ellipticals becomes steeper . however , there are already ( small ) differences between bcgs and regular ellipticals at @xmath151 , a surface brightness that probes the inner parts of galaxies . this , together with the relatively small difference in the scaling relations between @xmath152 and @xmath39 , suggests that there may be intrinsic differences in the dynamic structures between bcgs and the bulk of early type galaxies ( bernardi et al . 2007 ; lauer et al . 2007 ) . one possibility may be that the amount of dark matter in the central parts of bcgs and regular ellipticals is different ( von der linden et al . 2007 ) , or perhaps the relative importance of an - isotropy and rotation may be different . in this context , it is interesting to note that we find a large fraction of very luminous early type galaxies with disky isophotes . this appears to somewhat contradict with the usual expectation that the most luminous galaxies should show boxy isophotes . also notice that it is difficult to measure high - redshift galaxies to very faint surface brightness due to the cosmological dimming . thus a proper comparison between local and high - redshift samples will need to explicitly account for the surface brightness dependence we found in the paper . in this work we uncovered four strong dry merger candidate bcgs ( see fig . [ merger_sbp_eta.eps ] ) , three of which have two close nuclei within @xmath153kpc and with nearly identical radial velocities . we also found another 10 bcgs in our sample with possible dry merger signatures ( see fig . [ example_merger.eps ] ) , as advocated by bell et al . ( 2006 ) . this strongly suggests merging at the centre of clusters is an important physical process that affects a variety of properties of bcgs , both in photometry and kinematics . an analysis of the statistics for the fraction of dry mergers in bcgs is ongoing and the results will be reported elsewhere . we thank c. n. hao , simon white for discussions , and in particular a. von der linden for sharing data , helpful discussions and comments on a draft . we acknowledge dr . tod lauer , the referee , for a constructive report that improved the paper . this project is supported by the nsf of china 10333060 , 10273012 , 10640430201 , 10773014 and 973 programs no . 2007cb815405 and 2007cb815406 . sm acknowledges the chinese academy of sciences and the alexander von humboldt foundation . funding for the creation and distribution of the sdss archive has been provided by the alfred p. sloan foundation , the participating institutions , the national aeronautics and space administration , the national science foundation , the u.s . department of energy , the japanese monbukagakusho , and the max planck society . the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium ( arc ) for the participating institutions . the participating institutions are the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , the korean scientist group , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , university of pittsburgh , princeton university , the united states naval observatory , and the university of washington . bell e. f. et al . , 2006 , apj , 640 , 241 bender r. , 1988 , a&a , 193 , l7 bender r. , dbereiner s. , mllenhoff c. , 1988 , a&as , 74 , 385 bender r. , surma p. , doebereiner s. , moellenhoff c. , madejsky r. , 1989 , a&a , 217 , 35 bernardi m. et al . , 2003a , aj , 125 , 1817 bernardi m. et al . , 2003b , aj , 125 , 1866 bernardi m. , hyde j. b. , sheth r. k. , miller c. j. nichol r. c. , 2007 , aj , 133 , 1741 bertin e. , arnouts s. , 1996 , a&as , 117 , 393 binney j. , merrifield m. , 1998 , galactic astronomy ( princeton : princeton univ . press ) blanton m. r. et al . , 2003a , apj , 592 , 819 blanton m. r. et al . , 2003b , apj , 594 , 186 blanton m. r. et al . , 2005 , aj , 129 , 2562 blanton m. r. , roweis s. , 2007 , aj , 133 , 734 brough s. , collins c. a. , burke d. j. , lynam p. d. , mann r. g. , 2005 , mnras , 364 , 1354 cowie l. l. , binney j. , 1977 , apj , 215 , 723 davies r. l. , efstathiou g. , fall s. m. , illingworth g. , schechter p. l. , 1983 , apj , 266 , 41 de lucia g. , blaizot j. , 2007 , mnras , 375 , 2 desroches louis - benoit , quataert e. , ma chung - pei , west a. a. , 2007 , mnras , 377 , 402 de vaucouleurs g. , 1948 , ann . astrophys . , 11 , 247 faber s. m. et al . , 1997 , aj , 114 , 1771 faber s. m. , jackson r. e. , 1976 , apj , 204 , 668 fabian a. c. , 1994 , ara&a , 32 , 277 fabian a. c. , nulsen p. e. j. , 1977 , mnras , 180 , 479 fan x. et al . , 1996 , aj , 112 , 628 gallagher j. s. , ostriker j. p. , 1972 , aj , 77 , 288 gao l. , loeb a. , peebles p. j. e. , white s. d. m. , jenkins a. , 2004 , apj , 614 , 17 gonzalez a. h. , zabludoff a. i. , zaritsky d. , 2005 , apj , 618 , 195 graham a. , lauer t. r. , colless m. , postman m. , 1996 , apj , 465 , 534 hao c. n. , mao s. , deng z. g. , xia x. y. , wu h. , 2006 , mnras , 370 , 1339 ; erratum , mnras , 373 , 1264 jones c. , forman w. , 1984 , apj , 276 , 38 j@xmath135rgensen i. , franx m. , kj@xmath76rgaard p , 1995 , mnras , 276 , 1341 kj@xmath76rgaard p. , j@xmath135rgensen i. , moles m. , 1993 , apj , 418 , 617 koester b. p. et al . , 2007 , apj , 660 , 239 krick j. e. , bernstein r. a. , pimbblet k. a. , 2006 , aj , 131 , 168 laine s. , van der marel r. p. , lauer t. r. , postman m. , odea c. p. , owen f. n. , 2003 , aj , 125 , 478 lauer t. r. , 1985 , mnras , 216 , 429 lauer t. r. , 1988 , apj , 325 , 49 lauer t. r. et al . , 2005 , aj , 129 , 2138 lauer t. r. et al . , 2007 , apj , 662 , 808 lin y .- t . , mohr j. j. , 2004 , apj , 617 , 879 lugger p. m. , 1984 , apj , 286 , 106 matthews t. a. , morgan w. w. , schmidt m. , 1964 , apj , 140 , 35 merritt d. , 1985 , apj , 289 , 18 michard r. , 2002 , a&a , 384 , 763 miller c. j. et al . , 2005 , aj , 130 , 968 nieto j .- l . , capaccioli m. , held v. e. , 1988 , a&a , 195 , l1 oegerle w. r. , hoessel j. g. , 1991 , apj , 375 , 15 oemler a. , jr . , 1973 , apj , 180 , 11 oemler a. , jr . , 1976 , apj , 209 , 693 ostriker j. p. , hausman m. a. , 1977 , apj , 217 , l125 ostriker j. p. , tremaine s. d. , 1975 , apj , 202 , l113 patel p. , maddox s. , pearce f. r. , aragn - salamanca a. , conway e. , 2006 , mnras , 370 , 851 petrosian v. , 1976 , apj , 209 , l1 postman m. , lauer t. r. , 1995 , apj , 440 , 28 rest a. , van den bosch f. c. , jaffe w. , tran h. , tsvetanov z. , ford h. c. , davies j. , schafer j. , 2001 , aj , 121 , 2431 richstone d. , 1975 , apj , 200 , 535 richstone d. , 1976 , apj , 204 , 642 schombert j. m. , 1986 , apjs , 60 , 603 schombert j. m. , 1987 , apjs , 64 , 643 schombert j. m. , 1988 , apj , 328 , 475 s@xmath77rsic j. l. , 1968 , atlas de galaxies australes ( cordoba : observatories astronomica ) shen s. et al . , 2003 , mnras , 343 , 978 shimasaku k. et al . , 2001 , aj , 122 , 1238 smith g. p. , kneib j .- , smail i. , mazzotta p. , ebeling h. , czoske o. , 2005 , mnras , 359 , 417 stoughton c. et al . , 2002 , aj , 123 , 485 strateva i. et al . , 2001 , aj , 122 , 1861 tran k .- v . , van dokkum p. g. , franx m. , illingworth g. d. , kelson d. d. , schreiber n. m. f. , 2005 , apj , 627 , l25 van den bosch f. c. , ferrarese l. , jaffe w. , ford h. c. , oconnell r. w. , 1994 , aj , 108 , 1579 van dokkum p. g. , 2005 , aj , 130 , 2647 von der linden a. et al . , 2007 , mnras , 379 , 867 white s. d. m. , 1976 , mnras , 174 , 19 white s. d. m. et al . , 2005 , aap , 444 , 365 wu h. et al . , 2002 , aj , 123 , 1364 wu h. , shao z. y. , mo h. j. , xia x. y. , deng z. g. , 2005 , apj , 622 , 244 zibetti s. , white s. d. m. , schneider d. p. , brinkmann j. , 2005 , mnras , 358 , 949 zheng z. y. et al . , 1999 , aj , 117 , 2757	we investigate the photometric properties of the early type brightest cluster galaxies ( bcgs ) using a carefully selected sample of 85 bcgs from the c4 cluster catalogue with redshift less than 0.1 . we perform accurate background subtractions , and surface photometry for these bcgs to 25 @xmath0 in the sloan @xmath1-band . by quantitatively analysing the gradient of the petrosian profiles of bcgs , we find that a large fraction of bcgs have extended stellar envelopes in their outskirts ; more luminous bcgs tend to have more extended stellar halos that are likely connected with mergers . a comparison sample of elliptical galaxies was chosen with similar apparent magnitude and redshift ranges , for which the same photometric analysis procedure is applied . we find that bcgs have steeper size - luminosity ( @xmath2 ) and faber - jackson ( @xmath3 ) relations than the bulk of early type galaxies . furthermore , the power - law indices ( @xmath4 and @xmath5 ) in these relations increase as the isophotal limits become deeper . for isophotal limits from 22 to 25 @xmath0 , bcgs are usually larger than the bulk of early type galaxies , and a large fraction ( @xmath6 ) of bcgs have disky isophotal shapes . the differences in the scaling relations are consistent with a scenario where the dynamical structure and formation route of bcgs may be different from the bulk of early type galaxies , in particular dry ( dissipationless ) mergers may play a more important role in their formation ; we highlight several possible dry merger candidates in our sample . [ firstpage ] galaxies : e / s0s galaxies : cd galaxies : cluster of galaxies galaxies : photometry galaxies
we have previously observed that @xmath0 , @xmath1 and charged hadron yields are significantly suppressed especially for the high @xmath3 region ( @xmath3 @xmath7 @xmath8 gev / c ) in au+au collision at 200 gev compared with p+p collisions.@xcite@xcite@xcite@xcite since there is no suppression in d+au collisions at high @xmath3,@xcite it is understood that the suppression occurs due to the final state interaction at the collision such as the gluon radiation in the hot dense matter . another evidence for the suppression being a final state effect comes from the non - suppression of the direct photon yield in au+au collisions.@xcite to understand the character of the suppression more , the comparison between different system size ( p+p / d+au / cu+cu / au+au ) measurements has been studied , and an extended @xmath3 reach to 20 gev / c has been afforded by the long run4 au+au dataset . we newly measured the following spectra . * @xmath0 spectra with extended @xmath3 range in au+au at 200gev * @xmath0 and charged hadron spectra in cu+cu at 200gev * @xmath6 spectra in d+au and p+p at 200gev @xcite * newly finalized @xmath1 spectra in p+p , d+au , and au+au @xcite the phenix experiment consists of four spectrometer arms ( two central arms and two muon arms ) and a set of global detectors . each central arm covers the pseudorapidity range @xmath9 @xmath1 @xmath9 @xmath10 @xmath11 and 90 degrees in azimuth . charged particles are tracked by a drift chamber ( dc ) and pad chambers ( pc ) in each central arms . the electromagnetic calorimeters ( emcal ) are used to measure @xmath12 energy deposit and construct the invariant masses of @xmath0(@xmath13 ) , @xmath6(@xmath14 ) and @xmath1(@xmath13 ) . the fig.[fig : raa ] , shows the comparison of @xmath2 for @xmath0 and charged hadrons in 0 - 10 @xmath15 most central au+au and cu+cu collisions as function of @xmath3 . both @xmath0 and charged hadron are strongly suppressed in both au+au and cu+cu collision . the difference between @xmath0 and charged hadron for @xmath16 gev / c comes from the proton contribution . for more central collision , the suppression is getting stronger and the difference between @xmath0 and charged hadron is getting larger . above @xmath17 gev / c , @xmath18 becomes flat out to 20 gev / c . these results are consistent with the model predicting parton energy loss in the medium@xcite and the model predicting shadowing , cronin effect , and parton energy loss in the medium.@xcite in addition , @xmath1 is also suppressed in central au+au collisions and the suppression pattern is similar to @xmath0 . @xcite -6 mm for @xmath0 and charged hadron at 0 - 10 @xmath15 centrality bin in au+au ( left ) and cu+cu ( right ) collisions as a function of @xmath3 with theoretical prediction ( red @xcite and purple @xcite lines ) . the error bars are statistical error , and the boxes are systematic error . ] for @xmath0 and charged hadron at 0 - 10 @xmath15 centrality bin in au+au ( left ) and cu+cu ( right ) collisions as a function of @xmath3 with theoretical prediction ( red @xcite and purple @xcite lines ) . the error bars are statistical error , and the boxes are systematic error . ] -10 mm -20 mm in fig.[fig : raa_theory](a ) , the comparison of @xmath2 in au+au to that in cu+cu with similar @xmath4 is shown . the suppression is similar for similar @xmath4 at mid - centrality . though a universal @xmath4 scaling is roughly suggested,@xcite due to the size of the uncertainties , it is unclear whether a single scaling curve can exactly describe the suppression ( @xmath2 ) in both systems simultaneously . in fig.[fig : raa_theory](b ) , we show the integrated @xmath2 of @xmath0 at @xmath19 gev / c with two different theoretical curves @xcite@xcite as a function of @xmath4 . both models are consistent with the data from central to mid - central collisions . -10 mm -15 mm the ratio of @xmath1 to @xmath0 is @xmath20 in all systems and for all centralities as shown in fig.[fig : ratio_eta].@xcite . also , the ratio of @xmath6 to @xmath0 at p+p and d+au becomes flat at high @xmath3 as shown in fig.[fig : ratio_ks].@xcite therefore , the mesons are affected by the medium in the same way in different collision systems . these results are consistent with jet fragmentation at high @xmath3 . -5 mm to @xmath0 in au+au ( left ) , d+au ( right ) and p+p ( right ) as a function of @xmath21 at @xmath22 gev . the error bars are statistical error , and the boxes are systematic error . ] to @xmath0 in au+au ( left ) , d+au ( right ) and p+p ( right ) as a function of @xmath21 at @xmath22 gev . the error bars are statistical error , and the boxes are systematic error . ] -8 mm to @xmath0 in p+p and d+au as a function of @xmath21 . the opened symbols show the ratio of of @xmath23 to @xmath24 as reference.@xcite the error bars are statistical error , and the boxes are systematic error . ] -10 mm -10 mm we have studied @xmath0,@xmath1 , @xmath6 and charged hadron spectra in au+au , cu+cu , d+au and p+p at high @xmath3 . for @xmath0 and charged hadron , we observed the suppressions in both cu+cu and au+au collisions compared with p+p collisions , and no suppression is observed in d+au collisions . the @xmath2 comparison between au+au and cu+cu indicates that the suppression is almost the same for similar @xmath4 . a universal @xmath4 scaling of @xmath2 , independent of system , describes the data in an approximate sense , but it can not be confirmed exactly due to experimental uncertainties . in addition , the high @xmath3 @xmath0 suppression is flat out to 20 gev / c and its magnitude is quantitatively consistent with parton energy loss model calculations . @xmath1 has a similar suppression pattern as @xmath0 does . the ratio of @xmath0 to @xmath1 is independent of centralities and system size . similarly , the @xmath6 to @xmath0 ratio is constant within uncertainties at high @xmath3 for both p+p and d+au . these particle ratios are consistent with normal jet fragmentation . 12 s. s adler et al . , phenix collaboration , phys . 91 , 241803 ( 2003 ) . s. s adler et al . , phenix collaboration , phys . 91 , 072301 ( 2003 ) . s. s adler et al . , phenix collaboration , phys . 91 , 072303 ( 2003 ) . s. s adler et al . , phenix collaboration , phys . c 69 , 034910 ( 2004 ) . s. s adler et al . , phenix collaboration , phys . 94 , 232301 ( 2005 ) . k. adcox et al . , nucl . methods a499 , 469 ( 2003 ) . a. drees , h. feng , j. jia , phys . c 71 , 034909 ( 2005 ) . v. s. pantuev , hep - ph/0506095 ( 2005 ) . i. vitev , phys . rev 89 , 252301 ( 2002 ) . x. wang , phys . b 595 , 165 - 170 ( 2004 ) . v. ryabov for the phenix collaboration , ( in the same proceedings ) s. s adler et al . , phenix collaboration , update of `` high @xmath3 @xmath1 meson production in p+p , d+au and au+au collisions at @xmath25 gev '' soon to be submitted to plb s. s adler et al . , phenix collaboration , update of `` identified charged hadron production in @xmath22 gev p+p collisions '' soon to be submitted to prd	phenix has extended the measurement of the @xmath0 , @xmath1 , identified and inclusive charged hadron up to 20 gev / c , and extended the measurement to the cu+cu collision system . a strong suppression is observed for both @xmath0 and charged hadron yields in central au+au and cu+cu collisions . comparing au+au and cu+cu systems , @xmath2 becomes independent of @xmath3 above 5 gev / c . its centrality dependence is compared with two models in order to test for universal @xmath4 scaling that is independent of system ; results are inconclusive . the results are compatible with energy loss predictions . in addition , the ratio of @xmath1 to @xmath0 approaches , within uncertainties , a constant value of @xmath5 at high @xmath3 in p+p , d+au , and au+au , while the ratio of @xmath6 to @xmath0 is also consistent with a constant value at high @xmath3 in d+au and p+p . these results are compatible with normal jet fragmentation .
as the scale of architectures in integrated circuit design continues to be reduced , the dimensions of present - day interconnects are on the scale of tens of nanometers , and the interconnects for next - generation nanoelectronic devices may well be just a few nanometers in lateral dimension . furthermore , the incorporation of nanometer - scale components , such as active molecules , into integrated circuits will require interconnects of a similar scale . on these scales , self - assembled nanowire systems are becoming increasingly interesting@xcite , as conventional lithographic techniques reach their limits around 10 - 15nm@xcite and spm - based nanolithography@xcite methods lack scalability . moreover , the introduction of such `` bottom - up '' technology , based on naturally nanometer - scale components to complement or even replace the current `` top - down '' technology is likely to require completely different architectures . one example is the `` crossbar '' architecture@xcite , in which active molecules are used as devices at the junctions between two perpendicular nanowires . this architecture is designed to take advantage of the typical product of self - assembly schemes an array of parallel wires rather than relying on controlled positioning of individual wires . over the last ten to fifteen years , a number of systems have been identified which yield self - assembled nanowires on semiconductor surfaces , and stem from a mixture of serendipity and deliberate design . there are only a few systems which provide long , straight nanolines on flat terraces . in / si(111)@xcite ( amongst other metals ) and pt / ge(001)@xcite are examples of systems which undergo surface reconstructions leading to 1d atomic chains with metallic properties . on the si(001 ) surface , the bi nanolines@xcite appear to be the only example of an atomically - perfect , straight , self - assembled 1d nanostructure , but are not metallic . aside from fortuitous discoveries , symmetry - breaking by use of a vicinal surface to produce and array of steps can be used to grow long , straight nanowires of many metals , including au , bi et al.@xcite , and 1-d nanoscale structures can be designed by taking advantage of anisotropy in the heteroepitaxial strain between the si(001 ) surface and an appropriate deposited material . the rare - earth silicides are the archetype of this method@xcite . within the family of rare - earths ( as well as scandium and yttrium ) , a close lattice match can be obtained in one direction , with a variety of positive and negative lattice mismatchs in the orthogonal direction . by choosing the right material , metastable 1d nanoscale structures 3 - 10 nm wide and hundreds of nanometres long can be formed . these have some phenomenological similarities with the bi nanolines , as will be described . this review will survey the growing field of nanowires on semiconductor surfaces , with the focus on the bi / si(001 ) nanoline system , whose importance is explained in the next section . the study of semiconductor surfaces has been revolutionised over the last twenty years by the advent of two techniques which give real - space atomic scale information : scanning tunneling microscopy ( stm ) and electronic structure modelling ( in particular , density functional theory , or dft ) . the two techniques are complementary , in that stm gives an approximate answer to an exact question ( i.e. the structure of the real si(001 ) surface ) , while dft gives an exact answer to an approximate problem ( i.e. the lowest - energy structure of a computation cell containing a few si dimers , and a few layers of bulk - like si ) . while these techniques between them enabled the clear identification of the si(001 ) reconstruction , even today there is some controversy about the exact nature of the buckling of dimers on si(001 ) at low temperatures@xcite . in the same way , the understanding of the structure and properties of the nanolines surveyed in this paper has been greatly enhanced by a close collaboration between experiment and theory . the benefits of this synergistic approach will be drawn out throughout the discussion . the bi : si(001 ) nanoline system , which will be the focus of this review , was discovered about ten years ago . as can be seen from the examples in fig . [ fig : binanoline ] , they grow perfectly straight along the @xmath0 directions on the si(001 ) surface for hundreds of nanometres , apparently limited only by the terrace size of the underlying substrate . when they encounter a step edge , they will either grow out over lower terraces in a long , narrow promontory , or they will burrow into higher terraces , until a deep inlet is formed . they have a constant width , 1.5 nm or four substrate dimers , and are very stable . so long as the temperature is not so high that the bi can evaporate , they are stable against prolonged annealing , maintaining the same width . this is quite unlike the rare earth silicide systems which in most cases ( yttrium being the probable exception ) will coarsen into 3d epitaxial islands with long anneals . there are no other elements which produce perfect nanolines on the si(001 ) surface , although there is some evidence to suggest that sb nanowires with the same appearance as the bi nanoline form in the presence of surface hydrogen@xcite , and dft modelling that indicates that a sb nanoline would be stable@xcite . there has also been a suggestion that er may form nanolines of the same structure@xcite ( as opposed to silicide wires ) , but this remains unconfirmed . m@xmath11@xmath2 m . the bi nanolines run along @xmath0 directions , perpendicular to the si dimer rows , and can grow to over 1@xmath2 m in length . in the inset , the atomic structure , comprising a pair of bi dimers , about 6.5 apart , is shown . hydrogen termination of the bi nanolines has been used to enhance the apparent height of the nanoline . ] structurally , the bi nanoline is unique amongst nanoline systems it is neither a periodic reconstruction of the surface , nor the result of anisotropically strained heteroepitaxial growth of one bulk structure on another . in fact , it is somewhere in between . the nanoline structure is built around a pair of bi ad - dimers , on top of a complex subsurface reconstruction which is responsible for many of the nanoline s remarkable properties , such as the extreme straightness ; this will be explained in detail below . in this article , we intend to describe the formation , structure and properties of this system , setting it in the context of other nanoline systems , particularly the rare - earth silicide family . the paper is laid out as follows : in section [ sec : methods ] we consider the methods used to examine the systems , both experimental and theoretical ; in section [ sec : other - nanowires ] self - assembled nanowire systems _ other _ than the bi : si(001 ) nanowire system are described , with particular reference to the rare - earth silicide family ; in section [ sec : bi - nanowires ] the bi : si(001 ) nanowire system is examined in detail , and in section [ sec : conclusions ] the state of the field is considered and conclusions are drawn . in considering the nanowires in sections [ sec : other - nanowires ] and [ sec : bi - nanowires ] we consider first the formation and structure of the systems ( including the effects of annealing ) , followed by the electronic properties ( including conductance data where available ) and finish with the reactivity of the systems . since its invention in 1982 by binnig and rohrer at ibm , stm has allowed researchers to image dynamic surface processes , in real - space , with near - atomic resolution . stm as a technique bridges a variety of disciplines . in surface physics , stm is unique as a real - space probe of electronic density of states of individual species ; the stm tip can be a passive observer of surface chemical reactions , and also an active catalyst of these reactions ; and stm allows for the direct observation of many phase transformation processes , albeit in 2d rather than 3d . however , this is not meant to be a review of capabilities of stm ; for that , the reader is pointed to classic books on the subject@xcite . here a few comments are made , which are pertinent to the particular challenges of elevated - temperature ( ` hot ' ) stm for the study of epitaxial growth . surface reactions can be followed at room temperature by using quench experiments , in which a high - temperature surface is cooled rapidly to freeze in a snapshot of an evolving surface , for example in the growth of iii - v semiconductor surfaces@xcite . however , the quenching process may introduce spurious surface features , and so it is preferable to image the surface _ in - situ _ , i.e. at temperature . the ability to vary the sample temperature makes it possible to image a surface in the process of changing , rather than after it has changed . to give just one example , the phase transition between the ( 1@xmath11 ) reconstruction and the ( 7@xmath17 ) reconstruction on the si(111 ) surface , at around 1100k , has been observed directly in both directions , by careful control of the temperature across the transition@xcite . for useful observations of dynamic processes , the speed of the reaction must be matched to the rather slow imaging rate of an stm . one example is the motion of step edges@xcite . at around 300@xmath3c , step motion can be imaged as the addition or subtraction of 4 atoms between stm images . however , above about 500@xmath3c , the step edge will have moved significantly in the time taken for the stm to scan one line . thus from line to line , the step will be in different places , as it oscillates about a mean position . in stm , therefore , fast - moving steps have a streaky appearance above 500@xmath4c , as may be seen in fig . [ fig : nucleationstm ] . furthermore , mobile species may be invisible at elevated temperatures , either due to the decrease in contrast , or due to their mobility . surface chemical processes , such as the diffusion of si atoms on si(001)@xcite or the decomposition of a molecule@xcite , can be studied at the single - species level , and complete chemical reaction pathways from initial adsorption can be followed through to complete decomposition@xcite . kinetic information such as activation barriers can be extracted , and any metastable intermediate structures can be imaged , giving a unique insight into the reaction process . the ability to control the substrate temperature is particularly important in studies of epitaxial growth , where the kinetic pathways available at different temperatures will determine the surface morphology . in the case of si / si(001 ) and ge / si(001 ) epitaxy , it has been possible to image the surface simultaneously with the deposition of material , and therefore observe directly the nucleation of islands , and the growth of these nuclei into 1d , 2d and 3d islands , along with associated processes , such as the ripening of epitaxial islands , surface stress relief mechanisms , and the transition from island growth to step flow growth modes@xcite , all of which processes occur at different sample temperatures . however , these added abilities bring with them a penalty in sensitivity . there is generally a loss of resolution due to thermal noise above room temperature , and a loss of stability , with thermal drift making it hard to image the same area for extended periods . in many of the high - temperature stm images shown here , the nanolines have a curved appearance . they have no curvature themselves , the appearance comes from the drift during scanning . furthermore , imaging of epitaxial growth necessarily involves imaging a surface where there are many atoms loosely bound to the surface , or in the gas phase being deposited . in this situation , it is very easy for atoms to stick to the stm tip , which can have adverse effects on the imaging , and it is usually preferable to obtain stable imaging conditions rather than achieve the highest possible resolution , which is generally a more unstable situation . for the same reason , scanning tunneling spectroscopy ( sts ) , or current imaging tunneling spectroscopy(cits ) , becomes unfeasible at elevated temperatures . it is generally better to take a series of images of the same area at different bias voltages , giving voltage contrast of an object , although as the feedback loop is active in this situation , the tip - sample distance will vary , and so the information gained is not equivalent to sts / cits . thus these limitations mean that while high - temperature stm brings the unique ability to study evolving surface morphology in real - space , it is not always the best way to study growth surfaces . a combination of high - temperature studies , and room - temperature studies of quenched surfaces , is likely to provide a fuller description of a system . there have been a number of books on modelling techniques published recently@xcite , so this section will feature only a brief description of the techniques used for modelling of semiconductor surfaces , concentrating on electronic structure calculations ( both semi - empirical and _ ab initio _ ) . these techniques retain the quantum mechanics of the electrons while treating the ions classically . the first problem to consider is how to model a semi - infinite piece of material ( that is , a surface ) . the approximation that is used is to consider a small piece of appropriate material with boundary conditions . generally , one of two solutions is used : either a cluster of atoms , with the dangling bonds on the edges terminated in some suitable way ( e.g. hydrogen ) , or a slab of atoms with periodic boundary conditions in two or three directions . a cluster allows a smaller number of atoms to be used than a slab , but has the drawback that it is hard to allow for long - range strain effects ( such as those caused by reconstructions or steps on surfaces ) . a slab approach requires care : the vacuum gap between repeating images must be large enough to prevent interactions , and the slab itself must be sufficiently thick to allow strain relaxation . the slab approach is more commonly used , particularly with the methods described below . once the system to be modelled has been defined , the calculation technique must be chosen . putting aside quantum chemical methods for the sake of brevity ( they scale rather poorly with system size for more details , the reader is directed to one of many excellent books on the subject@xcite ) , we will consider the semi - empirical tight binding ( tb ) technique , and the _ ab initio _ density functional theory ( dft ) technique . tb@xcite postulates a basis set of atom - centred orbitals for the wavefunctions ; the schrdinger equation can then be rewritten as a matrix equation , with elements of the hamiltonian matrix formed from integrals between orbitals on different atoms . these hamiltonian matrix elements are _ fitted _ to either _ ab initio _ calculated data or experimental results . for simplicity , the basis is often assumed to be nearest neighbour only ( with a cutoff defined on the range of interactions ) and orthogonal . once the hamiltonian has been defined , the band energy of the system can be found by diagonalisation of the hamiltonian or other methods . the other energetic terms ( e.g. hartree correction , exchange and ion - ion interactions ) are represented by a repulsive potential ( often a pair potential ) . despite its apparent simplicity , tb is extremely effective , and generally qualitatively accurate , if not approaching quantitative . it has been shown@xcite that this tb formalism can be derived from dft via a set of well - defined approximations , which explains to some extent its success . it is well - suited for a rapid exploration of configuration space ( e.g. possible structures for some new surface feature@xcite ) provided that a suitable parameterisation exists for the bonds between different species . the fitting of parameterisations is non - trivial , and the transferrability of a given parameterisation ( i.e. its accuracy in environments far from those in which the fitting was performed ) is never guaranteed . generally the matrix elements themselves are fitted to a band structure ( or energy levels of a molecule ) , while their scaling with distance is fitted to elastic constants or normal modes of the system . tb retains quantum mechanics ( since the energy is obtained by solving the schdinger equation ) while approximating the most complex problems . dft@xcite starts with an exact reformulation of the quantum mechanics of a system of interacting electrons in an external potential : the result is a set of equations for non - interacting electrons moving in an _ effective _ potential with all the complex electron - electron interactions in a single term , known as the exchange - correlation functional ( it is a function of the charge density , which is itself a function of position , hence `` functional '' ) . unfortunately , the form of this functional is not known , and must be approximated , for instance with the local density approximation ( lda ) or one of the generalised gradient approximations ( gga ) . dft has been extremely successful in many areas of physics , chemistry , materials and , increasingly , biochemistry , particularly when combined with the pseudopotential approximation . the hard nuclear potential and the core electrons of each atom are replaced with a single `` pseudopotential '' , and only the valence electrons are considered , leading to a softer potential . this approximation is most effective for atoms where there is considerable screening of the valence electrons by the core electrons ( e.g. in si the 3s electrons are screened by the 1s and 2s shells , while the 3p electrons are screened by the 2p shell ) ; in first row elements and first row transition metals , this is a much smaller effect , leading to the development of `` ultrasoft '' pseudopotentials . these issues are discussed in more detail in many excellent books and reviews@xcite . the essential point to note is that dft is widely used for electronic structure calculations , and provides results which are accurate to within approximately 0.1ev . an important development in electronic structure techniques over the last ten years is that of linear scaling techniques@xcite . standard techniques for solving for the ground state , whether tb or dft , scale with the cube of the number of atoms in the system ( either because matrix diagonalisation scales with the cube of the matrix size , or because the eigenfunctions spread over the whole simulation cell , leading to cubic scaling when they are orthonormalised ) . this scaling places rather strong restrictions on the sizes of system which can be modelled , even on massively parallel machines : for _ ab initio _ methods , going beyond 1,000 atoms rapidly becomes prohibitive , though somewhat larger systems can be addressed . however , since electronic structure is fundamentally local ( consider bonding as an example ) , the amount of _ information _ in the system should be proportional to the number of _ atoms _ in the system . tight binding methods which take advantage of this ( e.g. refs.@xcite ) have been widely used for some time , allowing calculations on many thousands of atoms . the implementation of linear scaling dft algorithms has proved significantly harder , though recent efforts in nearly linear@xcite and linear scaling methods@xcite suggest that these techniques are starting to produce useful and general results . one of the challenges of using stm to examine semiconductor surfaces is that the current arises from both geometric and electronic structure , though this is less true at high biases ( where height or geometric structure will dominate ) . some technique is required to understand these changes , and to test proposed structures against experiment . the field of stm simulation is a complex one ( in part because the structure and composition of the tip is unknown ) ; the interested reader is referred to excellent reviews for further information@xcite . we will briefly summarise the simplest and most common approximation used , and discuss how the experimental - theoretical interaction can best proceed . the most commonly used approach is the tersoff - hamann approximation@xcite . this asserts ( via a careful series of approximations ) that the tunneling current is proportional to the local density of states ( ldos ) due to the sample at the position of the tip . in effect , the tip is assumed to have a flat density of states . this approach is directly equivalent to considering the projected charge density ( i.e. the partial charge density due to each band for bands within @xmath5 of the fermi level ) . it is qualitatively accurate , though will not reproduce the observed corrugation values correctly@xcite . the authors are of the opinion that for true success in investigating nanowires ( and other systems ) on semiconductor surfaces , a close collaboration between experiment and modelling is required . any successful collaboration depends on a number of factors , but the key factors , we believe , are frequent , easy correspondence , an understanding of the limitations and abilities of the appropriate techniques , and trust between the different sides in the collaboration . the last point extends to sharing of unpublished data , and extensive discussion of possible courses of action . the three authors have been working together in different combinations for ten years ( leading to around 20 joint publications ) , and we find that our collaboration is still immensely fruitful . while the bi / si(001 ) nanolines have remarkable structural qualities , they are not the only self - assembled nanowire system on semiconductor surfaces . in this section we present an overview of these other nanowire systems ( more details are presented elsewhere@xcite ) , and we will consider in more detail the most important of these systems , the rare - earth silicide family , focussing on their structure and formation , electronic properties and reactivity . a variety of metals have been found to self - assemble into nanowires or nanolines on the si(001 ) surface , though none with the perfection of the bi nanolines . the simplest way in which to form 1d structures on the si(001 ) surface is by epitaxial growth . an island nucleus is essentially an ad - dimer , which sits between the dimer rows . the two pairs of si dimers which support the ad - dimer are distorted by its presence , and so are attractive adsorption sites for further ad - atoms or ad - dimers@xcite . this process has been described as a surface polymerisation reaction@xcite . the long sides of the string , which are equivalent to a - type steps on this surface , have a very low sticking coefficient which keeps the island from broadening at lower growth temperatures . deposition of group iii and group iv elements on si(001 ) will therefore result in long 1d chains of dimers@xcite . however , there is little control over the length or structure of these wires , and they are unstable against annealing , eventually reorganising into compact islands . the pt / ge(001 ) system@xcite forms nanowires through a complex surface reconstruction , resulting in a system with a high degree of perfection , which approaches that of the bi nanolines . when annealed to high temperatures , arrays of atomically perfect nanowires of pt / ge form , each of which is 0.4 nm wide with a spacing of 1.6 nm between the wires . sideways growth of two - dimensional islands can also be blocked by introduction of missing dimer trenches . these form to relieve surface stress , either because of contamination by a small amount of a transition metal such as ni , or through heteroepitaxial growth of another material such as ge . the trenches tend to line up in a semiregular array , giving an approximate ( 2@xmath1n ) reconstruction in leed , with n@xmath68 - 12 ( though the regularity of the trenches is rather poor ) . this method has been used to deposit a variety of metals , such as fe@xcite , ga@xcite and in@xcite ( using the ni - based technique ) and the molecule styrene@xcite . the result is long - range , reasonably well ordered wires of the metals , though their properties have not been characterised . the 2d symmetry of the surface can also be broken deliberately by the use of vicinal substrates to produce regular arrays of steps@xcite which can then be decorated with materials such as gold@xcite and bi@xcite . this approach is used primarily on si(111 ) , although with an extreme orientation towards ( 112 ) directions the surface might be labelled si(557 ) or si(5 5 12 ) depending on the angle . the resulting step reconstructions comprise chains which are metallic and one - dimensional ; early observations suggested possible observations of luttinger liquid behaviour@xcite , though this has been disputed@xcite . their spacing can be controlled to some extent by varying the angle of miscut of the surface . as an atomic - scale extension of the idea of lithography , the si(001 ) surface has been passivated with atomic hydrogen ( forming the monohydride phase ) and individual hydrogen atoms removed with an stm tip to form atomic - scale patterns . the wire formed by removal of hydrogen atoms along ( or across ) a dimer row is known as a `` dangling bond '' wire ( or db wire)@xcite and is predicted to show conduction effects similar to conjugated polymers with polaronic and solitonic effects@xcite . an example of a dangling - bond wire is shown in fig . [ fig : hashizume](a ) . these wires have been reacted with a variety of adsorbates including iron@xcite , gallium@xcite , aluminium@xcite , silver@xcite and organic molecules such as norbornadiene@xcite . however , although for a single row of dangling bonds , a perfect atomic wire can be formed , as with ga in fig . [ fig : hashizume](b ) , for greater widths , as in fig . [ fig : hashizume](c , d ) , the dangling bond wire is more ragged , and the resulting wires ( here ag ) are typically composed of a large density of small , roughly spherical crystals , which exhibit many imperfections and boundaries . a more elegant implementation stems from careful application of molecular chemistry to produce a directed reaction : when the monohydride surface was exposed to styrene and _ one _ hydrogen atom removed , self - directed growth of lines of styrene resulted from a chain reaction between each adsorbed molecule and an adjacent hydrogen@xcite . such self - organised methods are important for the growth of molecular nanowires on surfaces . however , for all these systems there is a lack of control over the size and length of the nanowires . single nws grow in both @xmath7110@xmath8 directions , blocking each other s growth . finally , the nws can form bunches . both the single nws and the bunches can grow a second layer of material , unlike the bi nanolines . _ _ image courtesy prof . j.nogami , u. toronto , canada . _ _ ] a more promising materials system for nanowire growth on si(001 ) is that of the rare - earth metal silicides . there has been extensive research on these systems because of their good conductivity and low schottky barrier with silicon@xcite , though this work concentrated on the si(111 ) surface . moreover , in the hexagonal phase some rare - earth silicides show a good lattice match with the si(001 ) substrate in one direction , with a large mismatch in the other . heteroepitaxial growth of these materials would therefore be expected to be constrained in the high mismatch direction , and facile in the other , resulting in long 1d islands . this is indeed what happens . the behaviour of the rare - earth nanowire systems on si(001 ) has been investigated extensively , and nanowires have been observed under certain growth conditions . as the other major family of self - assembled nanowires on the si(001 ) surface , we give a fuller description of this family for comparison to the bi nanolines . .lattice mismatch between hexagonal phase of various rare - earth silicides and si(001 ) . data is presented in terms of lattice constant and percentage mismatch ( the si(001 ) surface has a lattice constant of 3.84 ) . where the indication `` hexagonal '' is given , it indicates that other phases are possible . _ courtesy of dr c. ohbuchi , nims , japan._. [ cols="<,<,<,<,<",options="header " , ] the most common form of rare - earth nanowire ( renw ) formed on si(001 ) is thought to result from the anisotropic strain between the alb@xmath9 crystal structure of the nanowire and the si(001 ) substrate , leading to fast growth along the less - strained direction , and extremely limited growth along the more - strained direction . the mismatch for various silicides in the alb@xmath9 hexagonal structure is given in table [ tab : rare - earth - mismatch ] . renws of this type have been reported for er@xcite , dy@xcite , gd@xcite and ho@xcite . for sm@xcite nws only appear on vicinal substrates ; other reports@xcite find that on flat surfaces only rectangular , 3d islands are formed after a ( 2@xmath13 ) layer ) . bundles of nws have been reported for yb@xcite ; again , early reports@xcite suggested that nws were not formed , and 3d islands resulted from annealing . though they are not rare - earth metals , similar nws have been reported for both sc@xcite and y@xcite , resulting from the same combination of crystal structure and anisotropic strain . although it forms a similar wetting layer on the substrate to yb , eu does not form nws@xcite . the resulting renws are found both individually and in bundles ; the individual wires have sizes which vary from element to element but are in general 5 - 10 nm wide and less than 1 nm high . the wires grow extremely fast , reaching lengths of up to 1@xmath2 m ; more interestingly they always run along @xmath10 directions and will grow out over step edges , drawing terraces with them , features that they share with bi nanolines . the large - scale image in fig . [ fig : silicide1 ] is very similar to that of bi nanolines at the same scale . for almost all materials , the surface of the wires shows a c(2@xmath12 ) reconstruction@xcite , though the yb nws , which are off - axis relative to the other renws , may have a ( 1@xmath11 ) surface@xcite . the si(111 ) surface matches the lattice of the re alb@xmath9 structure reasonably well in all directions ( with a mismatch of less than 2% for most metals)@xcite and so would not be expected to give anisotropic growth , but gdsi@xmath9 nws have been formed along step edges of a vicinal surface@xcite . these form because of a mismatch perpendicular to the step edge , and grow to over 1@xmath2 m long on appropriately prepared samples , with a width of 10 nm and height of 0.6 nm . other nw structures have been reported for transition metals on si(111 ) , for instance fe@xcite ( though these are rather short wires ) , ni@xcite and ti@xcite . the growth mechanism for these wires is not entirely clear : the resulting nws are 5 nm high and 5 nm wide ( nisi@xmath9 ) , 10 nm high and 40 nm wide ( tisi@xmath9 ) , though for both of these systems the nws appear to grow _ into _ the substrate somewhat . there is a technique for growing nws on si(001 ) , si(110 ) and si(111 ) explicitly relying on growth into the substrate , resulting in `` endotaxial '' nws , which has been applied to dy@xcite in the hexagonal alb@xmath9 structure , as well as co@xcite , which takes the caf@xmath9 structure . the driving force appears to be kinetic in these systems : the long direction of the nw has an interface which grows faster than the short direction , though this is partly dependent on the structure of the interface below the surface . the growth and formation of the renws is not understood at an atomic level : they form extremely quickly ( on experimental timescales ) so that , for instance , hot stm can not be used to observe formation of the lines . unlike the bi nanolines , which have a constant width , and do not change in width with annealing , the rare - earth nanoline family can take a variety of widths depending on the growth conditions . in the case of dysi@xmath11 , single nws of a consistent width form from the underlying ( 2@xmath17 ) reconstructed wetting layer , as in fig . [ fig : silicide2 ] . however , further annealing results in the formation of bunches of nanowires , which join together where they meet , and multilayer islands form . as strained coherent heteroepitaxial islands , the renws are only metastable ; prolonged annealing will cause them to coarsen , and form large 3d islands . overall , the deposition / anneal temperature must be in the window between 550@xmath4c and 650@xmath4c , and annealing time should not exceed 2 - 10 minutes . below 550@xmath4c or 2 minutes , the reaction will be incomplete and the nws formed will be immature . above 650@xmath4c or 10 minutes the systems tend towards lower energy , thermodynamically stable states . in particular , dy - si(001 ) will form 3d islands if annealed for long times ( e.g. 30 minutes for 0.86ml)@xcite with wires formed for shorter annealing periods . annealing er - si(001 ) initially results in nws , but as the annealing time is increased dislocations form in the nws@xcite . if the temperature is raised beyond 620@xmath4c the same behaviour is seen , leading ultimately to coarsening into islands@xcite . a large - scale peem study of ersi@xmath11@xcite found that large nws ( which may have been bundles of nanowires the technique does not have the resolution to distinguish ) were not affected by annealing on a coarse scale . however , gd shows more complex behaviour . annealing studies of thin films of gd found that both hexagonal and orthorhombic phases could form@xcite , and that with longer annealing times the orthorhombic phase grew at the expense of the hexagonal phase . long anneals of the gd / si(001 ) system , up to 620@xmath4c for one hour , resulted in the formation of previously unknown silicide structures , which are aligned perpendicular to hexagonal silicide nws on the same terrace@xcite . there are various useful observations of their behaviour which can be made : * a wetting layer with a characteristic reconstruction forms before the nanowires , and persists on the substrate with the nanowires ; specific reconstructions have been observed for : * * eu - si(001):(2@xmath13)@xcite , * * sm - si(001):(2@xmath13)@xcite , * * yb - si(001):(2@xmath13)@xcite and ( 2@xmath14)@xcite , * * nd - si(001):(2@xmath13 ) and ( 2@xmath14)@xcite , * * dy - si(001):(2@xmath14)@xcite , and ( 2@xmath15 ) and ( 2@xmath17)@xcite , * * gd - si(001):(2@xmath14)@xcite , and ( 2@xmath15 ) and ( 2@xmath17)@xcite * * ho - si(001):(2@xmath14 ) and ( 2@xmath17)@xcite * the only system for which this does not happen is er - si(001)@xcite . * the system must be heated during deposition or annealed post - deposition to allow reaction of the rare - earth metal with the substrate@xcite * annealing the system for too long , or depositing at too high a temperature , can result in 3d islands rather than nanowires ; this is discussed in more detail below . as different metals have different properties for each of these categories , we will discuss them briefly in turn . gd and dy@xcite form a ( 2@xmath14 ) reconstruction first , followed by a ( 2@xmath17 ) reconstruction@xcite just prior to nanoline formation . rather similar behaviour is seem for ho@xcite . the metal fraction in the two reconstructions is 3/8 and 5/14 ml respectively@xcite . for these three metals , the presence of the wetting layer is intimately connected with formation and stability of the nanowires . despite the wealth of studies of ersi@xmath9 nanowires , there are no reports of the structure and make - up of a wetting layer for this system ; there is some evidence@xcite that once more than @xmath60.05ml of er is deposited , long chains of surface dimers form , followed by nanowires , with no intermediate reconstruction . indeed , there are indications@xcite that the surrounding substrate is clean si(001 ) . another group of metals form closely related reconstructions : eu@xcite , yb@xcite , nd@xcite and sm@xcite all form ( 2@xmath13 ) and in some cases ( 2@xmath14 ) reconstructions ; by contrast to er , dy , gd and ho , these metals are strained in both surface directions on contact with si(001 ) ( which might be expected to reduce the periodicity of any substrate wetting ) and lack the main driving force seen before for nanowire formation . in general these metals do not form nanowires@xcite though there are certain conditions where yb can be made to form nanowires@xcite of some kind , which may well be grown off - axis ( in a different growth mode to the other silicide nanowires which grow with the c - axis aligned with the substrate ) . it is also possible to induce nanowire formation in situations where they would not normally form , either because of uniform lattice matching ( e.g on si(111 ) surfaces , as mentioned above ) or lattice mismatch in _ both _ surface directions using step edges . this technique has been successfully employed for sm on si(001)@xcite and gd@xcite and dy@xcite on si(111 ) . one of the original technological interests in rare - earth overlayers on si(111 ) was the good crystal growth possible due to the interface between the silicon and the overlayer , and the conductivity properties . the films have good conductivity and a low schottky barrier ( @xmath60.4 ev on n - si and @xmath60.8 ev on p - si@xcite ) . measurement of the conductivity of renws is a significant challenge . full two - point or four - point measurements requires either a unique stm instrument ( for instance one used to measure conductivity of cosi@xmath9 nanowires on si(110)@xcite ) , or alternatively the formation of nanoscale contacts to a nw@xcite . while significant progress has been made in this area recently , any measurements of conductivity will include the resistance of the interface between the nw and the contacts . scanning tunneling spectroscopy ( sts ) has been performed on various nws : dy and ho@xcite and gd@xcite . while sts does not measure the conductivity _ along _ the nw , it does measure the local electronic structure , and in all cases the nws are found to be metallic , while the surrounding reconstructed substrate is not@xcite . this is good evidence that the nws are taking on the bulk silicide structure , which is conducting , and the wetting layer is an intermediate state which relieves strain . conductivity measurements of transition metal silicides on si(111 ) and si(110 ) have been made@xcite . these studies exemplify the two techniques for measuring nw conductivity : nisi@xmath9 nws were contacted by gold pads@xcite , while four - probe stm measurements were made on cosi@xmath9 nws@xcite . these nws are relatively large compared to the renws discussed so far : 15 nm ( ni ) and 60 nm ( co ) wide . the cosi@xmath9 nws showed a high schottky barrier and conductivity equivalent to that of high quality thin films of cosi@xmath9 , while the nisi@xmath9 nws showed some signatures of quantum transport , though the conductivity of these wires was significantly lower than thin films ; this is likely due to the overgrowth of the sample with sio@xmath9 for transfer to the lithography apparatus . there is very little data on the chemical reactivity of the rare - earth nanowires . thin films on si(111 ) are susceptible to reaction with o@xcite ; the renws oxidise rapidly in air@xcite , and transition metal nws show signs of this ( confirmed by the conductivity of nisi@xmath9 nws on si(111)@xcite ) . in the tm system , the nws were overgrown with native oxide , and showed a significant decrease in conductivity relative to complete thin films of similar thickness , which is attributed to scattering at the nw surface / oxide interface . further information on the reactivity comes from the deposition of pt on a surface containing ersi@xmath9 nws on si(001)@xcite . the pt was deposited at room temperature and the sample was subsequently annealed . when stm images were taken they showed that the c(2@xmath12 ) surface reconstruction on the nws was no longer visible though the ( 2@xmath11 ) substrate reconstruction was still present . furthermore , the pt - covered nws were resistant to reactive ion etching and appeared stable when exposed to air for periods of up to 8 weeks ; the pt overlayer does appear to strain the nws , however , leading to delamination of the nws from the substrate if left untreated . clearly there is much work to be done understanding the reactivity and stability of these nws . the discovery that nanowires would form if a bi - covered si(001 ) surface was left to anneal around the bi desorption temperature was made quite by chance . in 1995 , in the materials department of oxford university , two of the authors were investigating the surfactant - assisted growth of ge / si(001 ) using bi as a surfactant , and were studying the properties of bi / si(001 ) . when a sample which had been left to anneal at _ ca . _ 500@xmath3c overnight was imaged , very large , flat terraces , and long , straight , bright lines ( longer than the scanning range of the stm ) , running across the surface were found . these were the bi nanolines . the first published image of the `` nanobelts '' was in 1997@xcite , and the first big study of their growth and properties came in 1999@xcite . in this section , we will discuss the atomic structure of the bi nanoline , and its physical and electronic properties . we show that many of the properties stem from its unusual structure . the reaction of the bi dimers with a variety of reagents , and the effect of burial of the nanoline will also be discussed . the physical structure is the key to understanding the properties of a nanoline system , and great effort has been devoted to identifying the structure of the bi nanoline . in this case , the structure was identified by a synergy of stm observations with tightbinding and dft simulations . the three different proposed models for the bi nanoline are shown in fig . [ fig : models ] . the two early models ( shown as the top two models ) share certain structural motifs . they contain a pair of bi dimers set into the surface layer of the si(001 ) crystal , whose compressive stress is relieved by missing dimer defects ( dvs ) . the first model proposed , shown in fig . [ fig : models](a ) , was based around a pair of bi dimers in the surface layer , separated by a missing dimer defect , so as to relieve the local stress of the bi dimers@xcite . this became known as the miki model . the second proposed model , shown in fig . [ fig : models](b ) , was based around a pair of bi dimers with defects on either side@xcite , which became known as the naitoh model . however , the third model is more complex : a reconstruction of several layers of si underneath the bi dimers produces the nanoline core . this was named the haiku model@xcite . of the three models , this structure is the only one to fit all the criteria which have been determined from stm observations , as well as having the lowest energy of the three . in this section , we will briefly review the historical process by which this structure was determined , and discuss the properties of the haiku structure . early stm images of the bi nanoline were taken at elevated temperature@xcite , in which the si dimers were not resolved , and hence the registry of the nanoline with the substrate was unknown . this data suggested that the nanoline contained two features , which were probably bi dimers , with a spacing of approximately 6.3 , with a total width of about 1 nm , which was approximately equivalent to the space of three si dimers . as the nanolines appeared bright at large bias voltages ( in both positive and negative bias images ) to dark at low voltages , it was also reasonable to assume that these bi dimers were situated in the surface layer , rather than in an adlayer or in a subsurface layer . it was thought that bi dimers embedded in the top surface layer would have considerable compressive stress , and on this basis , a model based around a 1dv , with bi dimers to either side , was proposed . this was the miki model@xcite . tightbinding calculations of the miki model@xcite found that it was more stable than the ( 2@xmath11 ) or ( 2@xmath1n ) bi reconstructions , and that the formation energy of a defect in the line was high : around 1.1 ev@xcite ( this number considers putting the bi which has been removed as an ad - dimer on the surface ; recent data@xcite shows that the defect energy falls to 0.11 ev if the bi dimer is placed in another miki model ; this point is discussed more fully in section [ sec : miki - model - revisited ] below ) . dft calculations@xcite agreed with those conclusions . the calculated ldos showed that the bi dimer states were further away from the fermi level than those of the si dimers and simulated stm images of the miki structure@xcite showed that the line would appear dark at low bias voltages , in agreement with stm results@xcite . testing by other methods appeared to confirm this structure . photoemission spectroscopy experiments@xcite found that the bi 5d core - level spectra of the bi nanowire was essentially identical to the spectra of the ( 2@xmath1n ) phase composed of bi ad - dimers . this suggested that the local chemical state and registry of bi adsorbates for both phases was the same , i.e. that the bi was in the form of dimers in the top layer of the structure . x - ray photo - electron diffraction ( xpd ) experiments@xcite found a good fit between the experimental xpd data and simulated xpd intensity peaks from the miki model . in particular , they confirmed the presence of bi dimers parallel to the si dimers and found the spacing between them to be @xmath66.3 , in agreement with the stm measurement , though it is important to note that the quoted distances are the result of a _ fit _ to the miki model . 10 nm stm image of the si(001):h surface . the h termination makes it easier to resolve the individual si dimers . a series of markers has been placed across the bi nanoline . from this , it can be determined that the nanoline occupies the space of _ four _ dimers . ] however , it was difficult to account for the extreme straightness of the nanoline with the miki model , as the calculated kinking energy was very small , around 0.1 ev . a kinetic argument was put forward , which suggested that there was a stronger preference for incoming bi dimers to line up at the end of the nanoline , and hence the nanoline would grow straight@xcite . furthermore , the diffusion constant of a bi surface dimer was high ( this diffusion was required for a kink to form in a miki model nanoline ) , and hence a line would remain straight once it had grown . despite this explanation , the issue of the straightness of the bi nanolines remained as a question mark over the miki model . subsequent room - temperature stm experiments@xcite proved decisively that the miki model could not be the structure of the bi nanoline , as it had the wrong registry with the surface . markers were placed on the surface of an image with the si dimers resolved , as shown in fig . [ fig : registry ] , and by counting across the nanoline , the width was determined definitively to be four dimers , and not three as in the miki model ; moreover the bi dimers lay _ between _ the si dimers of the substrate , not in the same position . ( this registry has been confirmed by every other stm measurement made@xcite ) . on the basis of the new stm data , the naitoh model was proposed@xcite . again this model was tested by atomistic calculations ; a detailed study of the dimensions and simulated stm appearance@xcite showed that again the nanoline would appear dark in stm at low bias voltages , but showed also that the spacing of the bi dimers was too narrow , approximately 5 , which could not be reconciled with the spacing of 6.3 measured from stm . moreover , the energy was considerably worse than the miki model@xcite , and the kinking energy was still low . there was therefore no satisfactory structural model at this time . inspiration for the structural model of the bi nanoline arose from the structural model proposed for the b - type double - height step of as - terminated ge@xcite . at this step edge , bond rotation forms a pair of 5 and 7-membered rings of ge , capped with as , shown in fig . [ fig : haikucartoon](a ) . while the haiku structure appears to be a complex reconstruction , in fact the construction from a cell containing two bi ad - dimers can be described schematically in two simple steps , as shown in fig . [ fig : haikucartoon](b)-(d ) . the first step is to rotate the second and third layer atoms underneath the bi ad - dimers , so that the si dimer atoms sink down into the surface . this creates the 5- and 7-membered rings also seen in the as / ge step structure . the central four atoms are then removed , and the two halves of the structure bonded together , thus creating the haiku structure . ( n.b . this is not proposed as the formation mechanism , but is simply given as a means of understanding the substructure . ) a linescan of the nanoline has been matched to the haiku model , using the background si dimers as reference marks@xcite . the positions of the bi dimers match up extremely well to the peaks in the linescan , giving strong confirmation of this model . other features of experimental linescans also agree well with simulated linescans for the haiku model@xcite . dft calculations of this new structure@xcite found that it was considerably more stable than the miki model , which has been confirmed by subsequent modelling@xcite . although the haiku structure shares several structural features with previous models the top of the structure has two bi dimers joined by rebonded second - layer si atoms , as in the miki model , while on the outside of the bi dimers , there are more rebonded si atoms , just as in the naitoh model the difference lies in the si substructure ; this mixture of 5 and 7-membered rings extends down many layers . consideration of the core of the nanoline reveals that it is a small triangular section of hexagonal silicon , embedded in the diamond cubic silicon substrate , rather like the `` endotaxial '' re nanowires mentioned in section [ sec : formation - structure ] . the \{111 } planes that delineate the core are shown in red in fig . [ fig : haikucore ] . the hexagonal core of the nanoline does not exist as a bulk structure for silicon ; it is particular to this nanoline structure . in this respect , the bi nanoline is quite unlike the epitaxial rare - earth silicide family , which are essentially one bulk crystal grown epitaxially onto another bulk crystal , taking their shape from the highly anisotropic mismatch between the two . this core structure is responsible for many of the properties of the bi nanoline . it will be difficult or impossible to kink , resulting in the extremely straight nanoline observed . as will be shown below in section [ sec : dez - sub ] , the tensile strain field results in the repulsion of other nanolines , step edges and missing dimer defects . as a result , the nanolines will not grow sideways , will not grow together , and will not coarsen into larger islands , unlike the silicide wires . the electronic structure of the haiku model has been studied in some detail@xcite , and again the nanoline becomes dark in low - bias images this effect is a property of the bi dimers rather than that of the nanoline . low - bias stm images reveal other electronic effects , which provide further confirmation of the haiku structure , while ruling out the miki structure . these features are discussed below in section [ sec : electronic - structure ] . however , despite the wealth of agreement of the predictions from the haiku model with experimental evidence , there has been no direct confirmation of the haiku structure . the bi nanolines are destroyed by exposure to air , while burial of the nanolines in a capping layer also changes their atomic structure@xcite ( though if done carefully the 1d character and dimerisation of the bi can be preserved ) , so that cross - sectional tem studies@xcite have not provided definitive confirmation of the core structure , and a recent x - ray standing wave(xsw ) study@xcite which relied upon capping with amorphous silicon and whose results cast doubt on the haiku model , can not be regarded as a counterindication . the difficulties of burial of the bi nanoline are discussed in section [ sec : burial ] . we note also that a recent suggestion@xcite , that the height of the bi dimers in the haiku model is too far above the plane of the surface si dimers based on xpd@xcite , is not a significant piece of evidence against the haiku model : the xpd data was _ fitted _ to the miki model , thus naturally giving the wrong data for the haiku . if a fit of the data to the haiku model were to give height data which contradicted the atomistic modelling , then the model would have to be revisited . despite the apparent success of the haiku model , the miki model has continued to attract a large amount of attention@xcite as a possible candidate for the nanoline structure . while in fact its incorrect registry with the si substrate unambiguously rules it out as a candidate nanoline structure@xcite , it is an energetically favourable structure , and might be expected to be present on the bi - covered si(001 ) surface . with the benefit of hindsight , it can be seen that the miki structure was indeed observed in early stm images of the bi - rich surface . [ fig : naitohmiki ] shows that annealing a bi - rich surface at 400@xmath3c produces a large density of quite straight dimer vacancy ( dv ) trenches . in the filled - states image , [ fig : naitohmiki](a ) , these appear to be composed of clean silicon . however . in the empty - states image , [ fig : naitohmiki](b ) , there is contrast between the dimers adjacent to the missing dimer trenches and the rest of the surface . this contrast does not occur on the clean surface at this bias voltage , and so it can be inferred that the trenches are decorated by bi dimers , i.e. the miki structure has been formed . as with the ge - induced dv trenches mentioned in sec . [ sec : other - nanowires ] , there is cooperative strain relief between a bi dimer and a 1dv . bi dimer / dv structures have the lowest energy / bi dimer of any structures except the bi nanolines . there is a kinking energy for these ( 2@xmath1n ) trenches of around 0.1ev@xcite . hence the formation of a semi - regular array of bi dimers reduces the energy further . c , followed by 10 min . annealing . ( a ) : -2.2 v,0.3 na . there are many straight missing dimer trenches . ( b ) : + 2 v , 0.3 na . brightening around the trenches in empty - states images indicates that they are decorated by bi dimers , i.e. the miki structure . ( c ) : a bi - decorated trench and bi nanoline on a h - terminated surface . a linescan along the dotted line ( d ) reveals that the bi - decorated trenches are significantly lower than the nanoline . ( a ) & ( b ) are reprinted from appl . 142 , naitoh et al . bismuth - induced surface structure of si ( 100 ) studied by scanning tunneling microscopy , p.38 , ( 1999 ) , with permission from elsevier.@xcite ] further proof of the bi decoration of 1dv trenches has been provided by a more recent experiment , in which a bi - rich surface was quenched from 550 @xmath3c at an early stage of the annealing process@xcite . here atomic hydrogen was adsorbed after cooling to enhance the contrast between bi and si significantly , as the h will adsorb easily to the si dimers , but does not adsorb on bi@xcite , as can be seen in fig . [ fig : registry ] . it was found that the h termination increased the relative _ apparent _ height of bi - related features by ca.150 pm , or 1.5 . an example image is shown in fig . [ fig : naitohmiki](c ) . a linescan across both a bi - decorated trench and a section of nanoline shows that the bi - decorated trench is quite similar to the nanoline , but is lower and has a deeper depression between the two bi dimers . stm linescans of the feature decorating dv trenches and the bi nanolines @xcite compare very well to simulated scans of the miki model and haiku model respectively@xcite , providing strong support for both these identifications . furthermore , in this way , it has been shown that the miki structure and the bi nanoline co - exist , demonstrating once again that the miki structure is not the nanoline , as has been recently suggested@xcite , but it is kinetically stable , even at high temperatures , due to the large activation barrier necessary to form the haiku structure . while rare - earth silicide wires have quite specific growth recipes in order that the nanowires do not coarsen , the bi nanolines are robust to a wide variation in the growth conditions . the essence of bi nanoline formation is that it is a competition between deposition and evaporation . a thick layer of bi on si(001 ) is only stable below the bi bulk evaporation temperature ( ca . 400@xmath3c ) . above this temperature , an epitaxial layer of 1 - 2 ml bi is stable , and annealing of this surface will result in a surface with bi nanolines co - existing with these bi islands , as shown in fig [ fig : bi2xn ] . the first - layer bi forms a ( 2@xmath1n ) structure , with n=4 or 5@xcite , and the second - layer bi forms as small groups of dimers , which can be either parallel or perpendicular to the underlying bi . c , and then annealed for 60 mins at 497@xmath3c , resulting in bi nanolines surrounded by a bi overlayer , which forms a ( 2@xmath1n ) reconstruction , with n=4 or 5 , to accomodate its large strain . left - hand images are empty - states , right - hand images are filled - states . the ( 2@xmath1n):bi does not grow over the bi nanoline at this temperature . ] the threshold temperature at which bi dimers on si(001 ) will start to evaporate is around 500@xmath3c , while the maximum temperature at which any bi is stable on the si(001 ) surface is around 600@xmath3c . deposition of bi in this temperature window , or deposition of bi at a lower temperature , followed by an anneal within this temperature window , will result in a surface comprising bi nanolines on an otherwise clean surface , as in fig . [ fig : recipes ] . however , the details of the growth recipe will determine the surface morphology at the end of the anneal , as is described below in sec . [ sec : recipes - sub ] . of particular technological interest is the growth of majority - domain surfaces after long anneals . the si(001 ) surface has two equivalent domains with the dimer rows running in orthogonal directions . on a typical si(001 ) surface , these two domains will have roughly equivalent total areas . single - domain surfaces form on vicinal wafers miscut along a @xmath0 direction , where the dimer rows tend to run perpendicular to the step edges . however , by long anneals of the bi - rich surface towards the high end of the temperature window , majority - domain surfaces can be formed , even on flat surfaces . in this case , extremely long bi nanolines grow , up to 1@xmath2 m , as can be seen in fig . [ fig : mcleanimage ] . c. ( b ) : deposition of 1 ml bi at around 550@xmath3c , followed by annealing at the same temperature . ( c ) : continuous exposure of 0.1 ml / min bi flux for 40 mins at 570@xmath3c . ( d ) : as ( c ) , but on a 0.5@xmath3 miscut sample . the growth recipes are summarised schematically in fig.[fig : recipesplan ] . ] high - temperature stm observations of the annealing bi - rich surface shows a consistent series of surface morphologies . immediately after the end of bi deposition , a relatively rough surface is obtained , with a high density of small islands , composed of a mixture of bi and si , as shown in fig . [ fig : recipes](a ) . these will quickly disappear , but at the same time , nanolines begin to form , either on the terraces , or on one of the mixed islands . the nanolines grow out over the step edges in long , finger - like islands known as promontories , with a bi nanoline surrounded by a thin strip of silicon . examples of the growth of promontories are seen in fig . [ fig : promontory ] , and more mature examples of this type of surface are shown in figs . [ fig : recipes]&[fig : mcleanimage ] . the nanolines will not only form long , narrow promontories , but also the reverse , a deep inlet where a nanoline in one terrance has grown through an up - step into an upper terrace . the surface between the bi nanolines contains a large density of mobile missing - dimer defects , apparent in stm as dark , fast - moving streaks . after further annealing , the missing dimer trenches begin to disappear , and the bi nanolines remain on a very flat , featureless surface . finally , the bi nanolines themselves evaporate . in the following sections , we will give describe the surface processes which drive this sequence of events , and account for these observations . the interactions of the bi nanoline and its associated strain field with defects and step edges in the terraces , and with each other , will be discussed and the possible nucleation mechanisms of this complex structure explored . as shown above , annealing a bi - rich surface at lower temperatures results in a surface containing a high density of missing dimer trenches , decorated with bi dimers , i.e. the miki structure . one feature of the bi nanoline is the high temperature required to form it , implying a significant activation barrier to its nucleation . while the large number of growth parameters , such as flux , total amount deposited , substrate temperature , miscut angle and azimuth , provides a wide range of different possible recipes , as summarised in fig . [ fig : recipesplan ] , there are essentially two routes to the growth of nanolines@xcite . one growth method for the nanolines is to deposit a monolayer of bi below this threshold temperature , and then anneal the sample until bi nanolines form . after around 1 hour at around 500@xmath3c , this will result in a surface with bi nanolines and a background bi - covered surface@xcite . continued annealing of this surface will result in the desorption of the background bi forming a surface similar to that shown in fig.[fig : recipes](a ) , and eventually a surface similar to that shown in fig . [ fig : recipes](b ) is obtained . the second growth preparation route involves the deposition of bi onto the surface within the desorption temperature window . due to the lower surface coverage of bi , and the higher temperature , a more ordered surface is obtained , and fewer , longer , bi nanolines form . the final surface is much the same , however , corresponding to that shown in fig.[fig : recipes](b ) . a high density of bi nanolines can be built up by continuous bi exposure for a long time , at a temperature in the upper half of the desorption window . in this regime , the nanolines are only stable for a few minutes in the absence of a bi flux and small differences in stability , such as the ends of a nanoline , will be significant . shorter nanolines , perpendicular to the prevailing direction , will tend to dissolve , so that a highly ordered surface with long , parallel nanolines is obtained , as in fig.[fig : recipes](c ) . by use of a vicinal surface , the preference for one domain can be enhanced further , and a surface close to a single - domain surface can be obtained , as in fig.[fig : recipes](d ) . a second method which produces a majority - domain surface , even on a surface without any intentional miscut , ( the miscut angle is @xmath12 ) @xcite , produces a surface as shown in fig . [ fig : mcleanimage ] . in this case , a substrate temperature around 530@xmath3c was used , 4.5 ml of bi were deposited , with an anneal for around 40 mins . comparison of fig.[fig : recipes](d ) and fig . [ fig : mcleanimage ] reveals that in the former case , the nanolines are parallel to the prevailing step edges , while in the latter , they are perpendicular . despite the large difference in morphology , the only differences between this recipe and that which produced fig.[fig : recipes](c , d ) are that continuous deposition for an extended period was not used , and the heating current used was ac , a strategy which was designed to reduce electromigration . whether this is is the cause of the majority - domain surface remains unclear , but such large - scale single - domain surfaces are ideal for nanoelectronics applications . c , a ml of bi is stable on the surface , and the ( 2@xmath15):bi reconstruction results . annealing this surface above 500@xmath3c , or deposition of bi onto the clean surface above this threshold results in bi nanolines on a clean background . high coverages of nanolines are obtained by higher - temperature anneals , with longer bi exposure times . ] while stm can observe only a relatively small part of the surface , the large - scale order of the surface can be determined using electron diffraction methods such as rheed . a series of rheed patterns at different azimuthal angles of a single - domain bi nanoline surface are shown in fig . [ fig : rheed ] . perpendicular to the nanolines , the rheed pattern comprises vertical lines , while parallel to the nanolines , a ring feature is seen . the strength of these features demonstrates the long - range order generated by the nanolines , and means that their growth and development can be followed in real - time . 0 ] direction , perpendicular to the nanolines , a series of long streaks is seen , as in ( a ) . along the [ 110 ] direction , parallel to the nanoline , as in ( c ) , a distinctive ring feature is seen , while at intermediate angles ( b ) , the ring opens up into a long arc . ] in contrast to the silicide nanowires , which form very rapidly at the growth temperature and are only stable for about 10 mins . at that temperature , there is a significant incubation period before the nucleation of the bi nanolines , and they will continue to grow for upwards of an hour after the deposition of bi has finished . thus the bi required to generate the nanolines must come from a reservoir of bi which is present on the surface . where is this bi located , and how does it affect the growth of the nanolines ? in images of the surface taken soon after the end of the bi deposition , for example , fig.[fig : recipes](a ) , there is often a high density of small islands on the surface . these islands have the same appearance as the rest of the surface ; they are not islands of bi dimers , as may be seen after deposition at lower temperatures , but appear to be mixtures of si and bi atoms . in general , such small islands are not stable at these temperatures , as has been seen in hot stm studies of si homoepitaxy on si(001)@xcite . during growth , these small islands are stabilised by the high flux of si atoms , but if the flux is cut off , a process similar to ostwald ripening occurs , in which these islands will decompose , and the material in them will move to step edges ( which may be regarded as a very large island ) . the similarity of the behaviour observed here suggests that these islands are sustained by a flux of si and bi , which is produced during bi deposition . for low coverages of bi , the lowest - energy sites for bi dimers which have substituted for si dimers in the top layer of the si(001 ) surface hereafter `` bi surface dimers''are bi surface dimer / missing - si - dimer complexes@xcite , where a bi dimer decorates one or both sides of a 1dv , as seen in fig . [ fig : naitohmiki ] . for each adsorbed bi dimer , therefore , up to four si atoms are ejected from the surface layer , producing a significant transient flux of si atoms ; this flux in turn produces the observed islands . at the end of bi deposition , this flux will die away , the islands will no longer be stable and will decompose , as occurs between fig.[fig : promontory](a)and ( b ) . a cartoon of this process is shown in fig . [ fig : reservoir](d ) . once the small islands have annealed away , the surface morphology is very flat , and the background between the nucleating bi nanolines appears featureless in hot stm , apart from missing dimer trenches . however , at this stage a large surface density of bi may be inferred from the fact that bi nanolines continue to nucleate and grow on this surface . the bi dimers embedded in the surface layer have approximately the same contrast as the si dimers@xcite , and are therefore invisible at elevated temperature , as in fig . [ fig : reservoir](a ) . at a large positive bias , however , the background bi becomes visible , as is shown in fig . [ fig : reservoir](b ) . top - centre of this image is a bi nanoline . there is a high density of linear grey features in the background , one of which is marked with a pair of white arrows . detail of these features can not be discerned at this temperature , although the background bi is noticeably lower than the nanoline . after quenching the sample to room temperature , and exposure to hydrogen , we can judge the true coverage and distribution of bi during the annealing process , as displayed in fig . [ fig : reservoir](c ) . there is a high density of bi surface dimers , and also many bi dimers are adsorbed to one or both sides of a missing dimer trench , forming the bi1dv or miki structures , as shown in fig . [ fig : naitohmiki ] . thus the bi reservoir consists not of mobile ad - dimers , but mostly of bi surface dimers , usually decorating the missing - dimer trenches to form the miki structure . c , with a sample bias ca . + 3.0 v. grey linear features , which run parallel to the bi nanolines , can be seen in the background . an example is marked by white arrows . ( c ) : stm image taken at room temperature , with the silicon passivated with hydrogen to increase the contrast between the si and bi . the grey linear feature can now be seen to comprise bi dimers decorating a missing - dimer trench , i.e. the miki structure . ( d ) : a cartoon of the formation of the bi reservoir . impinging bi atoms ( 1 ) exchange with si atoms , making bi surface dimers and bi!dv complexes ( 2 ) . the ejected si atoms difffuse and form islands ( 3).[fig : reservoir ] ] during the high - temperature anneal , the surface bi reservoir is continuously depleted , with some bi going to form nanolines , and the remainder evaporating back into vacuum . the activation barrier for evaporation is different for different surface adsorption sites . a rheed study@xcite found that the activation barrier for evaporation of bi from the nanoline was 0.25 ev higher than the barrier for evaporation of bi from a ( 2@xmath1n):bi overlayer . hence bi is mostly being lost from the background reservoir . however , it might be expected that missing dimers in the nanolines would be observed occasionally . in fact , this is not the case , and those nanolines which are observed at high temperatures always appear perfect . it is likely that , due to the large defect energy of the haiku structure ( 0.66 ev from dft calculations@xcite ) , any bi which evaporates from the nanoline is quickly replaced from the bi reservoir . thus at high temperatures , where bi can evaporate from the nanolines , their stability is dependent upon a sufficient quantity of bi in the reservoir to replace missing dimers . at the later stage of an anneal , no new bi nanolines nucleate , and those which are present cease to grow . this is an indication that the reservoir of bi is approaching exhaustion . finally , when the reservoir is exhausted , evaporation from the bi nanolines will take place , and they will disappear within a few scans , or sometimes within the time of one scan . a sequence showing the evaporation of nanolines is shown in fig.[fig : linebreakup ] . in general , the nanolines evaporate from the ends , rather than breaking up into pieces , which strongly suggests that the ends of nanolines are less stable than the middle of the nanoline , and that the energetic barrier to break the nanoline in the middle and create two ends is prohibitive , even at these temperatures . the most notable feature of the mature nanoline surface as seen in fig . [ fig : recipes](c ) and ( d ) as well as fig . [ fig : mcleanimage ] is known as the promontory . long , narrow islands of si surrounding bi nanolines grow out across lower terraces . a series of images showing the early stages of bi nanoline growth are shown in fig . [ fig : promontory ] . in ( a ) , a short nanoline has grown out over a lower terrace and the joining of the promontory and the pre - existing step edge has formed a hole in the upper terrace , as is marked in ( b ) . the growth of this promontory is blocked by the growth of a second nanoline , which has nucleated in the time interval between ( a ) and ( b ) , and is growing out across the next lower terrace . comparison of ( b),(c ) and ( d ) makes it it very clear that the promontories are the result of growth out over a lower step edge . by ( d ) , a common pattern of zigzagging bi nanolines which have blocked each other s growth is developing . the mechanism behind this promontory growth has not been determined ; in an early paper , it was suggested that the promontories form by etching of si around the bi nanolines@xcite . however , in fig . [ fig : promontory ] , there is no sign of etching taking place ; this is a pure growth phenomenon . it is likely that the growth of the nanoline is related to the high mobility of step edges at 500@xmath3c . nanolines will always continue to grow until they reach an obstacle . thus nanolines will stop at step edges . however , the position of the step edge at this temperature is not constant ; it fluctuates back and forth@xcite , as material moves randomly along it . this motion may be inferred from the jagged , streaky step edges in these high - temperature images , as discussed in section [ sec : hot - stm ] . therefore , each time the step edge moves forwards , the nanoline can extend to the new step position , and then when the step edge tries to move back again , its position is pinned by the presence of the nanoline . with each forward fluctuation therefore , the nanoline will grow forward , and thus a promontory will gradually form . a similar process could explain the retreat of an up step ahead of a growing nanoline . the notable feature of the promontory is that they have a fixed minimum width . at least 3 - 4 nm of si is maintained on either side of the bi nanoline . this fixed width is an example of the phenomenon known as the defect exclusion zone , or `` dez '' , which is discussed in section [ sec : dez - sub ] . since the nanoline has a strain field associated with it , it is likely to exhibit quite strong interactions with other surface features which exert a localised strain on the surface . such features might be adsorbed bi dimers , missing dimer defects , step edges , and other nanolines . using tight - binding simulations , where very large cells can be relaxed , the interaction between a nanoline and several features , such as 1dv defects and rebonded b - type step edges have been calculated . these data are plotted in fig.[fig : deznewfig ] . from elevated - temperature stm images , it is clear that there is a repulsive interaction between the nanolines and missing dimer defects and step edges . at high temperature , as in fig . [ fig : deznewfig ] , the surface is covered in black streaks , which indicate the positions of rapidly moving missing dimer defects . however , either side of the nanoline , there are no streaks . likewise , where a nanoline forms a promontory , the width of this empty area is continued as the width of the promontory . this characteristic width of defect - free silicon , around 3 - 4 nm either side of the nanoline , is known as the `` defect exclusion zone '' or dez @xcite . for a large separation , the total energy of a nanoline and a defect is the same as for the two structures independently . however , as the separation decreases , the total energy increases , so that at a distance of about 3.5 nm , there is a repulsive interaction between the nanolines of ca . 0.1 ev for both step edges and 1dvs , and at smaller distances , this energy increases rapidly . statistics about the position of 1dvs in the space between two nanolines have been gathered , as shown in fig . [ fig : deznewfig ] , taken from ref . they show that on average , most defects occur at least 3.2 nm away from the nanoline , which is in agreement with the calculated repulsion . for a pair of nanolines , there is a similar interaction , and the 0.1 ev threshold is also reached when the centres of the two nanolines are about 3.5 nm apart , or in other words , the gap between the nanolines is the width of one nanoline . clusters of nanolines with this approximate spacing are often seen , as in fig . [ fig : deznewfig ] . for nanolines adjacent to each other , the excess energy is about 1 ev per unit cell . despite this , nanolines are sometimes observed growing next to each other , as in fig . [ fig : recipes](c ) . the reason for this is that unlike defects and step edges , which are mobile , a nanoline can not move sideways . hence when two nanolines which have nucleated in different places grow past each other , they can not move sideways to reduce their interaction energy . the presence of nanolines in close proximity does indicate that this interaction energy is not so high as to present an insuperable barrier to the growth of two nanolines past each other . for bi surface dimers , which have a compressive stress field , decoration of missing dimer trenches is a natural way in which to relieve their stress . the gain in energy by this process is about 0.45 ev / bi dimer in tightbinding calculations@xcite . the interaction with the nanoline is also attractive . in images where there is a large surface density of bi dimers , they are often situated adjacent to the nanolines . this position is found to be 0.28 ev better in energy@xcite than elsewhere . this suggests that the nanolines are generally more stable in a surface with a significant amount of compressive stress , such as one which is covered in bi . indeed the energy of a nanoline in a bi - terminated surface is 0.8 ev better than on a si - terminated surface . c. a series of 81 nm stm images . during the interval from ( b ) to ( c ) of 14 min . , the bi line on the bottom extended to separate the big bi island and to make an inlet as marked by a white arrow . during the same period , a new bi line appeared as marked by a black arrow . ( d ) to ( e ) shows more details of observation corresponding this latter event , in the area marked by a white box in ( c ) . the time between images was 18 sec . in figure ( e ) a small bi island is situated at the place marked by a black arrow , in ( f ) the small bi island changed into a short bi line , and between ( f ) and ( g ) the new line extends in both directions towards the island and the inlet . it is noted that the side edges of the inlet , and the shapes of all the islands change image by image . reprinted from surf . 421 , miki et al . bi - induced structures on si(001 ) , p.397 , ( 1999 ) , with permission from elsevier . ] despite the success of the haiku structure in explaining all of the above surface phenomena , its identification raises as many questions as it answers . how would such a complex structure form ? how is it terminated ? if termination is energetically unfavourable , what would a nucleus structure look like ? experimentally , the atomistic nucleation process remains a mystery . there is an incubation time during the annealing process before nanolines start to nucleate , which suggests that there is a significant barrier to formation of these structures . on many occasions , the appearance of bi nanolines has been captured during stm observations , such as in fig . [ fig : nucleationstm ] . there is no apparent precursor state ; like athena from zeus brow , a short nanoline springs fully - formed from an empty patch of the surface . the nanolines then grow extremely rapidly . thus far , there has only been one proposed mechanism for the nucleation of a bi nanoline@xcite . it has been suggested that the initial step is for bi ad - dimers to fall into the missing dimer trenches on the hot si(001 ) surface , forming a row of pairs of bi dimers . such a structure would result in the increase in local compressive stress , which may provide a driving force for nanoline nucleation . we demonstrated earlier ( fig.[fig : haikucartoon ] ) that the 7 - 5 - 7 structural motif may be achieved by a simple bond rotation process . using this motif , the proposed mechanism suggests that the central four atoms between the two bi dimers in the proposed mechanism twist , so that a zigzag core structure is formed . from this the haiku structure evolves , with a maximum barrier in the calculation of less than 1.5 ev . the reaction pathway and associated energy diagram are shown in fig . [ fig : nucleation ] . this mechanism has a number of interesting features , but also a major limitation . the calculation takes place in a periodically repeated unit cell one dimer row wide , and is therefore considering the nucleation of an infinite nanoline , _ without _ ends . it is suggested that a long nucleus forms by the filling of several unit cells of a missing dimer trench , but this is an unstable situation@xcite . moreover , in our stm observations , we observe the appearance of very short nanolines , which then grow . the termination of these finite nuclei _ must _ therefore be considered . stm observations of nanoline evaporation show that nanolines do not break up in the middle , instead they evaporate from the ends . from this , it may be inferred that the energy of termination is quite high . while there has been no systematic study of termination , we have performed calculations on many and varied structures , all of which have been found to have a high energy . in tightbinding modelling , the energies were at least 1 ev / end , while those terminations which we have considered for the proposed zigzag nucleus structure have much higher energies , around + 1.6 ev / end , giving a total nucleus energy of over + 4 ev . this suggests both that understanding the termination of the nanolines is an important area for future research , and that nucleation structures should be considered in 2d , rather than in 1d ( which may not be feasible with a technique more accurate than tightbinding ) . given that the prevalent structures on the bi - rich surface are bi surface dimer/1dv complexes , alternative nucleation mechanisms which use these structures as a basis may be more fruitful and are being explored . the knowledge of the electronic structure of the bi nanoline has stemmed from two sources : variable - voltage stm@xcite and dft calculations of the electronic structure@xcite ( generally considered in terms of the local density of states ( ldos ) , or the tersoff - hamann approach to stm , which uses the ldos ) . knowledge of the atomic structure of the bi nanoline allows detailed calculations of its electronic structure to be made , which can then be compared to the observations made in stm . while the majority of the electronic contrast is expected to be dominated by the bi dimers in the nanoline , other details of the ldos have allowed a stronger identification of the haiku structure with the bi nanoline to be made , with the other possible candidate structures ruled out after comparison with the stm data . 10 nm images of a bi nanoline . the sample bias voltages used are -2.0v , -1.2v , -0.6v , -0.3v , + 1.5v and -0.8v , in ( a - f ) respectively . as the sample bias is reduced , between ( a ) and ( d ) , the nanoline changes contrast from light to dark relative to the surrounding si(001 ) . over this range , some of the dimers in the nanoline ( marked a in ( c ) and ( e ) ) exhibit a different voltage contrast . at very low biases , around -0.3v , an enhancement of the dimers around the nanoline , similar to that seen around a missing dimer defect in clean si(001)@xcite , is seen . this is visible ( and marked schematically as b ) in ( d ) . in ( f ) , the resolution is sufficient to see that the corrugation of the si dimers closest to the nanoline is increased , ( marked schematically by the dotted black lines ) suggesting a greater separation , and hence tensile strain . reprinted from surf . 527 , owen et al . interaction between strain and electronic structure in the bi nanoline , p.177 , ( 2003 ) , with permission from elsevier.[fig : varbias ] ] we begin by summarizing the data available from stm , which is illustrated in fig . [ fig : varbias ] : 1 . at high biases , the bi nanolines appear bright compared to the surrounding si(001 ) surface@xcite 2 . at low biases , the bi nanolines become dark compared to the surrounding si(001 ) surface@xcite 3 . at low biases , the si dimers neighbouring the bi nanoline show enhancement , or brightening@xcite similar to that seen around missing dimer defects on si(001)@xcite . these facts suggest various things : first , that the bi dimers are physically higher than the surrounding si(001 ) surface dimers , as at high biases the contrast in stm is determined primarily by geometrical structure ; second , that the _ local _ electronic structure of the bi dimers will show states further from the fermi level than the surrounding si dimers@xcite , as this determines the stm current at low biases ; finally , that the bi nanoline structure must be under tensile stress , which causes the si dimers near the nanoline to be pulled together , away from their relaxed surface structure , which is the origin of enhancement at low biases@xcite . examination of the detailed electronic structure of the miki and naitoh models@xcite and the haiku structure@xcite all show that the bi dimers become dark at low bias voltages by contrast to the surrounding silicon dimers . the calculated dos for the miki and haiku structures are shown in figure [ fig : miwados]@xcite . the @xmath13 peak comes from the up atoms of the substrate si dimers , as does the @xmath14 , while the @xmath15 peak comes from both the bi and si dimer bonds and @xmath16 from the bi atoms in the line models . when the substrate dimers are passivated with h , the peaks @xmath13 and @xmath14 almost disappear leading to peaks in the gap which arise from the bi dimers ( @xmath15 and @xmath16 ) . this data confirms earlier modelling and bias - dependent stm measurements@xcite discussed below . consideration of the electronic structure of the bi dimer reveals that it is rather unusual . in its native form , bi forms puckered hexagonal sheets , similar to the structure taken by as . full - potential lapw modelling of bi in a bulk diamond cell ( thus forcing it to take up tetrahedral bonding ) showed that there is very little hybridisation between the s- and p - orbitals@xcite . thus we expect bi to tend towards p@xmath17 bonding with a filled s - orbital ( as a lone pair ) and to be energetically most favourable with 90@xmath4 bond angles . this explains to some extent the haiku structure , where bond angles are almost 90@xmath4 . projection of the local densities of states calculated using pseudopotential - based dft calculations@xcite suggests that the bi dimer in the haiku structure has hybridisation between s and p@xmath18 , with p@xmath19 and p@xmath20 separated off . the flapw calculations showed that the relativistic effects of the full core increased the s - p splitting in the bi atom ; while relativistic effects can be incorporated in pseudopotentials , the haiku structure might be more stable , and show bonding closer to p@xmath17 , if these effects were taken fully into account . the bi - bi bond is a rather weak bond , with the bi - si bonds providing most of the energetic stability . this can be seen by considering the electronic structure of a bi ad - dimer on the clean surface . the electron localisation function ( elf)@xcite for this ad - dimer is shown in figure [ fig : bidimerelf](a)@xcite . the elf gives a quantitative analysis of the bonding properties of a system , with a value of 0.5 equivalent to the localisation seen in the homogeneous electron gas and a value of 1.0 equivalent to complete localisation . the figure shows an isosurface with elf=0.8 ( equivalent to reasonably strong covalent bonds ) . it can be seen that the si - si bonds are rather strong , as expected , and that the bi - si bonds are still visible at this level , though they are weaker than the si - si bonds . the bi - bi bond is not visible ( in fact it has a maximum value around 0.7 ) but the bi lone pairs are visible as rather diffuse , spherical objects ( by comparison to the up atoms of the substrate dimers ) . this implies that the bi - bi bond is rather weak , and the lone pairs would not be available for bonding@xcite . the energy barrier to breaking the bond in this bi dimer has been calculated with dft to be 0.17ev going from over the dimer row to over the trench between rows and 0.15ev going in the other direction , and the barrier is shown in fig . [ fig : bidimerelf](b ) . having the bi dimer over the dimer row is more stable than over the trench between rows by about 0.02ev . this implies that there should be a finite chance of imaging the haiku structure with the bi dimers out - of - phase with the underlying si dimer rows . however , we note that with the energy difference noted above , the dimer would be over the row 98% of the time at room temperature , and the likelihood that a whole line of dimers would flip , will be even smaller . there are occasional images with the dimers out - of - phase , but these are rare . , title="fig : " ] , title="fig : " ] in the next section , details of some elements which will attack the bi - bi bond will be presented , along with others that will not . here we consider the local structure of the bond . when the bi - bi bond is over the haiku core , it is almost perfectly relaxed , with bond angles extremely close to 90@xmath4 . insertion of any species into the bi - bi bond will distort these angles ( and the angles formed with the underlying si atoms ) , with a potentially large energy penalty . thus the bi - x - bi angle should be relatively close to 180@xmath4 with short bi - x bond lengths if an element x is to insert successfully into the bond . furthermore , the lone pair on the bi dimer , which might be expected to be a strong electron donor , is in fact rather passive , as it lies some distance below the fermi level@xcite . the implications of this for the chemical reactivity of the bi nanoline are discussed below in section [ sec : reactivity ] . the phenomenon of strain - enduced enhancement at low stm bias voltages has been explored before@xcite ; we reproduce the arguments here for convenience@xcite . the si(001 ) surface shows a dimer reconstruction which provides three bonds per surface si atom ( compared to two bonds per atom for the bulk terminated surface ) , but which pulls the bond angles and distances away from the ideal tetrahedral values . the net result is that the local band gap is reduced ( in simple chemical terms , the hybridisation towards sp@xmath17 is decreased , leading to a decreased splitting between the resultant hybrid bonds ) . if the surface dimers are then further distorted , their local states will move closer to the fermi level . the single missing dimer defect ( 1dv ) has second layer atoms bonding across the defect which distorts the dimers either side of the defect ( though lowering the energy because the second layer atoms have no dangling bonds ) . this distortion in turn causes the states near the fermi level on the neighbouring dimers to move towards the fermi level , leading to a brightening or enhancement at low biases . low - bias simulated images of the naitoh and miki structures@xcite show that the si dimers adjacent to the naitoh become _ darker _ than the surrounding si dimers as the fermi level is approached , while those on the miki model are unchanged . by contrast , with the haiku structure , the si dimers immediately adjacent to the nanoline become brighter closer to the fermi level@xcite , in good agreement with the stm images . a comparison of high and low - bias simulated images of the 3 structures are shown in fig . [ fig : simstm ] . however , the prediction from the miki model does agree very well with the appearance suggested from stm images@xcite , in which the shoulder dimers around a 1dv on a bi - rich surface become dark at low bias voltages . on the basis of these calculations , we suggest that the lack of enhancement at low bias voltages indicates almost complete relaxation of the strain of the 1dv by the presence of the bi dimers in the miki model . taking the principle of strain - induced enhancement further , the local subsurface strain could be determined by cross - sectional stm of the bi nanoline , of which there is none , or from ldos calculations , by projection of the charge density on a plane perpendicular to the ( 001 ) surface . we show the projected charge density from gga calculations for the clean si(001 ) surface , the haiku structure and the haiku structure with hydrogen on the si(001 ) surface in fig . [ fig : ldos]@xcite , on the ( 1@xmath210 ) plane , parallel to a si dimer row . by comparing in particular the clean si(001 ) surface and the clean haiku structure , the change in different states induced by the haiku can be seen . beginning with the clean surface , the localisation of the states close to the fermi level in the top few surface layers shows the strain induced by dimerisation . in the haiku structure , these states are even more strongly localised on the dimers immediately adjacent to the haiku structure , which causes their relative enhancement in stm . meanwhile the triangular si substructure of the haiku , c.f fig . [ fig : haikucore ] , is somewhat relaxed compared to the clean si dimers , although it is strained relative to the hydrogenated dimers and perfect bulk si . this goes some way to explaining the stability of the subsurface 5 - 7 - 5 ring structures , which serve as a highly effective relief mechanism for the epitaxial stress exerted by adsorbed bi . the absence of states close to the fermi level suggests that the nanolines are likely to block surface conduction perpendicular to the nanoline , and are not likely to act as a nanowire . the lack of any states on the bi near the fermi level is clearly the cause of the darkening of the nanoline relative to the si(001 ) surface seen in stm at low voltages . in the full charge densities(fig . [ fig : ldos](iv ) ) , the density of states associated with the bi dimers , combined with their higher physical height , explains their relative brightness in stm at higher bias voltages . in the case of the hydrogenated surface , the bi dimers have a similar charge density as in the clean surface , but in this case , the si @xmath22-bonds have been eliminated , so the bi dimer remains bright in stm at all biases@xcite . as described above in more detail , the bonding orbitals in the bi dimers have essentially a p@xmath23 bonding character , while the lone pair is a diffuse s - type orbital , some 10 ev below the fermi level@xcite . thus the bi dimer is in a very stable electronic state , and is not expected to be very reactive . moreover , this low - energy configuration is dependent upon the maintenance of bond angles close to 90@xmath3 , which will have an effect on the energetics of any insertion into the bi - bi dimer bond . however , the low - bias images shown above reveal that although the nanoline appears clean at high biases , some contrast between different dimers is visible at lower biases , suggesting that some chemical attack has occurred . we have proposed that the bi nanolines , although semiconducting themselves , might be used as templates for deposition of other materials , such as metals , active molecules , nanoparticles etc.@xcite . in order for the nanoline to have utility in this role , it must be possible to mask the substrate around the nanoline , so that adsorbants stick preferentially to the nanoline . it is also important , for any nano - electronics applications , to have some method for rendering the substrate insulating after growth of the nanoline . the chemical reactivity of the bi dimers , and the nanoline as a whole , is therefore crucial to any implementation of this concept . atomic hydrogen has been used as a mask on the bare si(001 ) surface for stm lithography experiments@xcite , in which an stm tip is used to remove lines or areas of hydrogen from the surface , and thus provide areas onto which metal will adsorb , as in fig . [ fig : hashizume ] . however , the writing process is slow , and lacks scalability . use of the bi nanolines as templates would remove the need for this writing , while providing long , and more uniform patterns . hydrogen is therefore a natural species to use as a mask around a bi nanoline . one of the most favourable properties of silicon as an electronics material is the high quality of its oxide film . oxidation of the silicon around the bi nanoline is therefore a natural way in which to isolate the nanoline electrically from the substrate . the usual method of oxidation of silicon is by exposure to molecular oxygen or to water at elevated temperatures , but this only gives microelectronics - quality oxide when grown at ca.800@xmath3c . since the bi nanoline is not stable above 600@xmath3c , ozone was considered as an alternative for low - temperature oxidation . ozone preferentially attacks the backbonds of the surface si dimers , forming a stable oxide , even at room temperature@xcite . more recently , the need for high - k dielectrics has driven the development of silicon nitride and oxynitride as alternatives to silicon dioxide . nitridation is usually done using ammonia , which breaks up on contact with the clean silicon surface above 200 k@xcite . as an added bonus , when ammonia reacts with silicon , the nitrogen moves below the surface , but the hydrogen remains on the surface , passivating it . it can therefore play a dual role , as insulating reagent , and as masking material . the reaction of these species with the bi nanoline are discussed in the sections below . the bi nanoline has been shown to be almost completely inert to attack by atomic hydrogen@xcite . fig.[fig : bihydrogenoxygen](a ) shows a close - up image of a nanoline on a si(001):h monohydride surface . in this image , nanolines were formed at 590@xmath3c , and the surface was cooled to 330@xmath3c , the surface was saturated with atomic h. the surface has the monohydride phase , with the si - si @xmath24-bond intact , and the h atoms saturating the dangling bonds . thus each si - h dimer is resolved as a pair of dots . some white dots on the si surface may be holes in the monohydride layer , stray bismuth atoms , or adsorbed -oh resulting from water impurity in the hydrogen . while the nanoline is almost completely unreacted , one dimer has become dark . this sort of defect is extremely rare . the preferential adsorption of the h onto the si is consistent with dft - lda modelling , which found that h on a bi dimer is 3 ev worse than h on a si dimer . there is therefore a strong driving force for any h which does adsorb onto the bi nanoline to diffuse off it , and onto the si . at 300 k. there is virtually no attack on the nanolines , while the substrate has been oxidised . ( c ) : after exposure to 150l o@xmath11 at 700 k and an anneal to 840 k , some gaps appear in the nanoline , but most of it is unaffected . ( d ) : exposure of a h - terminated surface with bi nanolines to air results in complete oxidation of the nanolines , which now show up dark relative to the h - terminated si.[fig : bihydrogenoxygen ] ] the oxidation properties of a monolayer of bi on si(001 ) shows an unusual changeover in behaviour with oxygen exposure . an eels / auger study@xcite found that for small exposures , the reactivity of a bi overlayer on si(001 ) is 100 times less than the reactivity of oxygen with the bare silicon surface . however , for large exposures , above ca . @xmath25 l o@xmath11 , the oxygen auger intensity continues to increase for a bi / si(001 ) surface , at a point where the signal for the clean si(001 ) surface has saturated . thus for very large exposures , the bi promotes oxidation of the silicon . moreover , the stoichiometry of the oxide formed on the bi / si(001 ) surface is closer to sio@xmath11 , which would imply better insulating properties . semiempirical calculations found that insertion of oxygen into the bi - bi bond is energetically much less favourable than for sb or as ( 0.32 ev vs ca . . moreover , insertion into a bi - si bond or a si - si bond is much more favourable ( 4.5 ev and 5.5 ev respectively ) , so that preferential attack of the silicon is expected . the reactions of molecular oxygen and ozone with the bi nanoline show a similar changeover in behaviour@xcite . the surface was exposed to ozone at room temperature and , for comparison , to molecular oxygen at 400@xmath3c . both surfaces were annealed up to 570@xmath3c , near the stability limit of the bi nanoline . the room - temperature ozone experiments , using exposures up to 20 l of o@xmath26 , revealed that the silicon background could be saturated without visible damage to the nanoline . in large - scale images , the bi nanoline appears to be completely unaffected , as shown in fig . [ fig : bihydrogenoxygen](b ) . however , in small - scale images , and by varying the imaging bias , some subtle differences between the appearance of various dimers in the nanoline could be seen@xcite . this variation may be due to attack of bi - si backbonds , as found to be favourable in mndo simulations@xcite , which would leave the bi - bi dimer intact , but perturb its electronic structure somewhat . annealing this surface introduced some defects into the nanoline . after annealing at 450@xmath3c , gaps appear in the nanoline , which comprise 1 - 2 unit cells . this trend continued with a further anneal to 570@xmath3c , as shown in fig . [ fig : bihydrogenoxygen](c ) . such piecemeal damage may be explained by the observation that the activation barrier for bi to desorb from oxidised si has been found to be 0.7 ev lower than for unoxidised si@xcite . the small contrast changes , which were seen after ozone exposure at room temperature may therefore indicate reaction with bi - si , or si - si backbonds around the nanoline , so making that unit cell more susceptible to evaporation during high - temperature annealing . molecular oxygen , when adsorbed at 400@xmath3c , behaved in a very similar way to ozone . again there was preferential attack of the silicon , leaving the nanoline mostly intact , but also gaps appeared in the bi nanoline after annealing for about one hour@xcite . while the bi nanoline is remarkably resistant to oxidation in uhv , it quickly oxidises when exposed to air , as shown by the following simple experiment . a h - terminated si(001 ) surface with bi nanolines was taken out of the imaging chamber into the uhv system loadlock , which was then vented to atmosphere for 30s , and then immediately pumped back down to uhv . the resulting surface is shown in fig . [ fig : bihydrogenoxygen](d ) . the nanolines , which previously were bright at all biases relative to the h - terminated background , are now black lines , indicating that they have suffered significant attack . this agrees well with the previous study , which found that for very large exposures to oxygen , the bi dimers promoted oxidation of the underlying silicon . the adsorption of ammonia onto si(001 ) as a precursor to silicon nitride or oxynitride growth has been studied by a variety of experimental and theoretical techniques , which concluded that at room temperature , the ammonia dissociated on the surface , forming nh@xmath11 and h groups@xcite . we have found that ammonia attacks the silicon preferentially at room temperature . an image of the bi : si(001 ) surface with ammonia termination is shown in fig.[fig : biammonia ] . within each dimer , the light grey end has been identifed with the nh@xmath11 group , while the dark grey end is the h atom@xcite . the apparent height difference in stm matches the physical height difference between these two species . there is some ordering of the nh@xmath11 groups , such that zigzag patterns and straight lines are seen running along the dimer rows . reaction of ammonia with the bi nanoline might be expected to proceed by the initial formation of a hydrogen bond between an ammonia h atom and the bi lone pair , followed by insertion into a bi - bi or bi - si bond . however , as discussed above , the lone pair on the bi dimer is essentially inert@xcite . during experiments on a surface which was pre - exposed to atomic h to saturate the si dangling bonds , for exposures of up to 70l of ammonia at substrate temperatures of up to 500k , the bi nanoline was largely unaffected . however , adsorption of ammonia onto a bi nanoline surface where the si dimers are not passivated does result in some reaction with the nanoline , even at room temperature . this difference may indicate that ammonia reacts first with the silicon , forming active nh@xmath11 species , which are mobile and can attack the bi dimers . calculations suggest that it is energetically favourable for an nh@xmath27 group to insert into the bi - bi dimer bond , and form a symmetrical bi - nh - bi feature . however , stm images of reacted bi dimers shows an asymmetric feature , not unlike the feature seen on background si dimers . examples may be seen in the inset to fig . [ fig : biammonia ] . nevertheless , the probability of attack by ammonia is low , providing the promise that ammonia can be used to grow silicon nitride without damage to the nanoline . 64 nm stm image of the ammonia - terminated bi : si(001 ) surface . the background si is completely terminated by a mixture of h and nh@xmath11 groups . dark patches on the bi nanoline reveal some attack by ammonia . the inset shows that reaction occures within the space of a single dimer , and the resulting feature is asymmetric , much like the termination of the background si . ] metal deposition onto the bi nanolines is currently under investigation . deposition of ag onto bi nanolines on the clean si(001 ) surface@xcite reveals a strong preference for adsorption onto the si , leaving the bi nanoline clean ( in complete contrast to the behaviour of ersi@xmath9 nws on si(001 ) discussed above in sec . [ sec : reactivity-1 ] ) . however , the passivity of the bi nanolines to atomic hydrogen and ammonia allows these species to be used as nanoscale masks . ag is known not to adsorb onto h - terminated si(001)@xcite as shown in fig . [ fig : hashizume ] , while in completely dewets from the si(001 ) surface , forming large droplets when h is adsorbed onto an in - covered surface@xcite . in all cases , following metal deposition onto an ammonia - terminated si(001 ) with bi nanolines , preferential adsorption onto the bi nanoline ( or onto background bi ) was seen , indicating that the ammonia - terminated surface is very effective as a mask against metal deposition . the reaction of the metal with the nanoline template is more complex , however . two types of behaviour have been seen with different metals , which may be described as wetting and non - wetting behaviour . wetting occurs when the interaction of the metal with the nanoline is stronger than the interaction of the metal with itself . nonwetting occurs when the interaction of the metal with itself is stronger . examples of each type of behaviour for different metals , in and ag , are shown in fig . [ fig : metal ] . adsorption of in results in significant intermixing of in and bi , forming a zigzag island structure , which is thought to result from a chain of in and bi atoms , as shown in the inset to fig . [ fig : metal](a ) . second - layer islands have a distinctive bright hexagonal feature , which is 19 or 2.5 dimer rows wide . this rational relationship between the in island dimension and the underlying si unit cell distance demonstrates that these in islands grow _ epitaxially _ on the nanoline template , raising the possibility of the growth of long single - crystal nanowires on the si(001 ) surface . by contrast , adsorption of ag results in the formation of small clusters of ag , around 0.5 nm in height , at a very early stage of deposition , as shown in fig . [ fig : metal](c ) . further deposition at room temperature results in an increase of the number density of these nanoclusters , but the modal size increases only slightly from 0.5 nm to 0.6 nm , with a strongly peaked height distribution . this behaviour may be thought of as non - wetting behaviour . while the bi nanoline has a number of interesting properties , for any nano - electronics device application it must be possible to passivate them against atmospheric attack . one way in which to do this is to bury them in a layer of epitaxial silicon . moreover , the burial of the bi atoms may lead to their acting as a 2d delta - doping layer , with close to atomic - layer precision in the depth placement . however , as with other nanostructures , for example burial of inas / gaas quantum dots in gaas , the process of burial is likely to change the shape and detailed structure c. most of the bi has segregated to the surface , and the nanolines have been destroyed . ( b ) : the bi nanoline surface ( a ) has been capped with a monolayer of bi at 400@xmath3c , before overgrowth . after overgrowth at 400@xmath3c up to position ` b ' , the growth layer was buried in amorphous si by further deposition of si at room temperature . in this case , nearly all the bi from the bi nanolines remains at the original surface position , ` a ' . ( c ) : as ( b ) , except that the capping layer material was sb , in place of bi . here most of the bi has segregated to the growth surface ( b ) , leaving sb at the original surface ( c ) , indicating that the bi has exchanged with sb from the capping layer . ] simple overgrowth of a surface containing bi nanolines results in the destruction of the nanolines@xcite . as shown in the secondary ion mass spectrometry(sims ) profile in fig . [ fig : sims](a ) , after overgrowth of si the vast majority of the bi has segregated to the surface of the grown film , although there is a small peak in the bi concentration at the original surface position . furthermore , x - ray analysis@xcite shows that the 1-d character of the nanoline is lost after capping , even by an amorphous overlayer deposited at room temperature . the mechanism by which the bi floats to the surface during growth is thought to involve exchange of subsurface bi atoms with surface si atoms , as bi is an effective surfactant during growth . in order to prevent this , a further surfactant layer of bi can be deposited around and over the bi nanolines , before si overgrowth@xcite . in this case , the exchange mechanism is inoperative , as the second - layer bi in the nanolines will not exchange with first - layer bi , and so it is possible to block the segregation of the bi nanolines to the growth surface . in the sims profile in fig . [ fig : sims](b ) , a thin crystalline si overlayer has been grown at 400@xmath3c after deposition of a bi layer , and this has been capped by depositing further si at room temperature , so as to bury the growth surface in amorphous si . in this case there , are two peaks in the sims profile . ` a ' refers to the bi density at the original bi nanoline surface , while ` b ' refers to the surface of the crystalline overgrowth . the surface of the overlayer has approximately 1 ml of bi on it , from which the quantity of buried bi can be determined . this is approximately 5% ml , which is typical for a bi nanoline surface . thus , using the bi surfactant layer , most of the bi nanolines have been preserved at the original surface . in order to gain more information about the exchange process , a surfactant layer of sb was also used , instead of the surfactant layer of bi . the resulting sims profile is shown in fig . [ fig : sims](c ) . in this case , there is a large peak of sb at the original nanoline surface position ` b ' , while a considerable fraction of the bi has segregated to the surface of the overlayer ` c ' . this result indicates that there is considerable exchange between the sb surfactant layer and the bi nanolines , and by this method , a layer of sb has been successfully buried , with a sharp peak at the original nanoline surface . while this exchange process is successful at burying the bi atoms , it does not provide any information about the structure of the buried bi . for device applications , it is important to retain the 1-d character of the original nanoline . it is also of interest whether the haiku structure will be preserved . stm can not observe subsurface features , and instead the structure of the buried bi nanolines was studied using an x - ray method@xcite . the results of this analysis@xcite indicate that two important elements of the bi nanoline structure have been preserved : the 1-d character , and the dimerisation of the bi atoms . tightbinding and dft calculations of the buried bi nanoline structure found that the haiku structure was not stable , and broke down into a hollow 1-d structure with bi dimers bonding to the silicon grown _ over _ the nanoline , and the si substructure reconstructing into something similar to a 2dv surface defect . a ball - and - stick model of this proposed structure is shown in fig . [ fig : buriedhaiku](a ) . it has significantly lower energy per bi atom than other proposed structures , including substitutional and interstitial bi defects . this structure may be compared to the cross - sectional tem image of buried bi nanolines in fig . [ fig : buriedhaiku](b ) . two nanolines are present within the tem image . the one on the left has strain associated with it , and a dark core , which may match with the proposed structure . the one on the right has triggered misfit dislocations along \{111 } planes , c.f . [ fig : haikucore ] . these dislocations , which indicate the presence of a large local stress , may suggest that the haiku substructure of the nanoline has survived overgrowth . a recent x - ray standing wave ( xsw ) study of a bi nanoline sample which had been capped with amorphous si at room temperature@xcite has cast doubt on the validity of the haiku model . however , more detailed analysis of lines buried with amorphous silicon@xcite showed that the dimerisation and one dimensional character of the nanolines is destroyed by burial in amorphous silicon ( and , as discussed above , burial in crystalline silicon changes the haiku reconstruction while leaving the one dimensional , dimerised character of the nanolines unaffected ) . furthermore , the presence of a significant quantity of bi at the growth surface may have an effect on the results obtained , unless it is removed@xcite . we have given a broad overview of the different material systems and experimental methods which result in self - assembled 1d nanostructures on semiconductor surfaces , particularly the technologically important si(001 ) surface . there are many different self - organising systems which result in 1d nanoscale features on the surface of si(001 ) and si(111 ) . aside from the pt / ge(001 ) system , none of these approach the length , and degree of perfection , of the rare - earth silicide wires and the bi nanolines . the metal / si(111 ) nanowire systems will tend to form 3 equivalent domains , limiting the long - range order of the nanowires , while the methods which rely on step - edge adsorption are limited by the spacing of step kinks ( though this can be controlled with careful sample preparation ) . the rare - earth silicide nanowire family are not nanowires _ per se _ ( in that they are a metastable state , not seen in a bulk structure ) , but are better regarded as a conventional heteroepitaxial system in which the lattice mismatch is anisotropic . even so , while many different rare earths superficially appear to behave in a similar fashion , review of the literature indicates that in fact several different nucleation and growth mechanisms are responsible for the formation of nanowires within this family , both kinetic and thermodynamic . the recent observation of nanowires with an orthorhombic crystal structure@xcite , rather than the hexagonal structure previously reported , suggests that there is much more to be learnt about this family of materials . experimental studies of all these systems has so far been largely confined to stm , apart from x - ray and tem studies of buried bi nanolines , and tem of the endotaxial re nanowires . a much greater understanding of the physical , chemical and electronic properties of these systems , would come from other experimental techniques , such as optical and vibrational spectroscopic techniques . ( for example , the strained haiku core of the bi nanoline may have a signature bond vibration frequency . ) the main focus of this review has been the bi nanoline system . we have described the physical and electronic structure of the bi nanoline , its reactivity with different materials , and the surface phenomena associated with it . although there has been no direct observation of the core structure of the haiku model , it provides a natural explanation for all the observed experimental phenomena , and is well - supported by experimental and theoretical data . the fundamental source of many of the properties of the bi nanoline is the subsurface core of si which has recrystallized into a hexagonal phase , lonsdaleite ( known from meteorites ) . the formation mechanism of this core structure remains unknown . a significant activation barrier is expected , given the high temperature and long incubation time required to form the nanolines , and the co - existence of the nanolines with the miki structure , which is less stable , but kinetically easy to form . while mechanisms proposed thus far have concentrated on the twisting of pairs of atoms , an alternative may well be a stacking - fault - like shift of the entire core from the cubic phase to the hexagonal phase . given their structure , it is clear that the bi nanolines are not a conventional heteroepitaxial system like the silicide family . there is no variation in width , either with annealing temperature or time only the number density of nanolines and the length changes with further deposition and annealing . the nanolines are only ever one unit cell wide , although clusters of nanolines are sometimes seen . the nanolines are not the result of simple adsorption onto symmetry - breaking features of the surface , such as step edges , or linear defects . nor are they a periodic reconstruction of the surface , as in the pt / ge wires@xcite . they might best be described as an adsorbate - stabilized surface recrystallization . the chemical properties of the nanoline stem from the behaviour of the bi dimer with its preference for p@xmath23 bonding and an inert s - type lone pair , rather than the sp@xmath23 hybridization of si dimers . although the bi - bi bond is weak , it is also passive , and only in shows any significant reaction with it , probably because the zigzag structure produced allows the bi atoms to retain their preferred 90@xmath3 bond angles . hydrogen does not attack the bi nanoline , while ozone will oxidise around the nanoline at room temperature , leaving it intact . likewise ammonia preferentially attacks the silicon , although some damage to the nanoline occurs at small fluxes . the determination of both the physical and electronic structure of the nanoline has demonstrated the importance of a close interaction between experiment and theory . we note that all the recent publications on the bi nanoline have been a collaboration between the two . the method of tightbinding has proved to be a valuable tool for quickly searching through a large variety of possible structures , as in for instance the haiku structure , or for relaxing a large number of atoms , as in modelling of the nanoline kink energies . at the same time , the tightbinding method retains a quantum mechanical description of the energetics , so that the calculated energies are in most cases in reasonable agreement with those calculated using dft . the bi nanoline has benefitted from an unusual degree of theoretical investigation . a similar level of theoretical work for the rare - earth nanowires is likely to deepen greatly the understanding of the different nucleation and growth phenomena that have been identified in that family of systems . with a significant body of work devoted to the fabrication and structure of the nanoline systems , attention should be paid to their possible applications . the length , straightness and perfection of self - assembled nanolines lend themselves to use as a nanoscale template . the most obvious candidate systems are the rare - earth nanowires , which are themselves metallic , so that an array of parallel nanowires would readily form the contacts for a nanoelectronic device , perhaps using the `` crossbar '' architecture . the er silicides appear also to act as preferential adsorption sites for metals such as pt , so that the properties of these contacts is not limited to that of the nanowires themselves . however , the growth of these nanowires on a semiconducting substrate , with no indication as yet that the substrate could subsequently be oxidised , might limit their application in nanoelectronics ( for which a high - quality insulating substrate is crucial ) . for the bi nanolines , oxidation of the surrounding silicon with ozone , followed by passivation of the surface with hydrogen , or nitridation using ammonia , is a plausible route for the fabrication of a single - nm 1d template on an insulating surface , considerably smaller than any of the rare - earth nanowires , and well within the size range where quantum scale effects might be expected to occur . by the deposition of metals , active molecules , nanoparticles , or other species onto this template , the fabrication of nanowires , or arrays of nanoparticles with interesting optical , electronic or magnetic properties could be achieved , with atomic precision . this area is being actively pursued by the authors . we are happy to acknowledge useful discussions with chigusa ohbuchi and wataru yashiro , as well as permission from various authors to reproduce figures as noted in the text . this study was performed through special coordination funds for promoting science and technology from the mext , japan ( icys and active atom - wire interconnects ) . drb is funded by the royal society .	a number of different families of nanowires which self - assemble on semiconductor surfaces have been identified in recent years . they are particularly interesting from the standpoint of nanoelectronics , which seeks non - lithographic ways of creating interconnects at the nanometre scale ( though possibly for carrying signal rather than current ) , as well as from the standpoint of traditional materials science and surface science . we survey these families and consider their physical and electronic structure , as well as their formation and reactivity . particular attention is paid to rare earth nanowires and the bi nanoline , both of which self - assemble on si(001 ) .
a large variety of independent astronomical observations have revealed that most of the mass in the universe is some form of non - baryonic , non - luminous , cold dark matter of as yet unknown composition . understanding the fundamental nature of dark matter is one of the biggest outstanding problems in cosmology and particle physics ( for a comprehensive review , see e.g. * ? ? ? compelling particle candidates for the dark matter , collectively termed weakly interacting massive particles ( wimps ) , exist in several well - motivated theoretical extensions of the standard model of particle physics like supersymmetry ( lightest supersymmetric particle , for a review see e.g. * ? ? ? * ) and universal extra dimensions ( lightest kaluza - klein particle , for a review see e.g. * ? ? ? * ) . a signal from dark matter could be detected in astronomical observations from wimp pair annihilation or decay . for example , wimp annihilation or decay generically leads to the production of gamma - rays ( as well as of other high energy particles ) , potentially at a detectable level with current telescopes @xcite . the unparalleled sensitivity of the _ fermi _ large area telescope ( lat ) at gev energies makes it an excellent instrument to look for this gamma - ray signal ( e.g. * ? ? ? in addition , the recent detection by pamela of a positron excess relative to cosmic - ray secondary production in the galaxy @xcite and a gentle but clear feature in the _ fermi_-lat positron - electron ( @xmath0 ) spectrum @xcite has led to intense discussion of the possibility that one or both of these signals are due to dark matter . if true , we expect a corresponding signal in gamma - rays stemming from both the annihilation or decay event itself , as well as from the secondary inverse compton ( ic ) up - scattering of background radiation by the high energy @xmath1 and @xmath2 produced . as far as dark matter annihilation is concerned , first year _ fermi _ observations of dwarf spheroidal galaxies , clusters of galaxies , the extragalactic gamma - ray background , and searches for gamma - ray lines have placed some of the strongest constraints yet on the possible wimp pair annihilation cross - section and have already excluded some annihilation models that might explain the pamela positron excess @xcite . in addition , several authors have recently pointed out that gamma - ray data ( including early _ fermi _ results ) also put constraints on the dark matter decay explanation of the pamela data , excluding some , but not all , of the relevant parameter space ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? these studies typically utilize _ fermi_-lat measurements of the diffuse galactic and isotropic gamma - ray emission or _ fermi _ sky maps and require that the expected decay signal from the milky way halo or the summed contribution from extragalactic dark matter not overproduce the observed gamma - ray fluxes . limits on the dark matter lifetime from the observation of dwarf galaxies with atmospheric cherenkov telescopes have also been derived in @xcite . interestingly , these analyses indicate that current gamma - ray constraints on the decaying dark matter scenario are generically weaker than those on annihilating dark matter , with respect to the regions of parameter space providing an explanation to the cosmic - ray electron - positron anomalies . moreover , as we review in the next section , several theoretical arguments imply that dark matter lifetimes in the range needed to explain the cosmic - ray anomalies are at least plausible , specifically in the context of supersymmetric grand unification @xcite . beyond the interest in scenarios where decaying dark matter might be related to the cosmic - ray anomalies , several particle physics frameworks feature dark matter candidates with a long , but finite , lifetime . we provide an overview of some of these models in the next section . the present study targets this general class of theories as well , and we consider here decay final states that can be regarded as representative for a wide range of scenarios . searches for dark matter decay with gamma ray observations are therefore not only phenomenologically motivated , as they can test the dark matter interpretation of the observed cosmic - ray anomalies , but they also probe theoretically compelling extensions to the standard model of particle physics . while already significant , much can be done to improve the current limits on decaying dark matter models . in this work , we show that novel and significant constraints can be placed on the dark matter decay lifetime for both quark - antiquark decay final states as well as for leptonic final states giving a good fit the cosmic - ray anomalies using _ fermi_-lat observations of isolated dark matter dominated objects like nearby clusters of galaxies , dwarf spheroidal galaxies , and m31 . in fact , the simulations of @xcite which compared the gamma - ray flux from dark matter annihilation and decay in nearby large - scale structures as reconstructed from constrained cosmological simulations predict that nearby clusters and filaments are good targets for searches for and constraints on a signal from dark matter decay , and here we derive such constraints with recent _ fermi _ observations . in 2 , we discuss and introduce selected dark matter decay models , in 3 we discuss the selection of targets and the modeling of their dark matter content , and in 4 we present our results on the constraints on the decay lifetime . while stable on the scale of cosmological structure formation , the particle making up the universal dark matter might well have a finite lifetime , @xmath3 . if this is the case , and if the dark matter particle decays into ( visible ) standard models particles , the decay process can be _ indirectly _ detected . depending on the particle lifetime , non - thermal secondary particles resulting from dark matter decay can affect key `` episodes '' in the early universe such as the synthesis of light elements or the decoupling of the cosmic microwave background ( cmb ) . in turn , light elemental abundances and the cmb spectrum can be used to set constraints on particle models where the dark matter is unstable ( see e.g. * ? ? ? there is no shortage of particle models where the universal dark matter is unstable , including unstable supersymmetric particles , moduli as dark matter particles , super - heavy dark matter candidates , axinos and sterile neutrinos @xcite . interestingly , particle candidates that would otherwise be completely un - detectable , such as gravitinos in supersymmetric extensions of the standard model , can in principle provide indirect clues on their nature if they decay @xcite . among the decay particle debris , gamma rays and antimatter ( such as energetic cosmic - ray positrons , anti - protons and anti - deuterons ) are ideal signals for indirect dark matter detection . as such , gamma - ray and cosmic - ray antimatter data have been extensively used to set constraints on the lifetime of meta - stable particle dark matter candidates @xcite . the recent detection of an exceptionally large flux of high - energy cosmic - ray positrons ( @xmath4 gev ) , compared to conventional models of galactic secondary positron production , reported by the pamela space - borne experiment @xcite prompted great interest in possible non - standard sources . another recently launched satellite experiment , _ fermi_-lat , also showed the existence of a gentle but clear feature in the spectrum of cosmic - ray electrons - plus - positrons at energies of several hundreds of gev , especially when supplemented with even higher energy data from the h.e.s.s . atmospheric cherenkov telescope @xcite . while possibly pointing to nearby astrophysical sources such as mature ( characteristic age @xmath5 yr ) pulsars ( see e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , the feature detected by _ fermi_-lat and the rise in the pamela positron fraction might share a common , more exotic origin : galactic dark matter @xcite . a large number of studies have , in fact , demonstrated that particle models can be constructed to accommodate both the positron rise and the _ fermi _ `` feature '' ( for short , we will hereafter dub these two as the `` cosmic - ray anomalies '' ) , invoking either the pair - annihilation or the decay of dark matter particles ( see e.g. * ? ? ? * ) . a dark matter annihilation interpretation of the cosmic - ray anomalies requires a very large pair - annihilation cross section , one or more orders of magnitude larger ( depending on the dark matter particle mass ) than the one that would explain via thermal decoupling the observed universal matter density @xcite . the attractiveness of the thermal relic dark matter scenario must therefore be traded - off , within this scenario , for large annihilation rates in today s cold universe , unless a very strong velocity dependence in the annihilation cross - section is invoked , ( see e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? dark matter decay as the source of the cosmic - ray lepton anomalies , on the other hand , requires the fine - tuning of the decay time - scale @xmath3 , but is otherwise a phenomenologically perfectly viable scenario . the required lifetimes are on the order of a few @xmath6 s , for particle masses in the tev range . in the broader picture of particle model - building , several scenarios predict that the symmetry that preserves the stability of the dark matter particle on scales much larger than the age of the universe is violated by sufficiently suppressed higher - dimension operators ( i.e. operators with mass - dimension greater than 4 ) that mediate the subsequent , slow particle decay . similarly to the generic prediction of proton decay , in the context of supersymmetric grand unification , operators with mass dimension 6 are expected to make supersymmetric dark matter unstable , with a natural time - scale @xmath7 where @xmath8 is the energy scale of grand unification interactions , as recently pointed out by @xcite and , subsequently , by @xcite . interestingly this yields the prediction @xmath9 which is precisely in the ballpark needed to explain the cosmic - ray anomalies . notice that gut - suppressed dimension-5 operators would produce lifetimes on the order of a second @xcite , and must thus be suppressed to avoid dark matter decaying long before structure formation . in addition to grand unification , another compelling reason to expect dark matter to decay with a lifetime on the order of @xmath10 s is provided by models , such as those elucidated in the analysis of @xcite , where the lightest supersymmetric particle is a gravitino and @xmath11-parity is ( very weakly ) broken ( see also * ? ? ? * ; * ? ? ? * ) . in this class of models , not only is the supersymmetric gravitino problem naturally solved , but the correct baryon asymmetry can naturally be generated by @xmath11-parity violating interactions via leptogenesis . another scenario that might in principle combine the origin of both the baryonic matter and the dark matter via leptogenesis , and provide a mechanism to explain neutrino masses and mixing is that of sterile neutrinos ( for a review see e.g. * ? ? ? recent analyses also pointed to hidden - sector gauge bosons and gauginos as natural particle frameworks where tev - scale particle dark matter decays with lifetimes on the order of @xmath6 s. a scenario featuring leptonic decay final states and a sufficiently long - lived dark matter as a consequence of a global symmetry broken by instanton operators originating from a non - abelian gauge group in the dark sector was also discussed in @xcite . if the decay of particle dark matter is to explain the cosmic - ray lepton anomalies , the required dominant decay final state must be close to one ( or more ) muon pairs , as shown e.g. in @xcite . nevertheless , among many of the scenarios mentioned above , the decay products will likely follow hadronization chains that are typically well reproduced in their major features by decay final states such as a quark - antiquark pair ( even though for instance in the scenario of @xcite the final state might be that resulting from @xmath12 or @xmath13 decay , the hadronic decay modes would still produce a gamma - ray spectrum qualitatively similar to that of a quark - antiquark pair ; see also @xcite for a discussion of weak corrections to annihilation or decay final states ) . for this reason , in the present study we will focus on two specific decay final states : @xmath14 as representative of those scenarios that offer a particle dark matter decay interpretation to the cosmic - ray anomalies ; and @xmath15 for more conventional decaying dark matter scenarios . for these two decaying dark matter final states , we will consider in detail the produced gamma - ray spectrum as a function of the dark matter particle mass , and model this spectrum with the dmfit package @xcite as implemented in the _ fermi_-lat data analysis pipeline . the consideration of different decay final states will be coupled to the choice of the best targets , as one particular target does not necessarily give the strongest constraints for all dark matter models . an important effect , as shown in @xcite , @xcite , and @xcite , is the diffusion of the high - energy electrons and positrons produced by dark matter decay . for sufficiently massive dark matter candidates , in fact , these particles copiously up - scatter low - energy background photons ( inverse compton ic emission ) into gamma rays detectable by the lat @xcite . the ic photon flux is however strongly dependent upon the electron diffusion and leakage out of the observed target , an effect that impacts both the spectrum and the predicted flux - intensity . we discuss this and other target - dependent considerations in the next section . a number of astrophysical targets are potentially good sources to search for a signal from dark matter annihilation or decay , including the galactic center , the milky way halo , nearby galaxies and dwarf spheroidal galaxies , nearby clusters of galaxies , and the summed contribution from unresolved extragalactic dark matter ( for a review in the context of the lat performance , see e.g. * ? ? ? underpinning which is the `` best '' target depends , however , on the type of signal ( annihilation or decay ) , on our understanding of the dark matter density in different objects , on the gamma - ray backgrounds , and on the dominant gamma - ray emission mechanism for a given dark matter particle model ( for instance prompt gamma - ray production versus secondary ic emission ) . extragalactic objects like clusters of galaxies and local group dwarf galaxies are particularly attractive targets for dark matter searches as they are highly dark matter dominated , isolated objects , and many of them lie in regions of low diffuse gamma - ray background , being located at high galactic latitudes . in addition , with the exception of m31 , the objects we discuss do not themselves host detectable gamma - ray emission , making them clean targets of study to constrain emission from dark matter . previous studies looking at constraints on dark matter models using _ fermi _ observations of dwarfs and clusters have focused on dark matter annihilation @xcite , while constraints on dark matter decay with _ fermi _ have primarily compared the observed diffuse gamma - ray backgrounds to the total summed emission expected from dark matter decay in the milky way halo or from unresolved extragalactic dark matter @xcite . in this paper , we constrain dark matter decay models using , instead , isolated extra - galactic dark matter sources , specifically nearby clusters , dwarf spheroidal satellites of the milky way , and the nearby galaxy m31 ( the andromeda galaxy ) . in some cases , these targets provide significantly stronger constraints on the dark matter decay lifetime than have been previously published , particularly for the case of nearby clusters ( see also the simulations of * ? ? ? * ) , which we discuss in the next section . clusters of galaxies are the most massive collapsed objects in the present - day universe and are highly dark matter dominated . apart from the central bright active galactic nuclei in a couple of clusters , clusters have thus far not been detected in gamma - rays , a fact that allows one to place limits on their possible dark matter related emission @xcite . we consider here the same sample of six , nearby clusters of galaxies as we studied in @xcite . these clusters are among the best candidates for dark matter searches and have published upper limits on their potential gamma - ray emission from dark matter ( annihilation ) based on the first 11 months of _ fermi_-lat observations for two illustrative ( annihilation ) final states , @xmath15 and @xmath14 , and for a grid of dark matter particle masses . remarkably , the expected gamma - ray spectral shape from dark matter decay for a given particle mass is the same as the gamma - ray spectrum from annihilation for a particle of half the mass with the same assumed final state ( although certain final states can only be available to decay processes , such as e.g. @xmath12 plus charged leptons , see e.g. * ? ? ? * ) , and we use the corresponding flux upper limits from @xcite to take into account the spectral dependence of the broad - band _ fermi_-lat upper limits . clusters are particularly good targets when considering decay to leptons , as is the case for models fitting the cosmic - ray anomalies , because in this case much of the gamma - ray emission stems from ic scattering of background radiation by the energetic @xmath16 produced @xcite . in clusters , @xmath16 lose energy through ic much faster than they can diffuse out of the system , unlike in smaller systems like dwarf galaxies @xcite . the _ fermi_-lat effective area is a strong function of energy and increases by nearly an order of magnitude from 100 mev to 1 gev , increasing more slowly at energies above 1 gev . in addition , the lat psf decreases with increasing energy ( effectively reducing the background for detection ) and the gamma - ray backgrounds are relatively soft . given the expected spectral shape for the dark matter decay models we consider , these facts mean that the tightest constraints come from the _ fermi _ data around @xmath17 gev ( for more detail see e.g. * ? ? ? * ; * ? ? ? at these energies the _ fermi_-lat point - spread function ( psf ) is around one degree , a radius which contains most of the virial mass of the clusters we consider . following @xcite , we assume an nfw profile for the cluster dark matter density distribution with cluster masses taken from @xcite and the concentration - mass relation of @xcite . the normalization of the expected gamma - ray flux from dark matter decay @xmath18 is @xmath19 where we integrate the dark matter density @xmath20 along the line of sight ( l.o.s . ) in the direction @xmath21 over a solid angle @xmath22 corresponding to a radius of one degree . the @xmath18 values for our cluster sample are listed in table 1 along with the masses , radii , and concentrations we employ . the choice of density profile and integration radius only matters at the level of a factor of a few in the derived @xmath18 since , as already mentioned , the size of the _ fermi _ psf includes much of the cluster virial mass ( see column 6 of table 1 ) and since compared to annihilation , which is proportional to density squared , decay is relatively insensitive to the inner dark matter density profile . local group dwarf spheroidal galaxies are very nearby objects which , in some cases , are even more dark matter dominated than clusters @xcite . similar to clusters , dwarf spheroidal galaxies have not been detected in gamma - rays @xcite . @xcite places limits on dark matter annihilation models based on the non - detection of eight dwarf spheroidal galaxies in the first 11 months of _ fermi_-lat data ; we will consider this sample of eight dwarfs in terms of limits on dark matter decay employing again the appropriate published _ fermi_-lat flux limits for assumed @xmath15 and @xmath14 final states and a grid of particle masses . for dwarf spheroidal galaxies , the dark matter halo mass can be estimated from the line - of - sight stellar velocity dispersion . dwarf masses are most strongly constrained near the half light radius as the mass at the half light radius , @xmath23 , is relatively insensitive to the unknown stellar velocity dispersion anisotropy @xcite ( see however also * ? ? ? outside of this radius the mass and dark matter density profile are less well constrained ( though the use of cold dark matter ( cdm ) priors reduces this uncertainty * ? ? ? the half light radii of the dwarf spheroidal galaxies we consider are smaller than the _ fermi_-lat psf at our energies of interest ( see column 5 of table 2 ) , so we will conservatively consider @xmath23 , taken from @xcite , as the lower limit on the dark matter mass probed by the _ fermi _ observations . we list the corresponding values of @xmath18 as well as @xmath23 and @xmath24 in table 2 . for comparison , in figure [ fig : dwarfsys ] we show the constraints we would get for a range of plausible dwarf virial masses assuming the density profile given in figure 3 of @xcite and taking the corresponding mass within one degree . as the nearest large galaxy , m31 ( the andromeda galaxy ) is also an excellent target for dark matter decay searches , as envisioned e.g. for the case of sterile neutrinos in @xcite . a significant detection of m31 has recently been reported in two years of _ fermi_-lat data @xcite . the gamma - ray luminosity of m31 is consistent with predictions for the luminosity from cosmic ray collisions with the ism and is likely not due to dark matter . we will conservatively take the detected flux plus @xmath25 @xcite as the upper limit on the possible dark matter induced gamma - ray emission from m31 in order to place limits on the dark matter decay lifetime , though in all likelihood any dark matter signal from m31 is significantly fainter . we consider both the flux resulting from the assumption that the gamma - ray emission is point - like and from the assumption of an extended source with emission following the m31 infra - red surface brightness . to find @xmath18 , we employ the nfw dark matter density profile from @xcite based on fits to surface brightness , velocity dispersion , and rotation curve data ( see table 1 ) . we also consider the isothermal profile of @xcite , for comparison . in this section we present our results on the limits on the decay lifetime for dark matter particles as a function of mass from gamma - ray observations of clusters ( [ sec : clusters ] ) , of dwarf spheroidal galaxies ( [ sec : dsph ] ) and of m31 ( [ sec : m31 ] ) . as a rule of thumb , it is simple to give an estimate for the expected _ fermi_-lat limits for the case of a standard decaying wimp : @xmath26 where @xmath27 is the limit on the gamma - ray flux from the observed target , and @xmath28 the integrated number of photons per decay within the relevant _ fermi_-lat energy range ( this number depends on the decay final state as well as on the decaying particle mass @xmath29 ) . figure 1 shows the limits we derive for the possible dark matter decay lifetime assuming a @xmath15 final state ( left panel ) and @xmath14 final state ( right panel ) derived from the 95% confidence level upper limits on the gamma - ray flux from nearby clusters of galaxies presented in @xcite for 11 months of _ fermi_-lat observations . for comparison , we show previous constraints on the decay lifetime derived from _ fermi _ all - sky maps and considering the dark matter contribution from both the milky way halo and unresolved extragalactic dark matter from @xcite ( see also * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? the shape of the constraints on the lifetime versus mass can be understood by considering the combined effect of how the gamma - ray flux depends on the decaying particle mass and of how the flux sensitivity depends on the resulting gamma - ray spectrum . increasing the particle mass @xmath29 , the particle number density , and hence the number of decays , decreases as @xmath30 . however , the number of photons per decay _ increases _ with @xmath29 more energy density is dumped into the decay products overall as the decaying particle mass is increased . the photons relevant to derive constraints from _ fermi_-lat data , however , are restricted to a specific energy range . particularly when secondary photons from inverse compton scattering are included , deriving a simple functional form for the integrated number of photons @xmath31 within the _ fermi_-lat energy range from the decay of a particle of mass @xmath29 versus the particle mass is non - trivial . suffice it to say that , neglecting ic photons @xmath28 grows with @xmath29 slower than linearly . this implies that , when using a _ fixed _ flux upper limit ( as we do for the case of m31 , see below ) we expect the limits on the dark matter lifetime to decrease as @xmath32 , with @xmath33 . a more subtle effect occurs if the flux limits are calculated themselves as a function of mass , as is done in the dedicated _ fermi_-lat data analyses presented in @xcite and @xcite where the simulated gamma - ray spectra from annihilation of particles into given final states were employed . the flux upper limits _ improve _ ( i.e. decrease ) with the particle mass , since most of the photons are produced at increasingly large energies where the instrumental point - spread function and effective area improve ( see e.g. fig . 2 in * ? ? ? * ) . this effect is even more acute when photons from ic are accounted for in this case for large enough masses @xmath34 gev , photons start to enter the _ fermi_-lat energy range and to dominate the total photon count ( see fig . 2 in * ? as a result of the combination of the flux limits improving with @xmath29 and of the photons counts decreasing with @xmath29 , we expect a maximum , at some intermediate value of the particle mass , for the constraints on the lifetime . this indeed what we find , unless , as in the case of m31 , we simply employ one fixed flux limit value . in the @xmath15 panel , the yellow band to the left shows a parameter space range corresponding to theoretical predictions based on the gravitino dark matter model of @xcite and of @xcite , where we posit that the gravitino lifetime @xmath35 goes as @xmath36 with @xmath37 the gravitino mass , and the model - dependent parameter @xmath38 ( in the figure , the left boundary corresponds to @xmath39 and the right boundary to @xmath40 ) . the right - most yellow band , in the @xmath15 panel , indicates a range for the mass - scale @xmath41 of a `` new physics '' dimension-6 operator that would mediate dark matter decay in the context of supersymmetric grand unification , according to the rate outlined in eq . ( [ eq : gut ] ) with @xmath42 . we take as an indicative range @xmath43 . in the right panel , we show decay lifetimes giving a good fit to the pamela positron excess and the combination of the pamela position fraction and the feature observed in the @xmath44 spectrum measured by the _ fermi_-lat ( for details on the determination of these contours , see * ? ? ? we also shade in yellow the more `` optimistic '' limits calculated by @xcite , where a specific , albeit rather arbitrary , galactic gamma - ray diffuse model is assumed and subtracted off of the _ fermi _ data . it can be seen from figure 1 that clusters of galaxies give significantly stronger constraints on the dark matter decay lifetime , both for a @xmath15 and a @xmath14 final state and for a large range of particle masses , than have previously been derived from gamma - ray observations . specifically , for a @xmath15 final state the decay lifetime is limited to be above @xmath45 secs for particle masses in the very broad mass range between 10 gev and 1 tev , while previous limits were about a factor of five worse worse . limits on specific particle theories with meta - stable dark matter , such as those shown in the figure with shaded yellow contours , also improve accordingly . for a @xmath14 final state ( right panel ) , the non - detection of clusters by _ fermi _ excludes essentially the entire best fit region for dark matter decay models fitting the combination of the pamela positron fraction and the _ fermi_-lat @xmath44 spectrum and much of the region fitting the pamela data alone . notice that this is in line with the predictions of @xcite , where the effect of substructures was also extensively discussed . only models where the dark matter particle is significantly lighter than 2 tev seem to pass the test of gamma ray emission from galaxy clusters . these results , therefore , cast doubt on the dark matter decay explanation of the rising positron fraction at high energies , particularly if this is taken jointly as an explanation of the _ fermi_-lat electron - positron spectrum . similar to dark matter annihilation , the strongest constraints come from gamma - ray observations of nearby groups or poor clusters like fornax , m49 , and ngc4636 , but all of the clusters in our sample lead to an improvement in the dark matter decay constraints . we note from table 1 that the ranking in terms of the brightest signal from dark matter of the various clusters is quite different for the cases of decay and annihilation , although the brightest object is in both cases the fornax cluster . notice that some uncertainty is present in the determination of the coefficient @xmath18 , as we discuss below . in figure [ fig : clussys ] , we compare the constraints on the dark matter decay lifetime determined above ( blue lines ) using an x - ray derived dark matter density profile to alternate choices for the density profile or cluster mass as an indication of the potential systematic error in our constraints for the fornax cluster ( left panel ) , which has the highest @xmath18 of the cluster sample , and for the coma cluster ( right panel ) , which has the highest mass of the cluster sample . we only explicitly show decay to @xmath14 , but the same relative offsets hold for other final states . for the fornax cluster , we plot the limits for two alternate estimates of the cluster mass : the mass from @xcite based on the velocity dispersion of cluster member galaxies and the mass from the independent x - ray analysis of @xcite based on rosat hri observations ( mass within one degree given from their published mass profiles ) . the dark matter density profile in the central fornax galaxy , ngc1399 , has also been investigated based on the radial velocities of globular clusters ( gc ) , but gc dynamics ( particularly the red gc ) probe primarily the central galaxy itself and do not reliably give the potential of the cluster as a whole . the mass of the coma cluster , the most massive and most distant cluster in our sample , has also been estimated based on weak lensing of background galaxies . in the right panel of figure [ fig : clussys ] , we show the change in the constraints on the dark matter decay lifetime for three different weak lensing estimates of the coma cluster mass as well as for the cluster mass estimated based on the velocity dispersion of the cluster galaxies @xcite . for both fornax and coma , the constraints for a variety of mass estimates change by less than a factor of two from the nominal constraints shown in figure 1 . even the most conservative fornax constraints represent a significant improvement over previous limits and exclude large portions of the parameter space that could explain the cosmic ray anomalies . for our nominal constraints we employ the mass - concentration relation of @xcite , because this relation was derived based on x - ray cluster observations ; however , we note that for a variety of mass - concentration relations from both simulations and observations @xcite the @xmath18 of fornax would change by less than 20% . a full accounting for systematic effects for all clusters in the sample is not statistically not feasible ( the observations not being directly comparable ) , nor do we wish to debate the merits of one set of observations over another . however , figure [ fig : clussys ] leads us to conclude that our constraints are robust within roughly a factor of two , and that they are not systematically biased high or low . as the fermi non - detection of most of the clusters in the sample excludes a significant fraction of the decay lifetime and particle mass combinations needed for a joint dark matter decay explanation to the cosmic - ray anomalies , our results lead to the conclusion that this possibility appears disfavored . figure [ fig : dwarfs ] shows the limits on the dark matter decay lifetime for nearby dwarf spheroidal galaxies using the _ fermi _ gamma - ray limits published in @xcite . for the dwarf constraints with a @xmath14 final state ( right panel ) , we conservatively only include gamma - ray emission from final state radiation and do not include ic emission , given the unknown effects of diffusion on @xmath1 and @xmath2 in dwarf scale systems @xcite . below we discuss how these constraints would improve with the inclusion of ic gamma - ray emission and specific assumptions for cosmic - ray diffusion . compared to clusters , the non - detection of dwarf spheroidal galaxies in gamma - rays gives weaker constraints at least for the conservative limits we derive assuming a mass of @xmath23 and no ic emission . however , at low masses the constraints still improve upon previous limits from the literature , in particular those we derive from the ursa minor dwarf galaxy . if we relax our conservative assumptions on gamma - ray emission from high - energy muon pairs , the constraints shown in the right panel of figure [ fig : dwarfs ] become considerably stronger , as shown in figure [ fig : dwarfsys ] for ursa minor ( we employ this specific dwarf galaxy as it gives the strongest constraints among the dwarf sample ) . in the left panel of figure [ fig : dwarfsys ] , we consider the limits on the dark matter decay lifetime if we include the expected ic gamma - ray emission for a @xmath14 decay final state for different assumed diffusion coefficients . following @xcite , the blue lines show the constraints for diffusion coefficients of @xmath46 @xmath47 s@xmath48 ( dot - dashed line ) and @xmath49 @xmath47 s@xmath48 ( dashed line ) which bound the values found for the milky way . in both cases , the energy dependence of diffusion is assumed to have a power - law dependence given by @xmath50 . the red line shows the constraints in the limit of no diffusion ( i.e. in the limit where the electrons and positrons loose all of their energy via ic prior to diffusing away from the system ) . the inclusion of ic emission significantly strengthens the constraints from ursa minor , but for reasonable diffusion coefficients these are still weaker than previous limits on the decay lifetime . in addition for our nominal constraints in figure [ fig : dwarfs ] , we conservatively assumed dwarf masses equal to @xmath23 even though the dwarf half - light radii are smaller than 1 degree . in the right panel of figure [ fig : dwarfsys ] , we compare the constraints for ursa minor for different assumptions on the mass profile . for illustration we show the limits on the decay lifetime for a @xmath14 final state with a diffusion coefficient of @xmath46 @xmath47 s@xmath48 , but of course the same relative offset also holds for other cases . employing the best - fit nfw profile parameters from fits to the stellar velocity data for ursa minor published in @xcite to calculate @xmath18 for a radius of 1 degree yields the dot - dashed line in figure [ fig : dwarfsys ] ( right panel ) , a factor of a few stronger than for @xmath23 . @xcite find that all of the milky way dwarf spheroidals are consistent with having the same total halo mass ( within the scatter ) of around @xmath51 m@xmath52 for a @xmath53cdm cosmology . looking for example at their figure 3 , for a range of two orders of magnitude in total halo mass between @xmath54 m@xmath52 and @xmath55 m@xmath52 and for an nfw profile consistent with a wmap5 cosmology , would give a mass within 1 degree for ursa minor between a few times @xmath56 m@xmath52 and a few times @xmath57 m@xmath52 . the decay constraints for this range of masses is reflected in the two dashed lines in figure [ fig : dwarfsys ] , where we show the results for the specific values of @xmath58 m@xmath52 and @xmath59 . depending on the assumptions on mass and diffusion , the ursa minor constraints become comparable to or stronger than previous limits . although faintly , the nearby spiral galaxy m31 is detected at gamma - ray frequencies , as recently reported by the _ fermi_-lat collaboration @xcite . the detected gamma - ray emission is compatible with a point - source , although a slightly higher test - statistics was reported if a spatial template based on the infra - red surface brightness was employed @xcite . here , to set constraints on the emission from dark matter decay we simply consider the 2-@xmath60 upper limit on the observed emission , conservatively summing the statistical and systematic uncertainties quoted by the _ fermi_-lat collaboration . a more in - depth study would resort to either attempting to subtract the astrophysical emission from m31 as expected from cosmic - ray interaction with the inter - stellar medium ( in analogy to the diffuse gamma - ray emission from our own milky way ) , or to utilize , for instance , a region of interest that cuts out the gamma - ray emitting region ( e.g. an annulus around the m31 central region ) . we defer such a study for a forthcoming analysis . m31 is an extremely interesting object to search for a signal from dark matter decay , as was for instance suggested , for the case of sterile neutrinos ( producing a monochromatic line at x - ray frequencies from the two - body decay into an active neutrino and a photon ) by @xcite . to estimate the object s brightness in gamma rays from dark matter decay , we employ the nfw dark matter density profile proposed by @xcite : table 1 shows that we obtain a @xmath18 value ( 1 degree radius ) for m31 which exceeds by almost a factor 2 the largest normalization factor for galaxy clusters , and by more than an order of magnitude the largest one we find among dwarf galaxies . despite being a potentially extremely bright source of gamma rays from dark matter decay , the constraints we derive from m31 are not as strong as those we found for clusters the reason evidently being that m31 is actually detected as a gamma - ray source , and we conservatively do not attempt to subtract off any astrophysical background . the left panel of figure [ fig : m31 ] illustrates the constraints we obtain for a @xmath15 decay final state . the black line refers to the nfw profile with the gamma - ray flux obtained with the assumption of a point - like source and @xmath18 as listed in table 1 . the green dashed line assumes the isothermal profile of @xcite and the same gamma - ray flux limit the closeness of the two lines illustrates the small uncertainty that stems from using two entirely different functional forms for the dark matter density profile . finally , for the blue dotted line we employ the extended source _ fermi_-lat flux and assume that the relevant mass coincides with the virial mass of m31 ( which we set to @xmath61 , from * ? ? ? the limits we obtain , while an improvement over what we get from dwarf galaxies and over previously available limits , are a factor of a few worse than those implied by the non - detection of galaxy clusters for the @xmath15 final state . the right panel of figure [ fig : m31 ] shows our results for the @xmath14 final state . we compare again with the regions favored by a dark matter decay interpretation of the cosmic - ray anomalies ( green and red shaded regions ) and with the current constraints ( grey region ) . the black lines refer to the nfw plus point - like source assumption , while the red lines refer to the case where we employ the virial mass and the extended source gamma - ray flux . given the relevance of diffusion for cosmic - ray electrons and positrons in the milky way , we do expect leakage of secondary @xmath16 out of m31 as well , and we model this process again in analogy to our own galaxy . we show with solid lines our results for the relatively small diffusion coefficient @xmath62 @xmath47/s , and with dashed lines the more conservative case of @xmath63 @xmath47/s ( as for dwarf galaxies , we set the rigidity dependence power to 1/3 ) . in either case , the limits we obtain do not improve over previous constraints nor do they approach the phenomenologically relevant regions for the cosmic - ray anomalies . in this paper , we analyzed how early _ fermi _ limits on the gamma - ray flux from clusters of galaxies , from nearby dwarf galaxies and the _ fermi _ detection of m31 impact the possible lifetime of a meta - stable particle dark matter . we showed that limits from dwarf galaxies and from m31 improve over previous constraints , for certain ranges of masses and final states . the best limits are however derived in the case of galaxy clusters , where we obtain an improvement of about a factor of five on the allowed decay lifetime for hadronic decay final states for a wide range of particle dark matter masses between 10 gev and 1 tev . furthermore , models where dark matter decays into muon pairs and explains the so - called cosmic - ray anomalies ( the positron excess reported by pamela and the feature observed by _ fermi _ in the electron - positron spectrum ) are found to over - produce gamma rays from clusters of galaxies compared to _ fermi _ observations for large regions of the best - fitting parameter space . while we can not completely exclude the possibility , our results imply that the dark matter decay interpretation of the cosmic - ray anomalies is in tension with gamma ray observations , especially if both the _ fermi _ and pamela anomalies are jointly explained , and this scenario may be ruled out completely if fermi does not detect gamma - ray emission from clusters in the next few years . _ fermi_-lat continues to take data in all - sky survey mode , and in the future longer exposures and a better understanding of the astrophysical backgrounds will continue to improve constraints on dark matter models and perhaps even reveal a signal from this elusive component of the universe . this work is partly supported by nasa grants nnx09at96 g and nnx09at83 g . sp also acknowledges support from the national science foundation , award phy-0757911 - 001 , and from an outstanding junior investigator award from the department of energy , de - fg02 - 04er41286 . ld was supported by the uc leads program . lccccccc fornax & 18.9 & 1.42 & 8.9 & 1.35 & 0.24 & 18.4 & 1 + coma & 94.9 & 19.38 & 5.7 & 3.22 & 0.50 & 16.6 & 5 + centaurus & 42.2 & 3.78 & 7.5 & 1.87 & 0.39 & 13.7 & 4 + m49 & 18.1 & 0.65 & 10.2 & 1.04 & 0.30 & 11.1 & 2 + awm7 & 70.4 & 6.08 & 7.0 & 2.19 & 0.55 & 10.2 & 6 + ngc4636 & 15.2 & 0.35 & 11.4 & 0.85 & 0.31 & 8.88 & 3 + m 31 & 0.77 & 0.0068 & 22.0 & 0.18 & 0.075 & 36.4 & + lcccccc ursa minor & 66 & 5.56 & 588 & 2.0 & 1.50 & 2 + ursa major ii & 30 & 0.79 & 184 & 2.8 & 1.03 & 3 + bootes i & 62 & 2.36 & 322 & 3.4 & 0.72 & 5 + sextans & 86 & 3.49 & 1019 & 1.5 & 0.55 & 7 + fornax & 138 & 7.39 & 944 & 2.6 & 0.45 & 8 + draco & 76 & 2.11 & 291 & 4.6 & 0.43 & 1 + sculptor & 79 & 2.25 & 375 & 3.7 & 0.42 & 4 + coma berenices & 44 & 0.197 & 100 & 7.7 & 0.12 & 5 +	we analyze the impact of _ fermi _ gamma - ray observations ( primarily non - detections ) of selected nearby galaxies , including dwarf spheroidals , and of clusters of galaxies on decaying dark matter models . we show that the fact that galaxy clusters do not shine in gamma rays puts the most stringent limits available to - date on the lifetime of dark matter particles for a wide range of particle masses and decay final states . in particular , our results put strong constraints on the possibility of ascribing to decaying dark matter both the increasing positron fraction reported by pamela and the high - energy feature in the electron - positron spectrum measured by _ fermi_. observations of nearby dwarf galaxies and of the andromeda galaxy ( m31 ) do not provide as strong limits as those from galaxy clusters , while still improving on previous constraints in some cases .
this paper proposes a new perspective to maximum likelihood(ml ) decoding in error - correcting codes as rational maps and shows some relationships between coding theory and dynamical systems . in section [ sec : cs ] , [ sec : lc ] , and [ sec : ml ] , we explain notations and minimum prerequisites of coding theory ( e.g. , see @xcite ) . the main results are presented in section [ sec : result ] . a mathematical model of communication systems in information theory was developed by shannon @xcite . a general block diagram for visualizing the behavior of such systems is given by figure [ fig : channel ] . the source transmits a @xmath1-bit message @xmath2 to the destination via the channel , which is usually affected by noise @xmath3 . in order to recover the transmitted message at the destination under the influence of noise , we transform the message into a codeword @xmath4 by some injective mapping @xmath5 at the encoder and input it to the channel . then the decoder transforms an @xmath6-bit received sequence of letters @xmath7 by some decoding mapping @xmath8 in order to obtain the transmitted codeword at the destination . here we consider all arithmetic calculations in some finite field and in this paper we fix it as @xmath9 except for section [ sec : ag ] . as a model of channels , we deal with a binary symmetric channel ( bsc ) in this paper which is characterized by the transition probability @xmath10 ( @xmath11 ) . namely , with probability @xmath12 , the output letter is a faithful replica of the input , and with probability @xmath10 , it is the opposite of the input letter for each bit ( see figure [ fig : bsc ] ) . in particular , this is an example of memoryless channels . then , one of the main purposes of coding theory is to develop a good encoder - decoder pair @xmath13 which is robust to noise perturbations . hence , the problem is how we efficiently use the redundancy @xmath14 in this setting . a code with a linear encoding map @xmath5 is called a linear code . a codeword in a linear code can be characterized by its generator matrix @xmath15 where each element @xmath16 . therefore the set of codewords is given by @xmath17 here without loss of generality , we assume @xmath18 for all @xmath19 and @xmath20 . we call @xmath1 and @xmath6 the dimension and the length of the code , respectively . because of the linearity , it is also possible to describe @xmath21 as a kernel of a matrix @xmath22 whose @xmath23 row vectors are linearly independent and orthogonal to those of @xmath24 , i.e. , @xmath25 where @xmath26 means the transpose matrix of @xmath22 . this matrix @xmath22 is called a parity check matrix of @xmath21 . the dual code @xmath27 of @xmath21 is defined in such a way that a parity check matrix of @xmath27 is given by a generator matrix @xmath24 of @xmath21 . the hamming distance @xmath28 between two @xmath6-bit sequences @xmath29 is given by the number of positions at which the two sequences differ . the weight of an element @xmath30 is the hamming distance to @xmath31 , i.e. , @xmath32 . then the minimum distance @xmath33 of a code @xmath21 is defined by two different ways as @xmath34 here the second equality results from the linearity . it is easy to observe that the minimum distance is @xmath35 if and only if there exists a set of @xmath36 linearly dependent column vectors of @xmath22 but no set of @xmath37 linearly dependent column vectors . for a code @xmath21 with the minimum distance @xmath38 , let us set @xmath39 , where @xmath40 is the integer part of @xmath41 . then , it follows from the following observation that @xmath21 can correct @xmath42 errors : if @xmath43 and @xmath44 for some @xmath45 then @xmath46 is the only codeword with @xmath44 . in this sense , the minimum distance is one of the important parameters to measure performance of a code and is desirable to design it as large as possible for the robustness to noise . let us recall that , given a transmitted codeword @xmath46 , the conditional probability @xmath47 of a received sequence @xmath43 at the decoder is given by @xmath48 for a memoryless channel . maximum likelihood(ml ) decoding @xmath49 is given by taking the marginalization of @xmath47 for each bit . precisely speaking , for a received sequence @xmath50 , the @xmath51-th bit element @xmath52 of the decoded word @xmath53 is determined by the following rule : @xmath54 in general , for a given decoder @xmath8 , the bit error probability @xmath55 , where @xmath56 is one of the important measures of decoding performance . obviously , it is desirable to design an encoding - decoding pair whose bit error probability is as small as possible . it is known that ml decoding attains the minimum bit error probability for any encodings under the uniform distribution on @xmath57 . in this sense , ml decoding is the best for all decoding rules . however its computational cost requires at least @xmath58 operations , and it is too much to use for practical applications . from the above property of ml decoding , one of the key motivation of this work comes from the following simple question . is it possible to accurately approximate the ml decoding rules with low computational complexity ? the main results in this paper give answers to this question . let us first define for each codeword @xmath59 its codeword polynomial @xmath60 as @xmath61 then we define a rational map @xmath62 , @xmath63,$ ] by using codeword polynomials as @xmath64 where @xmath65 . this rational map plays the most important role in the paper . it is sometimes denoted by @xmath66 , when we need to emphasize the generator matrix @xmath24 of the code @xmath21 . for a sequence @xmath43 , let us take a point @xmath67 as @xmath68 where @xmath10 is the transition probability of the channel . then it is straightforward to check that @xmath69 . namely , the conditional probability of @xmath50 under a codeword @xmath59 is given by the value of the corresponding codeword polynomial @xmath70 at @xmath71 . therefore , from the construction of the rational map , ml decoding ( [ eq : bitml ] ) is equivalently given by the following rule @xmath72 in this sense , the study of ml decoding can be treated by analyzing the image of the initial point ( [ eq : ip ] ) by the rational map ( [ eq : map ] ) . some of the properties of this map in the sense of dynamical systems will be studied in detail in section [ sec : ds ] . we will also discuss in section [ sec : discussions ] that performance of a code can be explained by these properties . for the statement of the main results , we only here mention that this rational map has a fixed point @xmath73 for any generator matrix ( proposition [ pp : fp ] ) . let us denote the taylor expansion at @xmath0 by @xmath74 where @xmath75 is a vector notation of @xmath76 , @xmath77 is the jacobi matrix at @xmath0 , @xmath78 corresponds to the @xmath51-th order term , and @xmath79 means the usual order notation . the reason why we choose @xmath0 as the approximating point is related to the local dynamical property at @xmath0 and will be explained in section [ sec : p ] . by truncating higher oder terms @xmath79 in ( [ eq : taylor ] ) and denoting it as @xmath80 we can define the @xmath81-th approximation of ml decoding by replacing the map @xmath82 in ( [ eq : mlrule ] ) with @xmath83 , and denote this approximate ml decoding by @xmath84 , i.e. , @xmath85 let us remark that the notations @xmath86 and @xmath87 do not explicitly express the dependence on @xmath81 for removing unnecessary confusions of subscripts . we note that there are two different viewpoints on this approximate ml decoding . one way is that , in the sense of its precision , it is preferable to have an expansion with large @xmath81 . on the other hand , from the viewpoint of low computational complexity , it is desirable to include many zero elements in higher order terms . the next theorem states a sufficient condition to satisfy these two requirements . [ thm : thm ] let @xmath88 . if any distinct @xmath81 column vectors of a generator matrix @xmath24 are linearly independent , then the taylor expansion ( [ eq : taylor ] ) at @xmath0 of the rational map ( [ eq : map ] ) takes the following form @xmath89 where the @xmath51-th coordinate @xmath90 of @xmath91 is given by @xmath92 first of all , it follows that the larger the minimum distance of the dual code @xmath27 is , the more precise approximation of ml decoding with low computational complexity we have for the code @xmath21 with the generator matrix @xmath24 . especially , we can take @xmath93 . secondly , let us consider the meaning of the approximate map @xmath83 and its approximate ml decoding @xmath87 . we note that each value @xmath94 in ( [ eq : ip ] ) for a received word @xmath43 expresses the likelihood @xmath95 . let us suppose @xmath96 . then , from the definition of @xmath97 , each term in the sum of ( [ eq : hosum ] ) satisfies @xmath98 when @xmath99 , this term decreases(increases , resp . ) the value of initial likelihood @xmath100 . in view of the decoding rule ( [ eq : mlrule ] ) , this induces @xmath52 to be decoded into @xmath101 , and this actually corresponds to the structure of the code @xmath102 appearing in @xmath103 . in this sense , the approximate map @xmath104 can be regarded as renewing the likelihood ( under suitable normalizations ) based on the code structure , and the approximate ml decoding @xmath87 judges these renewed data . from this argument , it is easy to see that a received word @xmath105 is decoded into @xmath106 , i.e. , the codeword is decoded into itself and , of course , this property should be equipped with any decoders . we also remark that theorem [ thm : thm ] can be regarded as a duality theorem in the following sense . let @xmath21 be a code whose generator(resp . parity check ) matrix is @xmath24 ( resp . @xmath22 ) . as we explained in section [ sec : lc ] , the linear independence of the column vectors of @xmath22 controls the minimum distance @xmath33 and this is an encoding property . on the other hand , theorem [ thm : thm ] shows that the linear independence of the column vectors of @xmath24 , which determines the dual minimum distance @xmath107 , controls a decoding property of ml decoding in the sense of accuracy and computational complexity . hence , we have the correspondence between @xmath108 duality and encoding / decoding duality . in corollary [ corollary : ag ] , we will consider this duality viewpoint in a setting of geometric reed - solomon / goppa codes . we show the second result of this paper about the decoding performances of the approximate ml ( [ eq : apmlrule ] ) . for this purpose , let us first examine numerical simulations of the bit error probability on the bsc with the transition probability @xmath109 . we also show numerical results on bch codes with berlekamp - massey decoding for comparison . the results are summarized in figure [ fig : berrate ] . and @xmath110 . horizontal axis : transition probability @xmath10 . vertical axis : bit error probability . @xmath111 : 3rd order approximate ml decoding . @xmath112 : berlekamp - massaey decoding . @xmath113 : ml decoding . , width=302 ] and @xmath110 . horizontal axis : transition probability @xmath10 . vertical axis : bit error probability . @xmath111 : 3rd order approximate ml decoding . @xmath112 : berlekamp - massaey decoding . @xmath113 : ml decoding . , width=302 ] here , the horizontal axis is the code rate @xmath114 , and the vertical axis is the bit error probability . the plots @xmath115 ( @xmath111 resp . ) correspond to the 2nd ( 3rd resp . ) order approximate ml ( [ eq : apmlrule ] ) , and @xmath112 are the results on several bch codes ( @xmath116 ) with berlekamp - messey decodings . for the proposed method , we randomly construct a systematic generator matrix in such a way that each column except for the systematic part has the same weight ( i.e. number of non - zero elements ) @xmath117 . to be more specific , the submatrix composed by the first @xmath1 columns of the generator matrix is set to be an identity matrix in order to make the code systematic , while the rest of the generator matrix is made up of @xmath118 random matrices generated by random permutations of columns of a circulant matrix , whose first column is given by @xmath119 the reason for using random codings is that we want to investigate average behaviors of the decoding performance , and , for this purpose , we do not put unnecessary additional structure at encodings . the number of matrices added after the systematic part depends on the code rate , and the plot for each code rate corresponds to the best result obtained out of about 100 realizations of the generator matrix . also , we have employed @xmath120 and 3 for the generator matrices of the 3rd and the 2nd order approximate ml , respectively . moreover , the length of the codewords @xmath6 are assumed to be up to 512 . from figure [ fig : berrate ] , we can see that the proposed method with the 3rd order approximate ml ( @xmath111 ) achieves better performance than that of bch codes with berlekamp - massey(@xmath112 ) . it should be also noticed that the decoding performance is improved a lot from the 2nd order to the 3rd order approximation . this improvement is reasonable because of the meaning of the taylor expansion . next , let us directly compare the decoding performances among ml , approximate ml ( 3rd order ) and berlekamp - massey by applying them on the same bch code ( @xmath121 , @xmath110 ) . the result on the bit error probability with respect to transition probability is shown in figure [ fig : berprob ] . this figure clearly shows that the performance of the 3rd order approximate ml decoding is far better than that of berlekamp - massey decoding ( e.g. , improvement of double - digit at @xmath122 ) . furthermore , it should be noted that the 3rd order approximate ml decoding achieves a very close bit error performance to that of ml decoding . although we have not mathematically confirmed the computational complexity of the proposed approach , the computational time of the approximate ml ( 3rd order ) is much faster than ml decoding . this fact about low computational complexity of the approximate ml is explained as follows : non - zero higher order terms in ( [ eq : hosum ] ) appear as a result of linear dependent relations of column vectors of @xmath24 , however , linear dependences require high codimentional scenario . hence , most of the higher order terms become zeros . as a result , the computational complexity for the approximate ml , which is determined by the number of nonzero terms in the expansion , becomes small . in conclusion , these numerical simulations suggest that the 3rd order approximate ml decoding approximates ml decoding very well with low computational complexity . we notice that the encodings examined here are random codings . hence , we can expect to obtain better bit error performance by introducing certain structure on encodings suitable to this proposed decoding rule , or much more exhaustive search of random codes . one of the possibility will be the combination with theorem [ thm : thm ] . on the other hand , it is also possible to consider suitable encoding rules from the viewpoint of dynamical systems via rational maps ( [ eq : map ] ) . this issue is discussed in detail in section [ sec : discussions ] . in any case , finding suitable encoding structure for the proposed decodings is one of the important future problem . the paper is organized as follows . in section [ sec : ds ] , we study properties of the rational map ( [ eq : map ] ) in view of a discrete dynamical system and show some relationships to coding theory . we also show that this discrete dynamical system is related to a continuous gradient dynamical system @xmath123 . section [ sec : taylor ] deals with relationship between a generator matrix of a code and its taylor expansion ( [ eq : taylor ] ) . the proof of theorem [ thm : thm ] , which is a direct consequence of proposition [ pp : jacobi ] and proposition [ pp : fl0 ] , is shown in this section . in section [ sec : ag ] , we apply theorem [ thm : thm ] to geometric reed - solomon / goppa codes in corollary [ corollary : ag ] with a simple example by using the hermitian curve . finally , we discuss the future problems on this subject as an intersection of dynamical systems and coding theory . in this section , we discuss ml decoding from dynamical systems viewpoints . let @xmath24 be a @xmath124 generator matrix for a code @xmath21 . we begin with showing the following easy consequence of linear codes , which will be used frequently throughout the paper . [ lemma : gi ] for @xmath125 , let us denote subcodes of @xmath21 with @xmath126 and @xmath127 respectively by @xmath128 then @xmath129 . from the assumption on the generator matrix , we have @xmath18 and let @xmath130 be the number of 1 in @xmath131 . then we have @xmath132 where the symbol @xmath133 means the number of combinations for taking @xmath134 elements from @xmath81 elements . however these summations of combinations are obviously equal because @xmath135 therefore @xmath129 . next , let us characterize the codewords in @xmath21 and the non - codewords in @xmath136 by means of the rational map ( [ eq : map ] ) . it should be remarked that a codeword polynomial @xmath70 for @xmath65 with @xmath137 , takes its value @xmath138 here we identify a point @xmath139 @xmath140 , with a point @xmath141 by a natural inclusion @xmath142 , and this convention will be used frequently in the paper . we also define a point @xmath143 as a pole of the rational map ( [ eq : map ] ) if @xmath144 . the boundary and the interior of a set @xmath145 are denoted by @xmath146 and @xmath147 , respectively . [ pp : fp ] the followings hold for the rational map ( [ eq : map ] ) : 1 . @xmath148 is a fixed point . 2 . let @xmath141 . then @xmath149 is a fixed point if and only if @xmath150 . the set of poles is given by @xmath151 especially , @xmath152 . for the statement 1 , let us note that @xmath153 for each codeword @xmath45 . then lemma [ lemma : gi ] leads to @xmath154 for the statement 2 , let us suppose @xmath150 . then , from the remark before this proposition , @xmath155 if @xmath156 for each @xmath51 . hence @xmath157 . on the other hand , if @xmath158 , then @xmath144 . it means that @xmath149 is a pole and can not be a fixed point . for the statement 3 , let us first note that @xmath159 if and only if @xmath160 for any codeword @xmath45 , because @xmath161 for @xmath143 . therefore @xmath162 , since @xmath163 for @xmath164 . let @xmath165 such that @xmath166 and @xmath167 for @xmath168 . then , if there exists a codeword @xmath45 such that @xmath169 , then the value of its corresponding codeword polynomial is @xmath163 . so , @xmath170 . on the other hand , if there is no such codeword , then @xmath144 , it means @xmath159 . from this proposition , the rational map ( [ eq : map ] ) has information of not only all codewords @xmath21 as fixed points but also non - codewords @xmath171 as poles . we call these fixed points codeword fixed points . the following proposition shows that all of the codeword fixed points are stable . [ pp : stable ] let a parity check matrix do not have zero column vectors , i.e. , there exists no codeword with weight 1 . let @xmath149 be a codeword fixed point . then the jacobi matrix of the rational map ( [ eq : map ] ) at @xmath149 is the zero matrix . hence , @xmath149 is a stable fixed point . let us denote the @xmath51-th element of the rational map ( [ eq : map ] ) by @xmath172 and denote the derivatives of @xmath173 and @xmath22 with respect to @xmath174 by @xmath175 and @xmath176 , respectively , for the simplicity of notations . in what follows , we will also use these notations for higher order derivatives in a similar way ( like @xmath177 ) . then the derivative @xmath178 is given by @xmath179 since @xmath150 , we have @xmath180 . let us consider the two cases @xmath181 and @xmath182 , separately . for @xmath181 , we have @xmath183 . similarly we have @xmath184 if @xmath185 . hence , @xmath186 in this case . on the other hand , if @xmath187 , from the assumption on the parity check matrix , we have no codeword @xmath188 such that the only difference from @xmath149 occurs at the @xmath51-th bit element , i.e. , @xmath189 for @xmath190 but @xmath191 . therefore , @xmath184 , and it again leads to @xmath186 . the case @xmath182 can be proven by the similar way . let us next discuss properties of the fixed point @xmath192 . [ pp : jacobi ] let @xmath193 be a generator matrix . then , the jacobi matrix @xmath77 of the rational map ( [ eq : map ] ) at @xmath0 is determined by @xmath194 for all @xmath195 . for the proof of this proposition , we need the following lemma . [ lemma : gij ] for @xmath196 with @xmath185 , let us consider the following subcodes of @xmath21 @xmath197 then the followings hold 1 . if @xmath198 , then @xmath199 2 . if @xmath200 , then @xmath201 the statement is trivial when @xmath198 , so we suppose @xmath200 . in the following we adopt the @xmath202 arithmetic for elements in @xmath203 . we can express @xmath131 , by using some permutations of rows if necessary , as follows @xmath204 in the following we only deal with the case @xmath205 , but the modification to the case @xmath206 is trivial . now we have the following two cases + case i : there exists @xmath207 such that @xmath208 , + case ii : not case i , i.e. , @xmath209 for all @xmath207 . + in the case i , let us fix @xmath210 in a message @xmath2 with @xmath211 , which corresponds to @xmath127 , and consider the numbers of codewords with @xmath212 and @xmath213 from the remaining message bits @xmath214 . then , from the assumption , there exists at least one non zero element in @xmath215 . hence , the same argument in lemma [ lemma : gi ] shows that the numbers of codewords with @xmath212 and @xmath213 under a fixed @xmath216 are the same . by considering all the possibilities of @xmath210 with @xmath217 , it gives @xmath201 next , let us consider the case ii . again by using a permutation if necessary , we have the following expressions @xmath218 where @xmath219 and @xmath220 because of @xmath200 . then we have @xmath221 however , the same argument in lemma [ lemma : gi ] implies @xmath201 the @xmath222 element in the jacobi matrix @xmath77 is given by @xmath223 from lemma [ lemma : gi ] , @xmath224 because @xmath225 where the first term comes from the codewords with @xmath213 and the second term comes from the codewords with @xmath212 . it is also easy to observe that @xmath226 , and @xmath227 . therefore the diagonal elements are @xmath228 . next , let us consider the case @xmath185 . in this case , if we have @xmath198 , then , from lemma [ lemma : gij ] , @xmath229 . on the other hand , if we have @xmath200 , then @xmath230 . this concludes the proof of proposition [ pp : jacobi ] . two corollaries follow from proposition [ pp : jacobi ] which characterize the eigenvalues and the eigenvectors of the jacobi matrix @xmath77 , and it clearly determines the stability and the stable / unstable subspaces of @xmath0 . to this end , let us denote by @xmath231 a graph whose adjacent matrix is @xmath77 . namely , the nodes of @xmath231 are @xmath232 , and an undirected edge @xmath222 appears in @xmath231 if and only if @xmath233 . [ corollary : eigenvalue ] suppose the graph @xmath231 is decomposed into @xmath81 connected components @xmath234 let @xmath235 be the number of nodes in the component @xmath236 , @xmath237 . then all the eigenvalues of @xmath77 are given by @xmath238 where the eigenvalues @xmath239 are simple and the @xmath31 eigenvalue has @xmath240 multiplicity . from proposition [ pp : jacobi ] , any two nodes in a same connected component have an edge between them . hence , it is possible to transform @xmath77 into the following block diagonal matrix @xmath241 where @xmath242 is determined by compositions of column switching elementary matrices , and @xmath243 is an @xmath244 matrix all of whose elements are 1 . the statement of the corollary follows immediately . from now on , we treat a jacobi matrix of the block diagonal form ( [ eq : elementarymatrix ] ) . obviously , it gives no restriction since , if necessary , we can appropriately permute columns of the original generator matrix in advance . let us denote the set of eigenvalues of @xmath77 derived in corollary [ corollary : eigenvalue ] by @xmath245 where the successive @xmath246 after each @xmath235 has @xmath247 elements . in case of @xmath248 , we ignore the successive @xmath249 for @xmath235 ( i.e. , @xmath250 ) . [ corollary : eigenvector ] the jacobi matrix @xmath77 is diagonalizable . furthermore , the corresponding eigenvectors @xmath251 for ( [ eq : eigenvalue ] ) under this ordering are given by the following @xmath252 where @xmath253 here the vectors @xmath254 have @xmath235 elements and the first element starts at the @xmath255-th row . the bold type @xmath256 expresses that the remaining elements in @xmath257 are @xmath31 . in case of @xmath248 , we only have @xmath258 ( with @xmath259 ) . it is obvious from corollary [ corollary : eigenvalue ] . we finally mention a relationship to a continuous gradient dynamical system . let us denote by @xmath260 a polynomial obtained by removing @xmath261 from a codeword polynomial @xmath70 of @xmath45 . by using this notation , the map ( [ eq : map ] ) can be also described as @xmath262 then it follows that @xmath263 this proves the following proposition . [ pp : gradient ] the rational map ( [ eq : map ] ) maps a point @xmath264 to the direction of @xmath265 with a contraction rate @xmath266 for each element @xmath267 . especially , @xmath264 is a fixed point of ( [ eq : map ] ) if and only if it is a fixed point of the continuous gradient dynamical system @xmath123 . next , we study a relationship between higher order terms in ( [ eq : taylor ] ) and a generator matrix @xmath24 , and prove theorem [ thm : thm ] . for this purpose , the key proposition is given as follows . [ pp : fl0 ] let @xmath88 . if any distinct @xmath268 column vectors of @xmath24 are linearly independent , then @xmath269 . let us denote a higher order derivative of an @xmath51-th element @xmath270 with respect to variables @xmath271 by @xmath272 for the proof of proposition [ pp : fl0 ] , we need to study higher order derivatives @xmath273 and @xmath274 . let us at first focus on higher order derivatives @xmath273 . we begin with the following observation , which characterizes the numbers of subcodes by means of column vectors of @xmath24 . it is noted that we adopt the @xmath202 arithmetic for elements in @xmath203 . [ lemma : c01 ] suppose @xmath275 are distinct indices , and let @xmath276 be subcodes in @xmath21 . then the followings hold 1 . if @xmath277 , then @xmath278 . if @xmath279 , then @xmath280 and @xmath281 . by a suitable bit permutation , if necessary , the sum of @xmath282 can be expressed as @xmath283 then an original message @xmath284 and its codeword @xmath285 satisfy the following @xmath286 in the case 1 ( @xmath287 ) , it leads to @xmath288 , so the conclusion follows from the same argument in lemma [ lemma : gi ] . the case 2 is trivial from the above expression of @xmath289 . the following lemma classifies the value @xmath273 based on the column vectors of @xmath24 . [ lemma : h ] let @xmath88 . then @xmath290 if either of 1 there exist same indices in @xmath291 2 . @xmath277 is satisfied . otherwise , that is @xmath291 are all distinct and @xmath292 , @xmath293 the condition 1 immediately implies @xmath290 since the degree of each variable @xmath294 in @xmath295 is 1 . hence we assume all the indices are distinct . let us define the following subcodes @xmath296 then @xmath273 can be expressed by @xmath297 suppose @xmath277 . then , by lemma [ lemma : c01 ] , we have @xmath298 . on the other hand , when @xmath81 is odd ( or even , resp . ) , @xmath299 therefore it concludes @xmath290 . the statement for @xmath292 is similarly derived from lemma [ lemma : c01 ] . next , we try to classify the value @xmath274 . [ lemma : ic01 ] suppose @xmath300 are distinct indices and let us define two subcodes in @xmath21 by @xmath301 then the following classification holds 1 . if @xmath292 , then @xmath302 , @xmath303 . 2 . if @xmath304 , then @xmath305 . 3 . if @xmath306 , then @xmath307 , @xmath308 . since we have @xmath309 the case 1 and 3 are trivial . so , let us assume @xmath304 . the proof is similar to that of lemma [ lemma : gij ] . by using a suitable permutation , let us express @xmath310 as follows @xmath311 here we only deal with the case @xmath205 again , since the modification for @xmath312 follows immediately from the following case ii . we have two situations + case i : there exists @xmath313 such that @xmath314 . + case ii : not case i , i.e. , @xmath315 for all @xmath313 . + in case i , let us fix @xmath316 with @xmath211 , which corresponds to @xmath317 , and consider the numbers of codewords with @xmath318 or @xmath319 for the remaining message bits @xmath320 . from the assumption , the @xmath321-th element of the vector @xmath322 is 1 , and , by applying lemma [ lemma : c01 ] to the subvector from @xmath323-th to @xmath1-th elements , the numbers of codewords with @xmath318 or @xmath319 are the same for each @xmath316 . hence we have @xmath305 . the proof for case ii is almost parallel to that of lemma [ lemma : gij ] , so we omit it . let us introduce the following notations , which are similar to those in the proof of lemma [ lemma : h ] , @xmath324 then , for odd @xmath81 ( or even @xmath81 , resp . ) , we have @xmath325 [ lemma : ip ] let @xmath88 . then @xmath274 is classified as @xmath326 where each condition is given by + c0 : there exist same indices in @xmath291 + c1 : @xmath327 and @xmath304 ( here @xmath327 means not c0 " ) + c2 : @xmath327 and @xmath328 + c3 : @xmath327 and @xmath292 the condition c0 immediately implies @xmath329 . hence we assume all the indices are distinct . the remaining proof follows directly from lemma [ lemma : ic01 ] for each case . first of all , let us study the case @xmath330 . by using the notations introduced before the lemma , we have @xmath331 therefore the condition c1 implies @xmath329 by lemma [ lemma : ic01 ] . on the other hand , if we assume the condition c2 , then @xmath307 and @xmath308 from lemma [ lemma : ic01 ] . if @xmath81 is even , @xmath332 . similarly , if @xmath81 is odd , @xmath333 . hence , we have @xmath334 for the condition c2 . for the condition c3 , the role of @xmath335 and @xmath336 changes each other from lemma [ lemma : ic01 ] , so it just leads to the opposite sign in @xmath274 to that for the condition c2 . next , let us study the case @xmath337 . without loss of generality , let us suppose @xmath338 . then we have @xmath339 therefore the condition c1 implies @xmath329 by lemma [ lemma : ic01 ] ( or lemma [ lemma : gij ] for @xmath340 ) . for the condition c2 ( @xmath81 must be @xmath341 ) , we have @xmath342 and @xmath343 from lemma [ lemma : ic01 ] . hence , by the same calculation as that for @xmath330 , we have @xmath334 . finally , let us consider the condition c3 . for @xmath344 , we have @xmath345 and @xmath346 by lemma [ lemma : ic01 ] again so , @xmath347 . for @xmath340 , we can not use lemma [ lemma : ic01 ] because of @xmath348 . however , @xmath349 holds from the assumption @xmath350 . hence a direct calculation shows @xmath351 , which is the formula for @xmath340 . before proving proposition [ pp : fl0 ] , let us show the following two lemmas . the proofs of them are easy application of induction . [ lemma:1/a ] the derivative @xmath352 is given by @xmath353 where the summation for the @xmath1-th term ( @xmath354 ) is taken on all combinations @xmath355 for dividing @xmath291 into @xmath1 groups . here @xmath356 represent a decomposition of @xmath357 ; @xmath358 [ lemma : fl ] let @xmath88 . then the derivative @xmath359 is given by @xmath360 where @xmath361 and @xmath362 are a decomposition of @xmath357 ; @xmath363 and the summations are taken on all the combinations of the decompositions . now we prove proposition [ pp : fl0 ] . + let us assume that any distinct @xmath268 column vectors of @xmath24 are linearly independent . then , from lemma [ lemma : h ] , [ lemma:1/a ] , and [ lemma : fl ] , we have @xmath364 on the other hand , from lemma [ lemma : ip ] , @xmath274 is 0 . the proof is completed . + finally , we are in the position to prove theorem [ thm : thm ] . + the formula for the case @xmath340 is given by proposition [ pp : jacobi ] . let us assume @xmath365 . from proposition [ pp : fl0 ] , all nonlinear terms with orders less than @xmath81 are zero , so we only study the @xmath81-th nonlinear terms . from the assumption and similar argument in the proof of proposition [ pp : fl0 ] , the derivative @xmath366 is given by @xmath364 then , the classification in lemma [ lemma : ip ] shows that @xmath367 in c0 and c1 . moreover , the condition c3 does not occur from the assumption . hence , only nonzero terms are derived from the condition c2 . it should be noted that the set of indices @xmath368 satisfying the condition c2 with @xmath51 is exactly the same as @xmath369 . hence , lemma [ lemma : ip ] results in @xmath370 for @xmath371 . finally , we have the following @xmath372 the identity in ( [ eq : hosum ] ) is derived by just substituting @xmath373 . it completes the proof . the purpose of this section is to derive corollary [ corollary : ag ] . this corollary shows that usual techniques in algebraic geometry codes can be applied to control , not only the minimum distance of the code , but also the approximate ml decoding . for the definitions of basic tools in algebraic geometry such as genus , divisor , riemann - roch space , differential , and residue , we refer to @xcite . we also request basic knowledge of algebraic geometry codes in this section ( e.g. , see @xcite , @xcite and @xcite ) . let @xmath374 be a finite field with @xmath375 elements . let us first recall two classes of algebraic geometry codes called geometric reed - solomon codes and geometric goppa codes . let @xmath376 be an absolutely irreducible nonsingular projective curve over @xmath374 . for rational points @xmath377 on @xmath376 , we define a divisor on @xmath376 by @xmath378 . moreover , let @xmath379 be another divisor whose support is disjoint to @xmath380 . we assume that @xmath379 satisfies the following condition @xmath381 for the sake of simplicity . here @xmath382 is the genus of @xmath376 . geometric reed - solomon codes are characterized by the riemann - roch space associated to @xmath379 @xmath383 where @xmath384 is the set of nonzero elements of the function field @xmath385 , and @xmath386 is the principal divisor of the rational function @xmath387 . the fundamental fact that the riemann - roch space @xmath388 is a finite dimensional vector space leads to the following definition . _ the geometric reed - solomon code _ @xmath389 of length @xmath390 over @xmath374 is defined by the image of the linear map @xmath391 given by @xmath392 . on the other hand , geometric goppa codes are defined via differentials and their residues . let us denote the set of differentials on @xmath376 by @xmath393 , and define for each divisor @xmath380 @xmath394 where @xmath395 is the divisor of the differential @xmath396 . _ the geometric goppa code _ @xmath397 of length @xmath398 over @xmath374 is defined by the image of the linear map @xmath399 given by @xmath400 , where @xmath401 expresses the residue of @xmath396 at @xmath402 . the following propositions are an easy consequence of the riemann - roch theorem . ( e.g. , @xcite , @xcite , @xcite)[pp : grs ] 1 . the dimension of the geometric reed - solomon code @xmath389 is @xmath403 and the minimum distance satisfies @xmath404 . 2 . the dimension of the geometric goppa code @xmath397 is @xmath405 and the minimum distance satisfies @xmath406 . the codes @xmath389 and @xmath397 are dual codes . it should be noted that the minimum distances of these two codes are controlled by the genus of @xmath376 and the choice of the divisor @xmath379 , and , as a result , they induce appropriate linear independence on their parity check matrices . for an application of theorem [ thm : thm ] , we need to derive expanded codes over @xmath203 from geometric reed - solomon and geometric goppa codes over @xmath407 . let @xmath408 be a code over @xmath374 with length @xmath390 , and @xmath409 be a basis of @xmath203-vector space @xmath374 . this basis naturally induces the map @xmath410 by expressing each element @xmath411 in @xmath412 as coefficients of @xmath203-vector space . then the expanded code of @xmath408 over @xmath203 is defined by @xmath413 a relationship between @xmath414 and @xmath415 is given by the following proposition . if a code @xmath414 has parameters @xmath416 $ ] , where @xmath390 is the code length , @xmath1 is the dimension , and @xmath36 is the minimum distance , then its expanded code @xmath415 has the parameters @xmath417 $ ] . now we apply theorem [ thm : thm ] to geometric reed - solomon / goppa codes . let @xmath418 , and let @xmath419 and @xmath420 be the expanded codes over @xmath203 of a geometric reed - solomon code @xmath389 and a geometric goppa code @xmath397 over @xmath374 . then , we have the following corollary . [ corollary : ag ] the expanded geometric reed - solomon code @xmath419 has the minimum distance @xmath421 . furthermore , there exists an @xmath81-th order approximate ml decoding with @xmath422 such that @xmath423 . the first statement is the property of a geometric reed - solomon code and its expansion . the second statement follows from theorem [ thm : thm ] and the duality of @xmath389 and @xmath397 . ( e.g. , @xcite , @xcite , @xcite ) + let @xmath424 be a power of 2 . the hermitian curve @xmath425 is given by the homogeneous equation @xmath426 and its genus is @xmath427 , because there are no singular points . it is known that the number of rational points over @xmath374 is @xmath428 . let us fix @xmath429 as an example . then , the following is the list of the rational points on @xmath425 : @xmath430 where @xmath396 is a primitive element of @xmath431 and @xmath432 ( i.e. , @xmath433 ) . let us suppose @xmath434 ( hence the code length is @xmath435 ) , and @xmath436 , @xmath437 . a basis of the riemann - roch space @xmath388 with @xmath438 is given by @xmath439 then , we can explicitly show a generator matrix of the expanded geometric reed - solomon code @xmath389 over @xmath203 as follows @xmath440 it should be noted that this case leads to a self - dual code @xmath441 . hence , its parity check matrix @xmath22 is the same as @xmath24 , and a direct calculation proves that the minimum distance of this code is 4 . it means that we have the @xmath442-rd order ml decoding in corollary [ corollary : ag ] . the explicit forms of @xmath443 are given as @xmath444 where @xmath445 . to conclude this paper , we address the following comments and discussions , some of which will be important for designs of good practical error - correcting codes . codeword fixed points in @xmath21 are stable from proposition [ pp : stable ] , and non - codeword poles in @xmath171 are unstable from proposition [ pp : gradient ] in the sense that nearby points to a pole in @xmath171 leave away from the pole . let us recall that each @xmath6-bit received sequence @xmath43 and its initial point @xmath67 are related by ( [ eq : ip ] ) and it is characterized that @xmath50 is the closest point to @xmath97 in @xmath446 . hence , if the received sequence @xmath50 is a codeword , then @xmath97 may approach to the codeword fixed point @xmath50 . this obviously depends on whether or not @xmath97 is located in the attractor region of the codeword fixed point , although it is actually the case when @xmath10 is small enough because of the stability . so far , a general structure of the attractor region for each codeword fixed point is not yet known . however this is an important subject since it is indispensable to give an estimate of error probabilities of ml decoding and its approximation . similar arguments also hold for a non - codeword received sequence and its repelling property . let us recall that the ml decoding rule is given by ( [ eq : mlrule ] ) which checks the location of the image for an initial point @xmath97 to the point @xmath0 . hence the local dynamics around the point @xmath0 will be important for decoding process . in the following , we explain the local dynamics around @xmath0 in two different cases : the jacobi matrix @xmath77 at @xmath0 is ( i ) identity or ( ii ) not identity . in the case ( i ) , the local dynamics around @xmath0 is precisely determined by theorem [ thm : thm ] . as explained after theorem [ thm : thm ] , the nonlinear dynamics around @xmath0 is closely related to the encoding structure of the code and the decoding process . in the case ( ii ) , let us suppose that @xmath243 in corollary [ corollary : eigenvalue ] induces unstable eigenvalues @xmath447 and let us focus on its stable / unstable eigenspaces @xmath448 , respectively . from its eigenvector , @xmath449 is given by the 1 dimensional subspace spanned by @xmath450 this plays a role to make @xmath451 to be equal , and it reflects the fact @xmath452 . it means that the unstable subspace @xmath449 points to codewords directions , i.e. , @xmath21 . on the other hand , @xmath453 is spanned by the stable eigenvectors @xmath454 in corollary [ corollary : eigenvector ] . contrary to the unstable eigenvector , these eigenvectors play a role to generate different elements in @xmath451 and point to non - codewords directions under time reversal , i.e. , @xmath171 . therefore , the fixed point @xmath0 can be regarded as an indicator to codewords in the sense that non - codeword elements shrink and codeword elements expand around @xmath0 . in both cases , further nonlinear analysis of center / stable / unstable manifolds of @xmath0 will be useful for finding suitable encoding rules and estimating the decoding performance for the approximate ml . from the above argument on the fixed point @xmath0 , it seems to be appropriate to design a generator matrix to be hyperbolic at @xmath0 , because @xmath0 separates expanding and shrinking directions properly and these separations have an affect on the decoding performance . however , if @xmath0 is hyperbolic , then the coding rate must satisfy @xmath455 by corollary [ corollary : eigenvalue ] . namely , the hyperbolicity prevents a code to have a high coding rate greater than half , although this is not a strict restriction in particular applications like wireless communication channels . therefore it is necessary to have a center eigenspace at @xmath0 for a code with the rate @xmath456 . the normal form theory in dynamical systems ( e.g. , see @xcite ) enables us to transform a map into a simpler form by using a near identity transformation around a fixed point . one of the essential points is that nonresonant higher order terms can be removed from the original map by this transformation . theorem [ thm : thm ] can be interpreted from the viewpoints of normal forms in such a way that an algebraic geometry code gives only zero nonresonant terms in the expansion of @xmath82 at @xmath0 . then , it leads to the following natural question whether a code whose rational map does not have resonant terms , but has nonresonant terms which are not necessarily zeros in its expansion is a good error - correcting code or not through a near identity transformation . at least , this class of codes contains algebraic geometry codes as a subclass , and a similar statement to corollary [ corollary : ag ] holds through near identity transformations . it seems to be valuable to mention a relationship to ldpc codes @xcite , @xcite , which are a relatively new class of error - correcting codes based on iterative decoding schemes ( for a reference to this research region , see @xcite ) . the iterative decoding schemes mainly use so - called sum - product algorithm for ml decoding and deal with a marginalized conditional probability in ( [ eq : bitml ] ) as a convergent point . although this coding scheme gives a good performance in some numerical simulations , mathematical further understanding of the sum - product algorithm and ml decoding is desired to design better coding schemes . from the viewpoint of dynamical systems , it seems to be natural to formulate the sum - product algorithm or ml decoding itself as a certain map , and then analyze its mechanism . the strategy in this paper is based on this consideration . the authors express their sincere gratitude to the members of tin working group for valuable comments and discussions on this paper . this work is supported by jst presto program .	this paper studies maximum likelihood(ml ) decoding in error - correcting codes as rational maps and proposes an approximate ml decoding rule by using a taylor expansion . the point for the taylor expansion , which will be denoted by @xmath0 in the paper , is properly chosen by considering some dynamical system properties . we have two results about this approximate ml decoding . the first result proves that the order of the first nonlinear terms in the taylor expansion is determined by the minimum distance of its dual code . as the second result , we give numerical results on bit error probabilities for the approximate ml decoding . these numerical results show better performance than that of bch codes , and indicate that this proposed method approximates the original ml decoding very well . maximum likelihood decoding , rational map , dynamical system + * ams subject classification . * 37n99 , 94b35
in the field of network science , real - world complex systems are abstracted as complex networks , in which nodes represent individuals and links denote the connections or interactions between individuals @xcite . nowadays , although we can obtain abundant data of various complex systems due to advanced technologies , it is demonstrated that larger parts of the data of the complex systems are still not available , and there are non - ignorable errors in the data that we obtain @xcite . thus , new methods are needed to process , correct , and make predictions from the data . link prediction methods aim to predict the missing or future links among network data @xcite . specifically , they estimate the existence likelihood of links between two nodes based on observed links and nodes attributes . link prediction has broad applications @xcite . for instance , it can be used in detecting potential interactions in protein - protein interaction network @xcite , recommending friends and goods in online social networks @xcite , exploring potential coauthor relationships in collaboration networks @xcite and so on . previous algorithms are basically from the field of machine learning including supervised learning @xcite , markov chain @xcite , and likelihood estimation @xcite . these algorithms heavily depend on attributes of nodes , and they do not seriously consider structural characteristics of networks . besides , their computation cost is inhibitive for large real - world networks @xcite . recently , the booming network science community gets deeper insights into the structure of complex networks @xcite , and further stimulates the research of link prediction @xcite . lots of prediction algorithms based on structural similarity are proposed , which can be classified into three types : local indices @xcite , quasi - local indices @xcite and global indices @xcite . for example , common neighbors ( cn ) @xcite , preferential attachment ( pa ) @xcite , adamic - adar ( aa ) @xcite , resource allocation ( ra ) @xcite , etc . are local indices , which only use the nearest neighborhood information . katz @xcite , leicht - holme - newman ( lhn ) @xcite , simrank @xcite and so on , which need the knowledge of the whole network topology , are global ones . quasi - local indices , including local path ( lp ) @xcite , local random walk ( lrw ) @xcite , superposed random walk ( srw ) @xcite , etc . need more topological information than local indices , but less topological information than global ones . generally , local indices have the lowest prediction accuracy , but their computational cost is the smallest among the three types of similarity indices . global indices are the opposite of local indices , while quasi - local ones are the trade - off between them . recently , information entropy has been employed to measure the complexity of the topological structures of complex networks @xcite . results showed that information entropy can better capture the topological difference than the other typical network measurements @xcite . moreover , information entropy has nature connection with link prediction problem in that the probability of a missing link between two nodes can be transformed into the corresponding information entropy . thus , researchers began to apply information entropy theory to link prediction problems in complex networks @xcite . for example , tan et al @xcite reexamined the role of common neighbors in link prediction by using the mutual information , and proposed a mutual information - based similarity index . xu et al @xcite derived the information entropy of a path , and studied the contributions of paths in link prediction based on path entropy , and finally provided a path entropy based similarity index . simulation results @xcite showed that the similarity indices based on information entropy have higher prediction accuracy than the other types of similarity indices . many complex networks contain the information of link weights @xcite , measuring the strength of connections between nodes , thus , it is reasonable to consider the link weights when designing link prediction algorithms to further improve the prediction accuracy . so far there have been a few tries in literature . murata and moriyasu @xcite improved the cn and aa indices by using the link weights information . bai meng etc @xcite developed the weighted version of lp index . l l y etc . @xcite particularly explored the role of weak ties in link prediction and found that weak ties can improve the prediction accuracy effectively . simulation results showed that the weighted version of these typical similarity indices indeed have a higher accuracy than the original ones . however , they still have a large space to improve . in this paper , we apply information entropy to link prediction in weighted networks . specifically , we reexamine the role of path in link prediction by considering both the path entropy and the path weight , and finally we propose a weighted similarity index . through simulation on weighted real - world networks we show the prediction accuracy of our index and make comparison with other typical weighted indices . quantitative measurement of complex networks is a hot topic in network science @xcite , and various measurements are proposed including degree , betweenness , closeness , k - core number and so on @xcite . however , most of the measurements are for nodes and links , and only a very few are particularly for paths such as path length @xcite and path attack centrality @xcite . in our latest work @xcite , we applied information theory to measure the importance of paths , and specifically we derived the entropy of a path . assuming that @xmath0 ( @xmath1 ) represents the event there is ( not ) a link between node @xmath2 and node @xmath3 in the network without degree correlation . then , the probability of @xmath0 is calculated as follows : @xmath4 where @xmath5 ( @xmath6 ) is the degree of @xmath2 ( @xmath3 ) , and @xmath7 is the total link number in the network . based on the definition of entropy and eq . ( 1 ) , we obtain the entropy of @xmath0 as follows : @xmath8 let s further consider a simple path @xmath9 of length @xmath10 . the occurrence possibility of @xmath11 can be calculated approximately as : @xmath12 which indicates that the occurrence probability of a path is approximately equal to the product of its links occurrence probabilities . thus , the entropy of path @xmath11 can be calculated as@xcite : @xmath13 eq . ( 4 ) indicates that the entropy of a path approximates to the sum of its links entropies . furthermore , path entropy takes both path length and node degree information into consideration . the longer path or the smaller node degree , the larger path entropy . through the basic definition of entropy , we know that if a path has a large path entropy , then its occurrence probability should be low , in other words , the existence of the path is important to the network . generally , path entropy is more discriminating than the other measurements . the reason is that when computing the path entropy , we not only consider the path length , but also node degrees , and even the order of the nodes in the path . in real society , the interactions between individuals are of different strengths for many complex networks , which are called weighted networks @xcite . for instance , in the air traffic network , the weight of a link is measured as the number of passengers in the related flight . in the router - level of the internet , the weights of links are generally correlated with the bandwidth of the physical connections or the cost for data transmission between routers . in social networks , weights of links are related to the interacting times or frequencies between individuals . however , there is no specific research about the weights of paths in networks . in traffic routing protocols , we usually choose the optimal path , of which the sum of the cost of all links is the smallest among the candidate paths @xcite . enlightened by this , here we define the weight of a path @xmath11 , as the sum of its links weights , which is : @xmath14 where @xmath15 is the weight of the link with end nodes @xmath16 and @xmath17 . in the framework of information theory , the probability of link existence between two nodes can be expressed with information entropy . then , the link prediction problem can be defined as the conditional entropy , which is as follows @xcite : @xmath18 where @xmath19 is the topological structure information we know , and based on which we make the prediction of the event @xmath20 . @xmath21 is the joint entropy , which quantitatively measures how much the existence of @xmath19 leads to the decrease of uncertainty of the event @xmath20 . in this paper , we consider the contributions of simple paths in link prediction . thus , we have @xmath22 , where @xmath23 is the set of all simple paths of length @xmath24 between @xmath2 and @xmath3 , and @xmath25 is the maximum length of simple paths we consider in the network . previous results showed that the longer the path , the less important the paths in link prediction . moreover , results demonstrated that the link weights can be used to improve the prediction accuracy . here we combine the path length , path weight , and path entropy to calculate the contribution of a path @xmath11 with length @xmath24 in the link prediction : @xmath26 where @xmath27 is a free parameter , which controls the influence of path weights . when @xmath28 , path weights are ignored . when @xmath29 , path with large weight is thought to have large contribution , otherwise , when @xmath30 , path with small weight is thought to have large contribution . @xmath31 is the optimal penalty factor that we found for suppressing the contributions of long paths . then , we give the definition of our prediction index , namely weighted path entropy ( wpe ) by considering the contributions of all simple paths , which is as follows : @xmath32 based on eq . ( 2 ) , ( 4 ) , ( 5 ) and ( 8) , we finally obtain the complete expression of wpe index as follows : @xmath33 assuming an undirected network @xmath34 , where @xmath35 , @xmath36 , and @xmath37 denote the sets of nodes , links and link weights respectively . note that in @xmath37 , @xmath38 , which means there is no direction for the weight between two nodes . to measure the predicting ability of an index , @xmath36 is usually randomly divided into two parts : a training set @xmath39 ( in this paper , @xmath40 of all links ) and a probe set @xmath41(@xmath42 of all links ) . clearly , @xmath43 and @xmath44 . two standard metrics auc and precision are applied to quantify the prediction quality . to calculate auc , @xmath45 times of independent score comparisons are made between node pairs in @xmath41 and @xmath46 . each time we select a node pair randomly from these two sets respectively and compare their scores . if there are @xmath47 times that the score of the link from @xmath41 is higher ( or smaller in the case of wpe ) than the link from @xmath46 , and @xmath48 times that they have the same scores , then , auc is calculated as : @xmath49 auc should be about @xmath50 if all the scores are generated from an independent and identical distribution . precision aims to measure the ability to predict top ranked links . if among the top @xmath51 links ranked by the scores , there are @xmath52 links belonging to @xmath41 , then precision is calculated as : @xmath53 note that for the wpe index , node pairs with small scores ( @xmath54 ) have large probability to be connected by links . thus , when calculating the precision , we rank the links based on their @xmath54 from the smallest to the largest , and thus the top ranked links have the smallest scores , which is the opposite to the other prediction indices . next , we introduce three indices and their weighted versions that we use for comparison . they are common neighbors ( cn ) @xcite , adamic - adar index ( aa ) @xcite and local path ( lp ) @xcite , which are defined as follows : @xmath55 @xmath56 @xmath57 where @xmath58 is the set of common neighbors of node @xmath2 and @xmath3 , and @xmath59 is the set of neighbors of node @xmath60 . @xmath61 is the adjacency matrix , and we set @xmath62 to obtain a near optimal prediction accuracy . in addition , their parameter - dependent weighted versions , wcn @xcite , waa @xcite , and wlp @xcite are respectively defined as follows : @xmath63 @xmath64 @xmath65 where @xmath66 . @xmath67 is the path of length three between node @xmath68 and node @xmath69 . @xmath24 and @xmath70 are the intermediate nodes in the path @xmath67 . cn and aa as well as their weighted versions are all taken as local indices which are only based on the nearest neighbors . lp and its weighted version are quasi - local indices since they consider the next - nearest neighbors . six weighted networks from disparate fields are used to test the accuracy for various prediction indices . the directions of links are ignored . the self - connections and multiple links are deleted from the network data . for the unconnected networks , we select the maximum connected components for experiments . the statistics of these networks are summarized in table 1 . @xmath71 les @xcite : this undirected network contains co - appearances of characters in victor hugo s novel ` les misrables ' . a node represents a character and a link between two nodes shows that these two characters appeared in the same chapter of the book . the weight of each link indicates how often such a co - appearance occurred . @xmath72 usair @xcite : the network of us air transportation , where nodes represent airports , links represent routes between airports , and the weight of a link is the frequency of flights between two airports . @xmath73 bomb @xcite : this undirected network contains contacts between suspected terrorists involved in the train bombing of madrid on march 11 , 2004 as reconstructed from newspapers . a node represents a terrorist and a link between two terrorists shows that there was a contact between the two terrorists . the link weights denote how ` strong ' a connection was , by considering the friendship and co - participating in training camps or previous attacks . @xmath74 bible @xcite : this undirected network contains nouns ( places and names ) of the king james version of the bible and information about their co - occurrences . a node represents one of the above noun types and a link indicates that two nouns appeared together in the same bible verse . the link weights show how often two nouns occurred together . @xmath75 florida @xcite : this network contains the carbon exchanges in the cypress wetlands of south florida during the dry season . nodes represent taxa and a link denotes that a taxon uses another taxon as food with a given trophic factor ( which is the basis of link weight ) . @xmath76 c.elegans @xcite : the neural network of the nematode worm c. elegans , where a node represents a neuron , a link joins two nodes if the corresponding neurons have synaptic contacts , and the weight represents the number of synapses between two neurons . .the topological statistics of six real - world weighted networks . @xmath77 is the number of nodes . @xmath78 represents the number of links . @xmath79 is the average degree . @xmath80 represents the degree heterogeneity , defined as @xmath81 . @xmath82 is the clustering coefficient . [ cols="^,^,^,^,^,^",options="header " , ] to investigate the ability of our prediction index , we perform experiments on the six weighted networks , and make comparison with three typical indices . both the unweighted and weighted versions of these indices are tested . note that the link weights of the networks are only considered when calculating the weighted versions of these indices . in addition , for the weighted versions of these indices , we adjust the control parameter @xmath27 and achieve the near optimal performances measured by auc and precision . for the pe and wpe indices , @xmath83 means only paths with length of 2 are used in the calculation , while @xmath84 indicates that paths with lengths of both 2 and 3 are used in the calculation . our simulation results are shown in table 2 and 3 , as well as fig . 1 and 2 . each value is the average of 100 independent runs . in the tables , the maximum performances for each network are marked in bold font . table 2 and table 3 generally show that it is necessary to consider the link weights since the weighted version of these indices achieves higher prediction accuracy than the unweighted ones when the appropriate control parameter is considered ( the near optimal parameters @xmath85 are given in the tables ) . furthermore , table 2 indicates that from the perspective of auc , the prediction accuracy of all the unweighted indices is already high for all the six real - world networks , so that the improvements of the weighted versions are not significant . however , as shown in table 3 , the improvements of the weighted versions are more obvious from the perspective of precision . moreover , through table 2 and 3 we see that for all the unweighted and weighted prediction indices , wpe always have the best performances for all the six real - world networks . especially , for the networks of bible , florida , c. elegans and usair , we can see from table 3 that pe and wpe are much better than the other prediction indices . it is worth mentioning that for pe and wpe , extra consideration of the contributions of longer paths is not always necessary and sometimes negative for link prediction . for instance , from table 2 and 3 we see that for some networks @xmath83 is better than @xmath84 . 1 and 2 show how the weight affects the prediction accuracy , for the weighted indices . note that for wpe , we focus on path weight defined as the sum of all the link weights in the path , while the other weighted indices only consider link weights . from fig . 1 and 2 , we can see that for wpe , generally both auc and precision increase first , and then decrease with the control parameter @xmath27 , and there are optimal parameters @xmath85 ( which are given in table 2 and 3 ) corresponding to the maximum performances . the curves of the other weighted indices have the similar change trends . these results indicate that the prediction accuracy is sensitive to path weights for wpe and link weights for the others , and we should balance the influences of the weak ties and strong ties by appropriately choosing the control parameter to achieve high prediction accuracy . moreover , we should be slightly biased to the contributions of weak ties , since the simulations results demonstrate that the optimal parameters are less than 1 for all the six networks except bomb ( for waa and wcn , c. elegans is also an exception ) . in summary , we study the link prediction problem in weighted complex networks from the perspective of information entropy . in fact , the likelihood of a link between two nodes can be converted into entropy , and small entropy corresponds to large probability of link existence . in this paper we consider the contributions of simple paths between node pairs in link prediction . specifically , we measure the contribution of an existing path by considering its length , entropy , and weight , which is further defined as the sum of link weights in the path . furthermore , we propose a weighted path entropy ( wpe ) index for link prediction by considering the contributions of all existing simple paths . through simulation on several real - world weighted networks , we found that wpe has higher prediction ability measured by auc and precision than the other typical weighted similarity indices . in fact , the pe index proposed in our previous work already has high prediction ability . thus , by appropriately considering the path weight , wpe further improves the prediction accuracy . through simulation , we also found that weak tie is more critical than the strong tie in link prediction . note that in our context , weak tie refers to path with small weight , while in the other works weak tie means small weight link . this work was supported by the national natural science foundation of china ( grant nos . 61304154 ) , the specialized research fund for the doctoral program of higher education of china ( grant no . 20133219120032 ) , the postdoctoral science foundation of china ( grant no . 2013m541673 ) , and china postdoctoral science special foundation ( grant no . 2015t80556 ) . barabsi a l 2016 _ network science _ ( cambridge university press ) newman m 2010 _ networks : an introduction _ ( oxford university press ) chen g r , wang x f and li x 2012 _ introduction to complex networks : models , structures and dynamics _ ( higher education press ) mayer - schnberger v and cukier k 2013 _ big data : a revolution that will transform how we live , work , and think _ ( houghton mifflin harcourt ) abbasi a , sarker s and chiang r h 2016 _ j. assoc . * 17 * 3 l l y and zhou t 2011 _ physica a _ * 390 * 1150 wang p , xu b w , wu y r and zhou x y 2015 _ sci . china inform . _ * 58 * 1 l l y , medo m , yeung c h , zhang y c , zhang z k and zhou t 2012 _ phys . rep . _ * 519 * 1 cheng w l and jian h r 2013 bioinformatics * 29 * 355 sherkat e , rahgozar m and asadpour m 2015 _ physica a _ * 419 * 80 newman m e j 2001 _ proc . sci . _ * 98 * 404 mohammad a h , vineet c , saeed s and mohammad z 2006 _ the proceedings of the fourth workshop on link analysis , counterterrorism and security _ 22nd april , 2006 , bethesda , usa sarukkai r r 2000 _ comput . netw . _ * 33 * 377 getoor l and diehl c p 2005 _ acm sigkdd explor . _ * 7 * 3 cui w , pu c l , xu z q and yang j 2016 _ physica a _ * 457 * 202 barabsi a l and albert r 1999 _ science _ * 286 * 509 newman m e j 2001 _ phys . rev . e _ * 64 * 025102 kossinets g 2006 _ soc . networks _ * 28 * 247 liben - nowell d and kleinberg j 2007 _ am . technol . _ * 58 * 1019 adamic l a and adar e 2003 _ soc . networks _ * 25 * 211 zhou t , l l y and zhang y c 2009 _ eur . j. b _ * 71 * 623 l l y , jin c h and zhou t 2009 _ phys . e _ * 80 * 046122 liu w p and l l y 2010 _ europhys . epl _ * 89 * 58007 katz l 1953 _ psychmetrika _ * 18 * 39 leicht e a , holme p and newman m e j 2006 _ phys . e _ * 73 * 026120 jeh g and widom j 2002 _ proceedings of the eighth acm sigkdd international conference on knowledge discovery and data mining _ , new york , usa , p. 538 sol r v and valverde s 2004 _ information theory of complex networks : on evolution and architectural constraints , in : complex networks _ ( springer , berlin , heidelberg ) pp . 189207 halu a , mukherjee s and bianconi g 2014 _ phys . * 89 * 012806 anand k and g. bianconi 2009 _ phys . e _ * 80 * 045102 tan f , xia y x and zhu b y 2014 _ plos one _ * 9 * e107056 xu z q , pu c l and yang j 2016 _ physica a _ * 456 * 294 zhu b y and xia y x 2016 _ plos one _ * 11 * e0148265 murata t , moriyasu s 2007 _ ieee / wic / acm international conference on web intelligence _ bai m , hu k and tang y 2011 _ chin . b _ * 20 * 128902 l l y and zhou t 2010 _ epl _ * 89 * 18001 pu c l and cui w 2015 _ physica a _ * 419 * 622 pu c , li s , michaelson a and yang j 2015 _ phys . _ * 379 * 1633 feigenbaum j , papadimitriou c , sami r and shenker s 2005 _ distrib . comput . _ * 18 * 61 this data is released by knuth d e 1993 , available at http://moreno.ss.uci.edu/data.html the data is released by batagelj v and mrvar a in 2006 , available at http://vlado.fmf.uni-lj.si/pub/networks/data brian h 2006 _ amsci _ * 94 * 400 the koblenz network collection 2015 , available at http://konect.uni-koblenz.de/ the data is released by ulanowicz r e , bondavalli c and egnotovich m s in 1998 , available at http://vlado.fmf.uni-lj.si/pub/networks/data/bio/foodweb/foodweb.htm watts d j and strogatz s h 1998 _ nature _ * 393 * 440	information entropy has been proved to be an effective tool to quantify the structural importance of complex networks . in the previous work ( xu et al , 2016 @xcite ) , we measure the contribution of a path in link prediction with information entropy . in this paper , we further quantify the contribution of a path with both path entropy and path weight , and propose a weighted prediction index based on the contributions of paths , namely weighted path entropy ( wpe ) , to improve the prediction accuracy in weighted networks . empirical experiments on six weighted real - world networks show that wpe achieves higher prediction accuracy than three typical weighted indices . link prediction , weighted networks , information entropy 89.75.hc , 89.75.fb , 89.20.hh
it is conventionally thought that the thermal x - ray components of neutron stars are originated from the initial residual heat when the stars cool @xcite , even before the discovery of galactic x - ray sources and pulsars . however , we will focus on this old problem in the regime of quark stars since there is no clear observational evidence to rule out quark stars or neutron stars . we demonstrate in this paper that the observed thermal emission of isolated pulsars could be well understood in a solid quark star model . the study of quark matter phases , both hot and cold , has been an interesting topic of research in recent years . in an astrophysical context , quark stars composed by cold quark matter have yet not been ruled out by the measured properties of pulsar - like compact stars @xcite . it has recently been proposed that realistic quark matter in compact stars could exist in a solid state @xcite , either as a super - solid or as a normal solid @xcite . the basic conjecture of normal solid quark matter is that de - confined quarks tend to form quark - clusters when the temperature and density are relatively low . below a certain critical temperature , these clusters could be in periodic lattices immersed in a degenerate , extremely relativistic , electron gas . note that even though quark matter is usually described as weakly coupled , the interaction between quarks and gluons in a quark - gluon plasma is still very strong @xcite . it is this strong coupling that could cause the quarks to cluster and form a solid - like material . in fact , various kinds of observations could put constraints on the state of the matter in a pulsar , and the cooling behavior is suggested since 1960s , even before the discovery of pulsars . can contemporary observations of x - ray thermal emitting and cooling pulsars be understood in the proposed solid quark star ( hereafter sqs ) model ? this is a question we try to answer in the paper , and we address that sqss could not be ruled out by the thermal observations . moreover , sqss might provide an opportunity to evaluate pulsar moments of inertia , if the thermal emission is also powered by spin . in this paper , we argue that contemporary observations of thermal x - ray emitting pulsars are consistent with the assumption that these sources are in fact sqss . the present analysis and calculation will concentrate on the observational temperature range , i.e. a few hundreds of to a few tens of ev . nonetheless , a phenomenological scenario for the whole thermal history of a strange quark star is also outlined , as described in 2.1 . the stellar residual heat would be the first energy source supporting the x - ray bolometric luminosities of sqss . 2.2 is involved on this topic , including both of the partial contributions of the lattices and the electrons . other energy reservoirs could exist as the stellar heating processes . this will be discussed in 2.3 , where different heating mechanisms are introduced , according to the different x - ray pulsar manifestations . in 3 , we compare the observations and the predictions given by the sqs pulsar model . our conclusions as well as further discussions are presented in 4 . mev ) when the star is born , and the state of the star could be the fluid of individual quarks . _ stage 2 _ : individual quark cluster phase . as the temperature decreases , individual quarks tend to form quark clusters because of the strong coupling between them . the state of the star could then be the fluid of quark clusters . _ stage 3 _ : solid quark star phase . as the temperature drops to a melting temperature , the fluid of quark clusters tend to solidify to form periodic lattice structure , such as bcc structure.,scaledwidth=80.0% ] the cooling process of a quark star can be quite complicated to model when starting from the birth of the star . theoretical uncertainties make it difficult to predict the exact temperatures where phase transitions occur . for illustrative purposes , we present the following scenario where cooling of a quark star takes place in approximately three stages ( see fig . [ fig : stg ] ) . the first stage occurs just after the birth of the quark star if its initial temperature is much higher than @xmath4 k ( 10 mev ) . the emission of neutrinos and photons will lead to fast cooling . hence , the star quickly enters the second stage where de - confined quarks begin to form quark clusters . as the temperature drops further , the fluid solidifies and the star is said to enter the third cooling stage or the sqs phase . here , quark clusters form periodic lattice structures ( e.g. the bcc structure ) . it is uncertain exactly how long the star will spend in each phase , or even if it will ever enter all three states . if the initial temperature is just around @xmath5 k or lower , stage 1 may be short lived or even non - existent . the quark - gluon plasma could be strongly coupled at birth and quark - clusters could be present immediately after formation . it is even possible that the melting temperature of solid quark matter could be @xmath6 k or higher . hence , a quark star would enter stage 3 immediately . the mechanisms for the emission of thermal photons and neutrinos in each specific stage would be quite different . a hot quark star ( in stages 1 or 2 ) would be a good radiator for thermal equilibrium photons with energy more than @xmath7 mev @xcite . meanwhile , the intense release of thermal energy would stimulate the generation and radiation of electron - positron pairs from the hot bare quark surface . the annihilation of the electron - positron pairs would generate the emission of photons inversely , and this plasma could be optically thick enough to produce a black body spectrum @xcite . the emission of neutrinos would be given rise to by the pair production process as well as the plasma process induced by the ultra - relativistic degenerate electron gas inside the star @xcite . the basic urca process would be another neutrino radiation component for stage 1 . when the star cools down and becomes solidified ( in stage 3 ) , thermal equilibrium photons ( usually soft x - ray ) could also be emitted from the bare sqs surface as free electrons transit in the levels of the energy bands of solid quark matter @xcite . the electron - phonon interaction as well as the interaction between electrons themselves might result in a metal - like spectrum . @xcite analyzed the spectrum of rx j1856 phenomenologically , and they did not conclude that there are significant differences between the metal - like spectrum and black body spectrum . hence , a black body thermal spectrum would be a good approximation for sqss . the neutrino emissivity of clustered quark matter is theoretically so far unknown . the pair production and the plasma process would also lead the neutrino cooling for sqss , if the stars enter the stage 3 at high temperatures , such as @xmath6 k. nevertheless , in the observational low temperature range @xmath8 k , the neutrino luminosity would be low enough so that photon cooling would be the dominant process for sqss . in general , in the low temperature range ( @xmath8 k ) which is the case considered by the present paper , the sqss cool down by losing thermal photons , and the sources of the energy that enable the thermal emission are firstly the residual thermal energy of the stars and secondly the energy input caused by the processes of stellar heating . the energy relation would then be described as @xmath9 where @xmath10 is the stellar bolometric x - ray luminosity , @xmath11 is the release rate of the stellar residual thermal energy , and @xmath12 is the luminosity of stellar heating . provided that the volume of a sqs is a constant , then the stellar residual thermal energy @xmath13 would only be the function of the stellar temperature @xmath14 , @xmath15 where @xmath14 is the value in the star s local reference frame . the heat capacity of the star @xmath16 comprises of the partial contribution of the lattice structure @xmath17 and that of the degenerate electrons @xmath18 , or @xmath19 . following debye elastic medium theory , the characteristic of the lattice heat capacity of solid quark matter could be evaluated by debye temperature , @xmath20 where @xmath21 is the reduced planck constant , and @xmath22 is the boltzmann constant . @xmath23 is debye cut - off frequency ( i.e. the maximum frequency of the wave that could propagate in a medium ) , which equals debye wave number @xmath24(@xmath25 is the number density of classical particles , or quark clusters for solid quark matter ) times the average sound speed in the medium , i.e. @xmath26 . for a sqs , the average sound speed could be the light speed approximately . a linear equation of state , extended to be used for a quark star , indicates that the pressure @xmath27 , where @xmath28 is the mass density of a quark star and @xmath29 is the speed of light . so an estimate could be @xmath30 . @xmath31 , where @xmath32 denotes the baryon number density of solid quark matter in the unit of @xmath33 . @xmath33 is the baryon number density of normal nuclear matter and equals 0.17 @xmath34 . we consider @xmath35 could be the typical values for a sqs . note that in the following calculation in this paper , we will adopt @xmath36 , since the variation of @xmath32 between @xmath37 would not cause the results to vary in orders . @xmath38 is the number of valence quarks in a quark cluster . we may expect that @xmath39 , since @xmath40 if quark-@xmath41-like clusters are formed @xcite , and @xmath38 could even be conjectured to be in the order of @xmath42 . debye temperature @xmath43 of a sqs is then @xmath44 k , which could be even higher than the temperature when a quark star is born . hence , the heat capacity in the low - temperature limit or the temperature - cube law , is applicable for the lattice in the third stage , i.e. @xmath45 where @xmath46 is the heat capacity per classical particle ( or quark cluster when referring to the solid quark matter)@xcite . thus @xmath47 , where @xmath48 is the total number of clusters in a star . for the stellar electron gas , those electrons distributed in the vicinity of the fermi surface will contribute significant heat capacity . the electron heat capacity would be evaluated by @xmath49 where @xmath50 is the number of electrons in a star , and @xmath51 is the fermi energy of the degenerate electron gas . in the extremely relativistic case , @xmath52 , in which @xmath53 is the number density of electrons . according to the calculation using the bag model , @xmath53 is typically chosen as one part in @xmath54 of the baryon number density of strange quark matter @xcite . in the 3 times nuclear density case , @xmath55 @xmath56 . @xmath51 is hence in the order of @xmath57 mev . in fig . [ fig : cv ] , the heat capacities of these two components are plotted . as can be seen , the electron heat capacity is much larger than that of the lattice . noting that it is really ambiguous about the melting temperature of a sqs , though we extended the calculation to 10 mev . we can now evaluate the cooling time scale for a sqs , if the star is only powered by its residual thermal energy . we choose the stellar mass of 1.4@xmath58 as a typical case . we then have @xmath59 where @xmath60 is the local stellar radius , @xmath61 is the stefan - boltzmann constant . on the right hand side of the equation ( [ eq : timesl ] ) , the first energy loss component origins from the radiation of thermal photons , the second and the third terms are the neutrino luminosities induced by the pair production and the interior ultra - relativistic electrons , respectively . the calculation about these two neutrino luminosities refers to the analytic formulae given by @xcite . the concerning about the high temperature here makes the consideration on the neutrino cooling necessary . although the actual neutrino emissivity of clustered quarks is hitherto unknown , these two components of neutrino emission could really take place for a sqs . thus , one could obtain an upper bound on the residual - heat - powered cooling time scale for a sqs via equation ( [ eq : timesl ] ) . the calculation showed that a 1.4@xmath58-sqs could only be sustained in @xmath6218 d , when it cools down from @xmath5 k to @xmath54 k. this would be the intrinsic distinction for sqss from neutron stars . neutron star cooling is mainly residual thermal energy powered , while thermal x - ray emission and cooling processes of sqss could however be sustained by heating processes . heating processes may play a significant role in the thermal evolution or the visibility of pulsars in the soft x - ray band . pulsars with different magnetospheric properties may be undergoing different heating mechanisms . luminous nonthermal radiation and bright pulsar wind nebulae ( pwne ) , these observational manifestations may imply active magnetospheres . strong star wind or relativistic particle flow would be ejected from the poles of such a pulsar . a certain amount of backflow of such plasma induce the stellar heating , or an energy input could take place at the polar caps and disperse to the bulk of the star @xcite . the heat flow @xmath63 could thus be @xmath64 where @xmath65 and @xmath66 are the radius and the surface temperature of the polar caps in the local reference frame respectively . we adopted the one - dimensional approximation for the gradient of temperature , i.e. @xmath67 . the luminosity of the return - current stellar heating @xmath12 could be a function of the spin energy loss rate @xmath68 and could generally be in the form of a power law . moreover , a phenomenological study on the bolometric luminosity and @xmath68 reveals that the power index could be 1/2 or 1 , namely @xmath69 or @xmath70 ( see appendix a for the details ) . the metal - like interior of a sqs could imply that its thermal conductivity @xmath71 could be the sum of the partial components of phonons , electrons , and static impurities , or @xmath72 where the subscripts ` p ' , ` e ' , and ` imp ' denote the partial components mentioned above , respectively . the electron thermal conductivity @xmath73 could be the dominant component for the metal - like solid material @xcite . so phonon contribution @xmath74 would be neglected , and , as a preliminary model , we omit the partial component of static impurities . the electron thermal conductivity would be written as @xmath75 where @xmath76 is the partial component contributed by the collision between electrons , while @xmath77 is that contributed by the collision between phonons and electrons . for their analytic formulae , we refer to @xcite with assuming that the results remain hold for sqss . some pulsars , showing steady long - term soft x - ray fluxes , being inert in nonthermal emission , lacking the proves on the existences of the associated pwne , ( and sometimes ) owning excesses of x - ray luminosities relative to the spin energy loss rates , could alternatively be inactive pulsar candidates . due to their little magnetospheric manifestations , sqss will not suggest the thermal x - ray radiation of these sources is of spin origin . one possible way to understand the origin of the energy is the accretion in the _ propeller _ regime . in this regime , a shell of atmosphere of matter may form surrounding the star . matter closing to the inner boundary of the shell may interact directly with the rotating stellar magnetosphere , as a result of which most of the matter will be expelled outward @xcite . nevertheless , a certain fraction of the accreting material , described by the accretion efficiency @xmath78 , may diffuse starward and fall onto the surface of the star finally . in the propeller regime , the stellar magnetosphere radius or alfvn radius @xmath79 would be on one hand larger than its corotation radius @xmath80 , but is on the other hand somewhat smaller than the light cylinder radius @xmath81 , i.e. @xmath82 where @xmath83 is the gravitational constant , @xmath84 , @xmath60 , and @xmath85 are the stellar mass , radius , and the magnetic strength at the poles , respectively . the accretion rate @xmath86 could be scaled by eddington accretion rate @xmath87 ; so @xmath88 could denote the accretion rate in this unit , i.e. @xmath89 . during the accretion of the material which can reach the stellar surface eventually , the energy of gravitation would be released . furthermore , when the two - flavor baryonic matter impact upon the surface of a sqs , it will burn into the three - flavor strange quark matter phase , and the latent heat of @xmath90 mev @xmath91 mev per baryon could be released in the phase transition @xcite . the luminosity of stellar heating in this situation could then be @xmath92 where @xmath93 is the mass of a proton . llccccccc + no . & source & t ( kyr ) & @xmath94 ( mk ) & @xmath95 ( km ) & @xmath96 ( mk ) & @xmath97 ( km ) & @xmath98 ( @xmath99 ergs s@xmath100 ) & refs . + + 1 & psr b0531 + 21 & 1 & @xmath101 & @xmath102 & & & @xmath103 & ( 1 ) + & ( crab ) + 2 & psr j1811 - 1925 & 2 & @xmath104 & & & & & ( 2 ) + & ( in g11.2 - 0.3 ) + 3 & psr j0205 + 6449 & 0.82 - 5.4 & @xmath105 & @xmath106 & & & @xmath107 & ( 3 ) + & ( in 3c 58 ) + 4 & psr j1119 - 6127 & @xmath108 & @xmath109 & @xmath110 & & & @xmath111 & ( 4 ) + & ( in g292.2 - 0.5 ) + 5 & rx j0822 - 4300 & 2 - 5 & @xmath112 & @xmath113 & @xmath114 & @xmath115 & @xmath116 & ( 5 ) + & ( in pup a ) + 6 & psr j1357 - 6429 & @xmath117 & @xmath118 & @xmath119 & & & @xmath120 & ( 6 ) + 7 & rx j0007.0 + 7303 & 10 - 30 & @xmath121 & @xmath122 & & & @xmath123 & ( 7 ) + & ( in cta 1 ) + 8 & psr b0833 - 45 & 11 - 25 & @xmath124 & @xmath125 & @xmath126 & @xmath127 & @xmath128 & ( 8) + & ( vela ) + 9 & psr b1706 - 44 & @xmath129 & @xmath130 & @xmath131 & & & @xmath132 & ( 9 ) + & ( in g343.1 - 02.3 ) + 10 & psr b1823 - 13 & @xmath133 & @xmath134 & @xmath135 & & & @xmath1360.30 & ( 10 ) + 11 & psr j0538 + 2817 & @xmath137 & @xmath138 & @xmath139 & & & @xmath140 & ( 11 ) + & ( in s147 ) + 12 & psr b2334 + 61 & @xmath141 & @xmath142 & @xmath143 & & & @xmath144 & ( 12 ) + & ( in g114.3 + 0.3 ) + 13 & psr b1916 + 14 & @xmath145 & @xmath146 & @xmath147 & & & @xmath148 & ( 2 ) + 14 & psr b0656 + 14 & @xmath149 & @xmath150 & @xmath151 & @xmath152 & @xmath153 & @xmath154 & ( 13 ) + & ( in monogem ring ) + 15 & psr j0633 + 1746 & @xmath155 & @xmath156 & @xmath157 & @xmath158 & @xmath159 & @xmath160 & ( 14 ) + & ( geminga ) + 15@xmath161 & & & 0.482 & @xmath162 & & & & ( 15 ) + 16 & psr b1055 - 52 & @xmath163 & @xmath164 & @xmath165 & @xmath166 & @xmath167 & @xmath140 & ( 14 ) + 17 & psr j2043 + 2740 & @xmath168 & @xmath169 & @xmath170 & & & @xmath171 & ( 16 ) + 18 & 1e 1207.4 - 5209 & @xmath172 & @xmath173 & @xmath174 & @xmath175 & @xmath176 & @xmath177 & ( 17,18 ) + & ( in pks 1209 - 51/52 ) + 19 & cxou j232327.9 & 0.3 & @xmath178 & @xmath179 & & & @xmath180 & ( 19 ) + & + 584842 ( in cas a ) + 20 & cxou j085201.4 & @xmath181 & @xmath182 & @xmath183 & & & @xmath184 & ( 20 ) + & -461753 ( in g266.2 - 1.2 ) + 21 & psr j1852 + 0040 & & @xmath185 & @xmath186 & & & @xmath187 & ( 21 ) + & ( in kes 79 ) + 22 & psr j1713 - 3949 & & 4.4 & 2.4 & & & 15 & ( 22 ) + & ( in g347.3 - 0.5 ) + 23 & rx j1856.5 - 3754 & @xmath188 & @xmath189 & @xmath190 & & & @xmath191 & ( 23 ) + 24 & rx j0720.4 - 3125 & @xmath192 & @xmath193 & @xmath194 & & & @xmath195 & ( 24 ) + 25 & rbs 1223@xmath196 & @xmath197 & 1.04 & 0.8 & & & @xmath198 & ( 25 ) + 26 & rx j0420.0 - 5022 & @xmath149 & @xmath199 & 1.4 & & & 2.7@xmath200 & ( 26 ) + 27 & rx j0806.4 - 4123 & & @xmath201 & @xmath1360.6 & & & @xmath202 & ( 27 ) + 28 & rx j1605.3 + 3249 & & 1.07 & @xmath1361.1 & & & @xmath203 & ( 25 ) + 29 & rbs 1774@xmath204 & & 1.04 & @xmath1361.1 & & & @xmath203 & ( 25 ) + + @xmath196 1rxs j130848.6 + 212708 @xmath204 1rxs j214303.7 + 065419 diagram for the x - ray pulsar sample . the solid lines indicate the death lines both for a typical pulsar ( with mass of 1.4@xmath205 and a radius of 10 km ) and low - mass sqss ( with mass of @xmath206 ) , with surface magnetic field @xmath207 g and potential drop in open field line region @xmath208 v. the death lines move up if decreasing @xmath209 and/or increasing @xmath210 @xcite . the hexagonals indicate the focused sources associated with detectable supernova remnants ( snrs ) , while the pentagons indicate those which exhibit without evident snrs . the triangles indicate 4 xdins with measurable temporal parameters , and the diamonds mark the ccos . the timing parameters for rx j0822 - 4300 are from @xcite , while the rest are from atnf pulsar catalog@xmath211 .,scaledwidth=70.0% ] we concentrate on those x - ray sources , which demonstrate significant thermal emission , own ordinary magnetic fields @xmath212 g , own comparatively young ages @xmath213 yrs , and have spins of a few tens of milliseconds to a few seconds . we mainly refer to the collation made by @xcite on cooling neutron stars , by @xcite on x - ray pulsars , by @xcite on x - ray dim isolated neutron stars ( xdins ) , and by @xcite on central compact objects ( ccos ) . the sample exhibited in table [ tab : data ] thus comprises top 17 active pulsar candidates , 7 xdins ( no . 23 - 29 ) and 6 ccos ( no . 5 and 18 - 22 ) . we note that the values of the surface temperatures adopted are determined by black body fits , according to the way that a sqs emits thermal photons . the xdins and ccos are considered to be magnetosphere - inactive pulsar candidates , since their quiescent manifestations on nonthermal radiation . ccos , moreover , would only be seen in the soft x - ray band without the evidences on the existences of the associated pwne . additionally , we note that the emission of the energetic crab pulsar is much likely to be overwhelmed by the nonthermal component originating from its luminous plerion , which hampers the detection to the stellar surface thermal radiation . as an estimate , we , however , adopted a @xmath214 upper limit to crab s surface temperature and an inferred radius with assuming a 2 kpc distance between crab and the earth @xcite . rx j0822 - 4300 , the central stellar remnant in puppis a , used to be analyzed by @xcite basing on the _ rosat _ observations in 1990s . they could not confirm whether the x - ray structures surrounding the star belong to the supernova remnant ( snr ) or are induced by the probable active magnetosphere . recent observations on it by _ chandra _ and _ xmm - newton _ telescopes , however , did not reveal the presence of the associated plerion , indicating an inert magnetosphere @xcite . because of the ambiguity of this source , we temporarily make it as one of the active source candidates . furthermore , we emphasize these x - ray sources that own measurable spin parameters in the @xmath215 diagram ( fig . [ fig : ppdot ] ) , by which a distribution of them can be read . other 10 x - ray sources with constraints on the upper limits of their bolometric luminosities are also denoted in the diagram @xcite . besides the death line for a typical pulsar , those for low - mass sqss are also indicated in the diagram . these death lines set boundaries of the ` graveyards ' for pulsars with different mass @xcite . the timing parameters are from atnf pulsar catalog @xcite , except for rx j0822 - 4300 , which refers to @xcite . for magnetosphere - active pulsar candidates , sqss would suggest themselves to reproduce the cooling processes , while , for magnetosphere - inactive pulsar candidates , possible approaches to understand their current x - ray luminosities would also be proposed by sqss . these two will be presented in 3.1 and 3.2 . clllcc + no . & source & @xmath216 ( s@xmath100 ) & @xmath217 ( s@xmath218 ) & @xmath219 ( ergs s@xmath100 ) & @xmath220 + + 1 & psr b0531 + 21 ( crab ) & 30.225437 & -3.862@xmath221 & 7.7@xmath222 & 4.1@xmath223 + 2 & psr j1811 - 1925 & 15.463838 & -1.052@xmath224 & & + 3 & psr j0205 + 6449 & 15.223856 & -4.495@xmath224 & 5.8@xmath225 & 4.6@xmath223 + 4 & psr j1119 - 6127 & 2.452508 & -2.419@xmath224 & 1.7@xmath226 & 8.5@xmath227 + 5 & rx j0822 - 4300 & 13.2856716499(3 ) & -2.6317(3)@xmath224 & 2.7@xmath226 & 2.8@xmath227 + 6 & psr j1357 - 6429 & 6.020168 & -1.305@xmath224 & 4.4@xmath225 & 1.2@xmath227 + 7 & rx j0007.0 + 7303 & 3.165922 & -3.623@xmath228 & 3.5@xmath229 & 8.8@xmath223 + 8 & psr b0833 - 45 ( vela ) & 11.194650 & -1.567@xmath224 & 1.4@xmath225 & 4.5@xmath223 + 9 & psr b1706 - 44 & 9.759978 & -8.857@xmath228 & 4.2@xmath225 & 1.1@xmath227 + 10 & psr b1823 - 13 & 9.855532 & -7.291@xmath228 & 3.2@xmath225 & 1.1@xmath227 + 11 & psr j0538 + 2817 & 6.985276 & -1.790@xmath230 & 4.3@xmath226 & 9.3@xmath200 + 12 & psr b2334 + 61 & 2.018977 & -7.816@xmath230 & 1.8@xmath222 & 1.7@xmath200 + 13 & psr b1916 + 14 & 0.846723 & -1.523@xmath230 & 1.8@xmath222 & 5.9@xmath200 + 14 & psr b0656 + 14 & 2.598137 & -3.713@xmath230 & 1.1@xmath231 & 1.7@xmath232 + 15 & psr j0633 + 1746 ( geminga ) & 4.217640 & -1.952@xmath230 & 3.2@xmath225 & 9.9@xmath227 + 16 & psr b1055 - 52 & 5.073371 & -1.501@xmath230 & 6.9@xmath226 & 1.5@xmath232 + 17 & psr j2043 + 2740 & 10.402519 & -1.374@xmath230 & 7.1@xmath229 & 3.5@xmath227 + + @xmath196 @xmath233 g @xmath234 ) if two isotropic black body emission components the hot component for polar caps and the warm one for the bulk of the star are defined for sqss , then the equation describing cooling processes would be written as ( cf . equation ( [ eq : coolinggeneral ] ) ) @xmath235 and the relation between @xmath14 and @xmath66 is given by equation ( [ eq : h ] ) . the luminosity of stellar heating @xmath12 could either follow the 1/2-law or the linear - law ( see appendix a ) . cooling behaviors would thus be calculated by assuming sqss rotate as orthogonal rotators and slow down as a result of magnetic dipole radiation with magnetic field strength at the poles of @xmath236 g. we provided a parameter space to fit the observational temperature - age data , and the comparison between the observations and expectations are shown in figs . [ fig : ccos ] ( for 1/2-law case ) and [ fig : cclr ] ( for linear - law case ) . some pulsars demonstrate two black body components in their thermal spectra , e.g. vela pulsar and the three musketeers , implying that the temperature inhomogeneity on these pulsars could be significant . we thus carried out a temperature - difference fit at the same time , as have been shown in the _ right _ panels in both figs . [ fig : ccos ] and [ fig : cclr ] . it is worthy of noting that , for geminga , if the photon index of the power - law ( pl ) component is thawed when fitting its phase - resolved spectra , the entry of the hot black body component could not improve the fits remarkably , or it may even becomes an artifact . this may mean the fluctuation of the magnetospheric emission during a spin cycle might mislead the understanding on its x - ray spectra , as analyzed by @xcite and their results are denoted by no . 15@xmath161 in table [ tab : data ] . if this situation holds , the surface temperature fluctuation on geminga could be tiny so that undetectable . we , hence , set a rough upper limit to the temperature difference for geminga by an order lower than its surface temperature , as denoted by 15@xmath161 in the ` temperature difference - age ' figures . the proposed relation between the rotational kinetic energy loss rate and the bolometric luminosity could inversely provide direct measurement to the moments of inertia of the active x - ray sources . the results are exhibited in table [ tab : moi ] and could be understood in the sqs regime , especially low - mass sqss ( cf . the adoption of the coefficients @xmath237 and @xmath238 in figs . [ fig : ccos ] and [ fig : cclr ] ) . we note that for sqss with 0.01@xmath205 , 0.1@xmath205 , 1.0@xmath205 and 2.0@xmath205 , the moments of inertia @xmath239 ( i.e. values scaled by @xmath240 g @xmath241 ) are @xmath242 , @xmath620.01 , @xmath620.55 and @xmath621.74 respectively . ccc + mass & @xmath243 @xmath205 yr@xmath100 ) & @xmath244 ergs s@xmath100 ) + + @xmath245 & 2.7 & 1.7@xmath246 + @xmath247 & 5.7 & 4.7@xmath246 + @xmath248 & 12.4 & 2.0 + + as has been discussed in 2.3.2 , one probable energy source to be responsible for the soft x - ray emission of magnetosphere - inactive pulsar candidates lies in the accretion in the _ propeller _ regime . the accretion could be either to the interstellar medium or to the fallback disk . in this regime , however , the stellar magnetosphere radius could hardly be given , even though the sources spindown rates are detectable . because the spindown rate in this case would be dominated by the accretion behavior . equation ( [ eq : coolinggeneral ] ) for this case could be specified as @xmath249 with considering equation ( [ eq : polarpowerd ] ) and the stellar residual thermal energy is too small to be taken into account . table [ tab : lum ] lists the luminosities of sqss under the accretion scenario . it could be concluded that the sources current observational thermal x - ray luminosities would be interpreted by considering the parameters @xmath88 , the accretion rate in the unit of that of eddington , and @xmath78 , the accretion efficiency , such as @xmath250 , @xmath251 . we collate the thermal observations of 29 x - ray isolated pulsars and , in the sqs regime , for the magnetospherically active pulsar candidates , establish their cooling processes ( figs . [ fig : ccos ] and [ fig : cclr ] ) , while for the magnetospherically inactive or dead pulsar candidates , interpret the x - ray luminosities under the accretion scenario . ( table [ tab : lum ] ) . fitting to the thermal spectra for x - ray pulsars using the black body model often results in such small emission sizes that could naturally be interpreted by the sqs model . a recent study on the cassiopeia a cco shows that the fitted emission size is still significantly smaller than a typical neutron star radius , even if the hydrogen atmosphere model is employed in the fitting @xcite . the uncertain estimates on the distances of pulsars may introduce considerable errors in the fitting about the x - ray source emission sizes . however , if these obtained emission radii can be trusted , sqss , because of the possibility of being low - mass , could provide an approach to understand these observational manifestations . we note that for sqss with mass of @xmath252 , @xmath253 and @xmath254 , their radii are @xmath621.8 , @xmath623.8 and @xmath628.3 km respectively . on the other hand , a linkage between pulsar rotational kinetic energy loss rates and bolometric x - ray luminosities is explored by sqss ( see appendix a ) , and the resulting estimates on pulsar moments of inertia is exhibited in table [ tab : moi ] . we hence conclude that the phenomenological sqs pulsar model could not be ruled out by the thermal observations on x - ray isolated pulsars , though a full depiction on the thermal evolution for a quark star in all stages could hardly be given nowadays , as the lack of the physics in some extreme conditions ( as has been discussed in 2.1 ) . sqss have significant distinguishable interiors with neutron stars , but the structures of the magnetosphere between these two pulsar models could be similar . therefore , sqss and neutron stars would have similar heating mechanisms , such as the bombardment by backflowing particles and the accretion to surrounding medium . however , a full comparison between sqss and neutron stars including their cooling processes as well as the heating mechanisms has beyond the scope of this paper , and should be interesting and necessary in the future study . the various performances of x - ray pulsars may indicate their current states or properties and imply their possible evolutionary history . in addition , if these x - ray sources are actually sqss , the possible formation mechanism could be an interesting topic . we thus extend a discussion in 4.3 . spin has always been a significant energy source for active pulsars , by which multiwave bands nonthermal radiation are driven , including those of the pulsar themselves as well as those of the surrounding plerions . some phenomenological studies demonstrate certain regularities for such an active pulsar population . a brief summary is that the nonthermal x - ray luminosity and spin energy loss rate own a relation of @xmath255 @xcite , and the @xmath256-ray luminosity is proportional to the square root of @xmath68 or @xmath257 @xcite . we , additionally , note that @xmath258 , where @xmath259 is the potential drop along the open field region . therefore , it seems that for such a population the more younger the larger @xmath68 and @xmath259 that a pulsar owns ( cf . [ fig : edot ] , _ bottom _ panel ) and the more luminous of the radiation in the hard x - ray and @xmath256-ray bands . besides nonthermal emission , spin may also be an energy origin for pulsar thermal radiation as the relations of @xmath2 or @xmath260 could exist observationally ( see appendix a for details ) . sqss predominantly follow these relations to accomplish their cooling processes , since their residual thermal energy could be quite inadequate to sustain a long term x - ray thermal emission . xdins and ccos could be representative populations of magnetosphere - inactive pulsars . nevertheless , their energy origin that could power the x - ray emission is still an enigma . the bolometric luminosities of 1e 1207.4 - 5209 , rx j1856.5 - 3754 and rx j0720.4 - 3125 exceed their spin energy loss rates by a factor of @xmath261 , @xmath262 and @xmath263 , respectively ; they appear not to be powered by spins . ccos could be a group of weakly - magnetized and long - initial - period pulsars @xcite , because of which their potential drops along the open field lines could be much less than @xmath264 v , so that the primary plasma could not be fully accelerated , resulting in inactive magnetospheres . the black body fits to the thermal spectra of ccos always result in extremely small emission sizes @xcite , implying the possible existence of low - mass sqss . therefore , the @xmath68 of ccos might be very low , so that effective magnetospheric emission and plerions could not be driven . ccos could thus be the representatives of a group of natal inactive pulsars . accretion to the fallback ejecta of the associated snrs could power their soft x - ray luminosities , and this scenario have been mentioned by several authors ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? xdins , owning comparatively larger spindown ages and stronger magnetic fields , might be the descendants of magnetars . considering similar properties between them @xcite , one could not exclude that xdins are still being powered by the decaying magnetic fields . however , observations have revealed the probable existence of residual disks around such pulsars , and thus their radiation may be of accretion origin @xcite . in this case , xdins could be the evolved products of active pulsars , and thanks to the accretion so that they are still visible after their death . most xdins have absorption features in their spectra , being similar to those of ccos . hence , if they are really undergoing accretion in the propeller regime , their spindown rates @xmath265 could then be of accretion origin rather than causing by the magnetic dipole braking . thus their magnetic fields could be much lower than the values determined by canonical magnetic dipole radiation , since the absorption features could be electron cyclotron lines . therefore , the common properties of absorption features between xdins and ccos could imply that xdins might be older ccos . xdins and ccos have not manifested themselves as radio pulsars , their radio - quiet demonstrations could be an intrinsic property rather than being of beaming origin , since they might be dead or natal inactive pulsars . so they could be grouped as radio - quiescent pulsars . for rx j0007.0 + 7303 , an active pulsar candidate , its lack of radio signal could be the result of an unfavorable geometry , or its radio beam sweeps away from the earth as has been analyzed by @xcite . the similar thermal manifestations of accreting pulsars ( e.g. xdins and ccos ) and cooling pulsars may have caused a confusion about distinguishing between the two classes @xcite . however , the activeness of the magnetosphere could provide a way to achieve an identification . if pulsars are in fact sqss with rapid rotation , cooling pulsars are then likely to be undergoing the spindown - powered heating evolution , and multi - bands nonthermal emission originating from luminous magnetospheres and even pwne will accompany such coolers during the processes . in contrast , accreting pulsars may have not demonstrated such characteristics observationally . in this context , the cco in the snr cassiopeia a ( source no . 19 in table [ tab : data ] ) should be an accretor because of the lack of pwn and nonthermal power - law component @xcite . possible low - mass compact stars is a direct consequence of the suggestion that pulsars could be quark stars @xcite . however , the probable existence of quark stars , spanning large mass range , could naturally raise such a question : how did these stars be created ? indeed , this could be another severe problem given by the observations . being different from solar - mass quark stars ( which could originate from core - collapse supernovae ) , low - mass quark stars could be the central remnants left by the detonation of the accretion - induced collapse ( aic ) of white dwarfs ( wds ) @xcite . the combustion from the hadronic matter to the strange quark matter could be the engine of such an explosion . during this phase transition , as high as @xmath266 mev per baryon could be released as the latent heat @xcite . the combustion could hence be in such a high efficiency that % 10 of the rest mass is liberated . for a wd approaching the chandrasekhar limit , its mass and radius are @xmath267 , @xmath268 cm respectively . the gravitational energy is then @xmath269 ergs . for a successful explosion of such a wd , the necessary mass of a quark star @xmath270 would hence be obtained by @xmath271 as @xmath272 . so , in this case , low - mass quark stars with a few tenth of or a few hundredth of solar mass could form , although one should be aware that the exact value of the quark star mass depends on the region where the detonation wave inside the wd can sweep though as well as the latent heat of the phase transition , which would be strongly model - dependent . these nascent quark stars would keep bare if they has not suffered massive super - eddington accretion @xcite . an accreted ion ( e.g. a proton ) should gain enough kinetic energy to penetrate the coulomb barrier of a quark star with mass @xmath273 and radius @xmath274 , or @xmath275 . the coulomb barrier @xmath276 is model - dependent , which could varies from @xmath7 mev to even @xmath277 mev @xcite . considering @xmath278 for low - mass quark stars , a mass lower - limit @xmath279 would then be determined as @xmath280 where the density @xmath281 , with @xmath282 the nuclear density , and @xmath283 mev . apart from the formation of the low - mass quark stars discussed here , other astrophysical processes , such as the collision between two strange stars , could also be the potential sources to give birth to the quark stars with smaller mass than that of the sun @xcite . ( _ top _ panel ) and @xmath284 ( _ bottom _ panel ) of active pulsar candidates . in the _ top _ panel , the fits are carried out both for group a ( including top 17 pulsars listed in table [ tab : data ] ; they are marked by dark points and their numbers ) and group b ( including the members in group a and the other 10 sources whose upper limits on the bolometric luminosities could be defined observationally ; the 10 sources are marked by grey points and letters ) . red lines are the fitted results for group a , while blue lines are those for group b. in both groups , solid lines provide the best fits to the data , while the dashed lines give the fits by freezing @xmath285 at 1 . we note that the 10 sources with upper limits on their bolometric luminosities defined are taken from @xcite , and they are a. b1509 - 58 , b. b1951 + 32 , c. b1046 - 58 , d. b1259 - 63 , e. b1800 - 21 , f. b1929 + 10 , g. b0540 - 69 , h. b0950 + 08 , i. b0355 + 54 , j. b0823 + 26.,scaledwidth=70.0% ] ccccc + group & @xmath285@xmath196 & @xmath286 & corr . coef.@xmath204 & @xmath287 ( d.o.f)@xmath288 + + a. & @xmath289 & @xmath290 & & 0.3277(14 ) + & ( 1 ) & @xmath291 & 0.7487 & 0.8607(15 ) + b. & @xmath292 & @xmath293 & & 0.6081(24 ) + & ( 1 ) & @xmath294 & 0.7730 & 0.9964(25 ) + + @xmath196 values in the parenthesis are frozen during the fits . the errors are in 95% confidence level . @xmath204 correlation coefficient of the data between bolometric luminosity and @xmath68.@xmath288 reduced @xmath295 , or @xmath295 per degree of freedom ( d.o.f ) . in the solid quark star(sqs ) regime , cooling of magnetosphere - active pulsars depends predominantly on the stellar heating as a result of the lack of sufficient residual heat as has been analyzed in 2.2 . the pulsar activity induced stellar heating would have intrinsically set up a linkage between the x - ray bolometric luminosity and the spin energy loss rate . we , hence , present here a phenomenological study on such a relation . the sample of active pulsars firstly includes the top 17 sources listed in table [ tab : data ] . @xcite summarized x - ray pulsars , in which other 10 sources upper limits on the bolometric luminosities could be defined . these 10 pulsars are meanwhile considered . [ fig : edot ] ( _ top _ panel ) illustrates these sources bolometric luminosities @xmath296 as a function of their spin energy loss rates @xmath68 , and table [ tab : fit ] lists the fitted parameters led by the function of @xmath297 the fits are carried out both only for the top 17 pulsars taken from table [ tab : data ] ( as group a ) and for the whole sample with the other 10 sources included at the same time(as group b ) . the best fits , moderate clearly , suggest a 1/2-law on the relation between pulsar bolometric x - ray luminosities and the spin energy loss rates , or @xmath298 with a coefficient @xmath299 in the unit of ergs@xmath300 s@xmath301 . if @xmath285 is fixed at 1 , a linear - law is then implied , i.e. @xmath302 with a coefficient @xmath303 , or @xmath238 is the conversion efficiency . the fits obtain @xmath304 , which is similar to the nonthermal x - ray case @xcite . for the 1/2-law case , the conversion efficiency turns out to be the function of @xmath68 , which would be @xmath305 which could in turn , besides the fits , provide a reference to the value of the coefficient @xmath237 by a natural constraint of @xmath306 . taking b0823 + 26 , the minimum @xmath68 in our sample as an example , its @xmath237 would be less than @xmath307 ergs@xmath300 s@xmath301 . additionally , we note here that we also present a age-@xmath68 relation in the _ bottom _ panel of fig . [ fig : edot ] , which illustrates a natural trend for spin - powered pulsars . the authors are grateful to prof . fredrick jenet for his valuable comments and advices , especially for his efforts on improving the english expression of this paper . we wish to thank mr . weiwei zhu for his detailed introduction to the observational situation about the sources psr j1811 - 1925 and psr b1916 + 14 . the authors also thank the colleagues in the pulsar group of peking university for the helpful discussion . this work is supported by nsfc ( 10778611 , 10973002 ) , the national basic research program of china ( grant 2009cb824800 ) , and by lcwr ( lhxz200602 ) .	we present a theoretical model for the thermal x - ray emission properties and cooling behaviors of isolated pulsars , assuming that pulsars are solid quark stars . we calculate the heat capacity for such a quark star , including the component of the crystalline lattice and that of the extremely relativistic electron gas . the results show that the residual thermal energy can not sustain the observed thermal x - ray luminosities seen in typical isolated x - ray pulsars . we conclude that other heating mechanisms must be in operation if the pulsars are in fact solid quark stars . two possible heating mechanisms are explored . firstly , for pulsars with little magnetospheric activities , accretion from the interstellar medium or from the material in the associated supernova remnants may power the observed thermal emission . in the propeller regime , a disk - accretion rate @xmath01 % of the eddington rate with an accretion onto the stellar surface at a rate of @xmath1 could explain the observed emission luminosities of the dim isolated neutron stars and the central compact objects . secondly , for pulsars with significant magnetospheric activities , the pulsar spindown luminosities may have been as the sources of the thermal energy via reversing plasma current flows . a phenomenological study between pulsar bolometric x - ray luminosities and the spin energy loss rates presents the probable existence of a 1/2-law or a linear law , i.e. @xmath2 or @xmath3 . this result together with the thermal properties of solid quark stars allow us to calculate the thermal evolution of such stars . thermal evolution curves , or cooling curves , are calculated and compared with the ` temperature - age ' data obtained from 17 active x - ray pulsars . it is shown that the bolometric x - ray observations of these sources are consistent with the solid quark star pulsar model . _ pacs : _ 97.60.gb , 97.60.jd , 95.30.cq pulsars ; neutron stars ; elementary particles
in experiments on ultracold trapped fermi gases , there are many situations where the system is out of thermal equilibrium . the first one is of course the trapping and cooling stage , i.e. , before the system has reached its equilibrium state which is usually the starting point for the actual experiment . then , in some experiments the system is excited in order to observe its dynamical behavior . for instance , many experiments studied collective oscillations of the system @xcite , another example being a recent experiment at mit where the collision of two atom clouds ( both in equilibrium ) was studied @xcite . finally , often the system is not imaged directly during the experiment , but only after the trap was switched off and the system has expanded for a certain time , in order to increase its size . the modeling of such time - dependent processes from the theoretical point of view can be quite complicated . for practical reasons , only semiclassical approaches are suitable for the description of time - dependent phenomena involving typically several @xmath0 atoms in a three - dimensional , non - uniform geometry . in some cases , it is possible to use hydrodynamic approaches : superfluid hydrodynamics describes the expansion @xcite and the collective modes @xcite of superfluid systems at zero temperature . hydrodynamics is also applicable in the normal - fluid phase if the mean time between collisions is much shorter than all other time scales of the process under consideration , so that the system can always be considered to be in a local equilibrium @xcite . superfluid and normal hydrodynamics can be combined to two - fluid hydrodynamics in order to describe superfluid systems at finite temperature @xcite . however , in many cases hydrodynamic approaches are not sufficient . in all cases where it is important that even locally the distribution @xmath1 of the atoms is not an equilibrium one , the boltzmann equation allows a very general description , provided the system is in the normal phase . if the system is superfluid , a more elaborate theory is necessary which couples the dynamics of the quasiparticle distribution function to the dynamics of the superfluid order parameter @xcite . in the past , several authors used the boltzmann equation for the investigation of collective oscillations in normal - fluid trapped fermi gases @xcite . in most cases , the boltzmann equation was not solved directly in order to find the distribution function @xmath2 , but semi - analytical approximate solutions were found by using the scaling ansatz @xcite or the method of moments @xcite . these methods rely explicitly or implicitly on the assumption that the collision term can be treated in the relaxation - time approximation , with a single relaxation time @xmath3 which is independent of the position in the trap . an exception is the work by toschi et al . @xcite , where the boltzmann equation was solved numerically , using a test - particle method very similar to the one we are using here . the test - particle method for the solution of the boltzmann equation has been used for many years in nuclear physics for the simulation of heavy - ion collisions @xcite . in the context of trapped atoms , it was also used for the simulation of the dynamics of the thermal cloud in a bose - einstein condensate @xcite and ( without collision term ) of the normal - fluid component in a superfluid fermi gas @xcite . for the collision term of the boltzmann equation , it is important to know the cross section which in principle can be modfied by in - medium effects . in the work by riedl et al . @xcite , it was shown that by using in the boltzmann equation instead of the free cross - section the in - medium one , the agreement between theoretical and experimental frequencies and damping rates of different collective modes is deteriorated . the reason is that the in - medium cross section is larger than the free one , so that the relaxation time is dramatically reduced . in our previous work @xcite , our aim was to include in addition to the in - medium cross section medium effects into the mean - field potential . this mean field resulted in better density profiles and allowed us to understand the shift of the quadrupole frequency in the collisionless regime at very low temperature observed in @xcite . however , it did not help to improve the agreement between theory and experiment in the region of higher temperatures , where the properties of the collective modes are completely dominated by collisions . hence , one of our motivations for the present work was to check the validity of the relaxation - time approximation which is implicitly made in the method of moments . here we will restrict ourselves to the case without mean field and with the free cross section . as we will show in sec . [ sec : modes ] , the numerical solution of the boltzmann equation gives indeed a significantly longer relaxation time than the method of moments . as we will show , this discrepancy is due to the restriction of the method of moments to second - order moments in the existing literature @xcite . once fourth - order moments are included , the results of the method of moments and of the numerical solution are in good agreement . however , already in the simplest case of a spherical harmonic trap without mean field , the inclusion of fourth - order moments is a very tedious task , while the numerical method can be generalized to more realistic cases . in addition , there are some other reasons why we felt the necessity for a numerical method . for example , there are damping effects due to the anharmonicity of the trap potential which can not be described by the method of moments . another advantage of the numerical method is that it offers the possibility to simulate not only the oscillation of the cloud , but also the subsequent expansion after the trap has been switched off . in sec . [ sec : numericaldescription ] of the present paper , we give a detailed description of the method . in particular , we explain in detail how the collisions are simulated , since our method is somewhat different from that of ref . moreover we discuss some tests we made in order to estimate up to which precision we can trust our simulation . then , in sec . [ sec : modes ] , we come to the main point of our article and calculate the properties of some collective modes for a system in a spherical harmonic trap . while the sloshing and breathing modes are rather trivial , the frequency and damping rate of the quadrupole mode are very sensitive to the collisions . we compare the numerical results with those of the method of moments . finally , in sec . [ sec : conclusions ] we summarize and give an outlook to future studies . throughout the paper , we will use units with @xmath4 ( @xmath5 reduced planck constant , @xmath6 = boltzmann constant ) . the strength of the interaction is characterized by the dimensionless quantity @xmath7 , where @xmath8 is the scattering length . concerning the fermi momentum @xmath9 and the fermi energy @xmath10 we follow the usual convention that these quantities are defined by the corresponding ones of an ideal fermi gas at zero temperature , i.e. , @xmath11 , @xmath12 being the atomic mass , and @xmath13 , where @xmath14 is the number of atoms and @xmath15 the trap frequency . temperatures will be measured in units of the fermi temperature @xmath16 ( since @xmath17 ) . we study a two - component ( @xmath18 ) gas of fermionic atoms of mass @xmath12 in a potential @xmath19 with attractive interaction @xmath20 . we assume that the system is in the normal phase and that it can be described semiclassically by phase - space distribution functions @xmath21 . in this paper , we will restrict ourselves to the case that the distribution functions of both spin states are equal ( @xmath22 ) , but the generalization of the method to the cases of different distribution functions , more than two components , or components with different masses is straight - forward . the time evolution of the distribution function @xmath2 is governed by the boltzmann equation @xcite @xmath23\,,\ ] ] where the left - hand side ( lhs ) describes the particle propagation , with @xmath24 and @xmath25 $ ] on the right - hand side ( rhs ) denotes the collision term which will be discussed later . the potential felt by the particles is the trap potential that contains a static part and a time dependent one ( which will be used to simulate the excitation of the collective modes ) @xmath26 . the density per spin state is related to the distribution function by @xmath27 and the number of atoms is given by @xmath28 the basic idea of the test - particle method ( also called pseudoparticle method ) for solving the boltzmann equation consists in replacing the continuous distribution function by a sum of delta functions , @xmath29 where @xmath30 is the number of `` test particles '' . this allows one to express the average of an arbitrary single - particle observable @xmath31 in the simple form @xmath32 in order to sample the six - dimensional phase space , it is necessary to choose a sufficiently large number of test particles @xmath30 ( usually @xmath33 ) . neglecting the collision term @xmath25 $ ] for the moment , it is easy to see that eq . ( [ eq : defdistri ] ) satisfies the boltzmann equation ( eq . ( [ eq : eqboltzmann ] ) ) if the positions @xmath34 and momenta @xmath35 of each test particle @xmath36 follow the classical equations of motion , eq . ( [ eq : eqmotion ] ) . in practice , the delta functions in eq . ( [ eq : defdistri ] ) can pose some problems . for instance , they do not result in a continuous density @xmath37 . therefore it is often useful to replace them by gaussians of width @xmath38 and @xmath39 in position and momentum space , respectively : @xmath40 with @xmath41 the widths @xmath38 and @xmath39 must be adapted such that they smooth out the fluctuations due to the finite number of test particles , but not the structure of the distribution function @xmath2 . the statistical fluctuations are of the order of @xmath42 , i.e. , the first condition is equivalent to @xmath43 the second condition implies of course that @xmath44 and @xmath45 , where @xmath46 and @xmath47 are the thomas - fermi radius and the fermi momentum , respectively , but this is not always sufficient . at low - temperature , it is crucial to resolve the rapid change of the distribution function around the fermi surface , i.e. , @xmath48 in practice , as the computation time increases as @xmath49 , it turns out that the conditions ( [ conditionn ] ) and ( [ conditiont ] ) can not simultaneously be satisfied at too low temperatures . in the absence of collisions , the numerical task consists only in solving simultaneously the classical equations of motion ( [ eq : eqmotion ] ) for the @xmath30 test particles . we do this by using the velocity verlet algorithm @xcite , which contrary to the original verlet algorithm @xcite uses the positions @xmath50 and velocities @xmath51 as starting point for the time step from @xmath52 to @xmath53 . the propagation from @xmath52 to @xmath54 is done according to @xmath55 where @xmath56 is the acceleration of the @xmath36-th test particle . if it is written in this way , it is obvious that the velocity verlet algorithm is identical to the leap - frog algorithm @xcite . note that the accelerations @xmath57 can be reused in the next time step , so that the algorithm needs only one evaluation of the acceleration per time step , but nevertheless its global error is of the order @xmath58 . this allows us to obtain a good accuracy for reasonable time steps @xmath59 . a good test of the particle propagation is to check the energy conservation : typically we find @xmath60 for each test particle and for all times considered . the rhs of the boltzmann equation ( [ eq : eqboltzmann ] ) describes the collisions between particles of opposite spin . it thus depends on the scattering cross section @xmath61 and reads @xcite @xmath62=\int \frac{d^3 p_1}{(2 \pi)^3 } \int d\omega \frac{d\sigma}{d\omega } -f^\prime f_1^\prime ( 1-f ) ( 1-f_1 ) ] \ , . \label{eq : collisionterm}\end{gathered}\ ] ] in the first term , @xmath63 and @xmath64 are the incoming momenta , @xmath65 and @xmath66 are the outgoing ones , @xmath67 is the solid angle formed by the incoming relative momentum @xmath68 and the outgoing relative momentum @xmath69 , and @xmath70 , @xmath71 , etc . in the second term , the role of incoming and outgoing momenta is exchanged . momentum and energy conservation implies @xmath72 and @xmath73-wave scattering , in which the cross section is isotropic , i.e. , @xmath74 . in principle the cross section is modified by medium effects @xcite , but in the present paper we will only use the free cross - section ( i.e. , the cross - section for the scattering of two atoms of opposite spin in free space ) which is given by @xcite @xmath75 where @xmath76 . in our numerical simulation , the collision term is included by allowing the test particles to collide with each other . the cross section of the test particles , @xmath77 , is related to the cross section of the atoms by @xmath78 ( since @xmath30 test particles represent @xmath79 atoms of a given spin ) . whether a pair @xmath80 of test particles collides in a time step @xmath52 or not is determined as follows : first , we determine if the two particles are at their closest approach in the present time step . explicitly , if we write @xmath81 and @xmath82 , the closest approach is reached at @xmath83 and we check if @xmath84 . if yes , we calculate the corresponding minimal distance by @xmath85 and check if @xmath86 . in this case , the collision is classically allowed . we then propagate both test particles to @xmath87 , change the direction of their relative velocity @xmath88 in a random way ( thus conserving the total momentum and the total energy ) , and propagate them back to the original time @xmath52 . finally , in order to take into account the pauli - blocking factors in eq . ( [ eq : collisionterm ] ) , we calculate the occupation numbers @xmath89 and @xmath90 at the new positions and momenta ( @xmath91 etc . ) using eq . ( [ eq : defdistri ] ) with gaussians instead of delta functions , see eq . ( [ eq : gaussians ] ) . with probability @xmath92 the collision is allowed and we keep the new positions and momenta , otherwise the collision is blocked and we keep the old ones . we checked that the total energy is still well conserved when collisions are switched on : typically we find better than @xmath93 for all times considered . before the simulation can start , the test - particle positions and momenta have to be initialized . here we assume that the system is initially in equilibrium . a suitable equilibrium distribution is given by the distribution function within the thomas - fermi or local - density approximation ( lda ) , @xmath94 since it is a stationary solution of the boltzmann equation @xcite . this distribution has two parameters , namely the inverse temperature @xmath95 and the chemical potential @xmath96 . the temperature @xmath97 is an input parameter , whereas the chemical potential @xmath96 is determined by demanding that the integral of eq . ( [ eq : equilibrium ] ) over @xmath98 and @xmath63 gives the right number of atoms . having determined the chemical potential @xmath96 , we randomly generate the test - particle positions and momenta in such a way that the probability to be at position @xmath98 and to have momentum @xmath63 is proportional to @xmath99 . in practice , we do this by first generating the positions according to the density profile obtained from eq . ( [ eq : density ] ) with @xmath100 . then we generate the momenta according to @xmath101 . in this subsection we describe two main tests we made to be sure that our code is reliable . here , we assume the potential to be static and , as in the rest of the paper , we use a spherical harmonic potential @xmath102 this potential defines naturally a time scale @xmath103 , a length scale @xmath104 , an energy scale @xmath15 , and so on . let us consider the energy distribution of the atoms , @xmath105 in equilibrium , the distribution should be given by @xmath106 , where @xmath107 is the density of states ( including the degeneracy factor 2 ) . in the present case of a spherical harmonic oscillator , we have @xmath108 . in the absence of collisions , energy conservation automatically implies that the distribution stays constant , but in the presence of collisions this test is a non - trivial check of the pauli blocking in the simulation . within the test particle method , @xmath109 is obtained by counting the test particles in energy bins . in fig . [ fig : distri ] energy distribution of the atoms ( divided by the density of states ) for @xmath110 ( see text for details ) . the system consists of @xmath111 atoms with a scattering length @xmath112 ( @xmath113 ) . the parameters of the simulation are : @xmath114 , @xmath115 , @xmath116 , and @xmath117.,width=302 ] we show , for @xmath110 , the initial fermi distribution ( solid line ) and the stationary distribution obtained in the numerical simulation after @xmath118 ( filled circles ) . the agreement between the distribution generated by the simulation and the initial fermi one is not perfect , but satisfactory . in order to show that this is not a trivial result , let us see what happens if we switch off the pauli blocking in the simulation of the collision term . in this case , already after a relatively short time @xmath119 , the distribution in the numerical simulation ( empty circles ) has converged to a boltzmann distribution with the same number of atoms and total energy ( dashed line ) . so , the stability of the fermi distribution in our full simulation shows clearly that pauli blocking is correctly implemented . the small deviations from the ideal fermi distribution are a consequence of the fact that with the chosen widths of the gaussians ( @xmath115 and @xmath116 ) , the condition ( [ conditiont ] ) is not well satisfied at @xmath110 . when we did the same kind of comparison at higher temperatures , we found that the agreement between the simulation and the fermi distribution improves : at @xmath120 , it is already perfect . the test described above is independent of the actual number of collisions . in order to check the latter , let us look at the collision rate @xmath121 , where @xmath122 denotes the number of collisions of test particles per unit time . although in equilibrium the net effect of collisions is zero , the collision rate in equilibrium is a good test for the simulation because it can be compared with the exact result [ eq . ( [ eq : mcblock ] ) , see appendix [ app : collrate ] ] . for testing purposes , it is useful to compare also the total rate of allowed and blocked collisions with the exact result [ eq . ( [ eq : mcnoblock ] ) ] . in fig . [ fig : collrate ] , the collision rates ( with and without blocking ) ) and eq . ( [ eq : mcnoblock ] ) ( solid lines ) , for a gas of 10000 atoms with interaction strength @xmath123 ( top ) and @xmath124 ( bottom).,title="fig:",width=302 ] + ) and eq . ( [ eq : mcnoblock ] ) ( solid lines ) , for a gas of 10000 atoms with interaction strength @xmath123 ( top ) and @xmath124 ( bottom).,title="fig:",width=302 ] of the simulation are shown together with the exact results as functions of the temperature for two different values of the scattering length . in the case of relatively weak interaction , @xmath125 ( upper panel ) , we see that the agreement between the simulation and the exact result is excellent for temperatures above @xmath126 . below that temperature , the collision rate in the simulation with pauli blocking gradually becomes too high since the finite widths of the gaussians ( @xmath127 , @xmath128 ) do not satisfy any more the condition ( [ conditiont ] ) and act in the pauli - blocking factors like an enhanced temperature . in the rest of this paper , we will therefore restrict ourselves to temperatures above @xmath129 . near unitarity ( @xmath130 ) , we consider only temperatures above @xmath131 because this is close to the superfluid transition temperature at unitarity @xcite . as it can be seen in the lower panel of fig . [ fig : collrate ] , the agreement between the collision rate obtained in the simulation and the exact one is satisfactory in the temperature range considered . the agreement is not as good as for @xmath132 at high temperature because of the larger cross section which leads to collisions between test particles which are further apart . the sloshing mode is an oscillation of the center of mass of the system . it plays a special role because in a harmonic trap it is undamped and its frequency is equal to that of the trap , independently of the number of atoms , of the temperature , and of the interaction between the atoms ( kohn mode @xcite ) . this is why it is often used for the experimental determination of the trap frequency @xcite . within the test - particle method , this general theorem is satisfied and it is easy to see why : let us first neglect collisions . from the equations of motion of the individual test particles in the harmonic potential ( [ eq : harmonic ] ) , @xmath133 it is evident that the averages @xmath134 and @xmath135 obey analogous equations of motion , @xmath136 let us now consider the effect of a collision of two test particles . of course , the trajectories of the colliding test particles will not obey any more the original equations of motion ( [ eq : eqmotionharmonic ] ) , but the collision has absolutely no effect on the averages : since the positions do not change during the collision , @xmath134 remains unchanged , and since the total momentum of the two colliding test particles is conserved , the average @xmath135 is not changed either . so , the equations of motion ( [ eq : eqmotionsloshing ] ) for the averages @xmath134 and @xmath135 remain valid in the presence of collisions . their solution is of course an undamped oscillation of the center of mass @xmath134 with frequency @xmath15 . this is confirmed by the numerical result shown in the upper panel of fig . [ fig : sloshingbreathing ] . by displacing all test particles by @xmath137 in the @xmath138 direction . bottom : simulation of the breathing mode . the mode was excited by changing at @xmath139 all test - particle momenta according to @xmath140 ( @xmath141 ) . both simulations were done for a system of @xmath142 particles at @xmath143 and @xmath144.,title="fig:",width=302 ] + by displacing all test particles by @xmath137 in the @xmath138 direction . bottom : simulation of the breathing mode . the mode was excited by changing at @xmath139 all test - particle momenta according to @xmath140 ( @xmath141 ) . both simulations were done for a system of @xmath142 particles at @xmath143 and @xmath144.,title="fig:",width=302 ] a couple of experiments studied the damping of the longitudinal and radial breathing modes @xcite in elongated traps . in a spherical trap , there is only one breathing mode ( monopole mode ) , corresponding to an oscillation of the mean - square radius @xmath145 around its equilibrium value @xmath146 . in a spherical harmonic trap , this mode is undamped and its frequency @xmath147 is independent of the number of collisions , like in the case of the sloshing mode . again , this is easy to see . consider the average kinetic and potential energies , @xmath148 and @xmath149 . in equilibrium , both are equal ( virial theorem ) . now let us assume that the system is compressed or expanded , such that @xmath150 . using again the equations of motion ( [ eq : eqmotionharmonic ] ) , one obtains @xmath151 obviously , these two equations describe an undamped oscillation with frequency @xmath147 . let us now look if they stay valid in the presence of collisions . since the collisions do not change the positions of the particles and conserve the total kinetic energy , it is clear that @xmath152 and @xmath153 are not affected . now let us write the difference of @xmath154 before and after a collision of two test particles @xmath36 and @xmath155 : @xmath156 where @xmath157 and @xmath158 are the relative momenta ( e.g. , @xmath159 ) before and after the collision . in the original collision term as written in eq . ( [ eq : collisionterm ] ) , particles have to be at the same position to collide , i.e. , @xmath160 , such that @xmath154 is not changed . in our simulation this is somewhat different , since the test particles can collide at a distance of up to @xmath161 . this adds a small noise to @xmath154 . in all practical cases , however , this noise is completely negligible . as an example we show in the lower panel of fig . [ fig : sloshingbreathing ] the oscillation of the mean - square radius of the cloud as a function of time . as one can see , it is a perfectly undamped harmonic oscillation with frequency @xmath147 . for the theoretical investigation of collective modes , it is convenient to consider a system which is in equilibrium until it is excited by a short pulse at @xmath139 . formally , this is achieved by adding to the time - independent trap potential a perturbation term of the form @xmath162 the reason for this choice , which is of course different from the experimental way of exciting a collective mode , is the following : provided the perturbation @xmath163 is small enough ( such that the system reacts linearly to it ) , the response to a perturbation with arbitrary time dependence , @xmath164 , can easily be obtained by folding the result for the perturbation ( [ eq : deltapulse ] ) with the function @xmath165 . by integrating the boltzmann equation over the ( infinitesimal ) duration of the pulse , one can show that the effect of the perturbation ( [ eq : deltapulse ] ) is to change the distribution function as @xmath166 where @xmath167 and @xmath168 denote the limits @xmath169 from above and below , respectively . in the numerical simulation , this means that all test particles get a kick at @xmath139 , @xmath170 whereas their positions are not changed by the perturbation . from now on we will study the quadrupole mode as an example for a collective mode with non - trivial properties . we write the perturbation as @xmath171 corresponding to a kick at @xmath139 of @xmath172 and @xmath173 . the parameter @xmath174 determines the amplitude of the perturbation . if @xmath174 is chosen too small , it is difficult to separate the oscillation of the mode from fluctuations ; if it is chosen too large , one is not in the linear - response regime . all the following results were obtained with @xmath175 , corresponding to moderate amplitudes . by varying @xmath174 within reasonable limits , we checked that the amplitude of the resulting oscillation scales linearly with @xmath174 . ) with @xmath175 for different temperatures . the system has @xmath176 atoms and @xmath177 . lower panel : imaginary part of the corresponding fourier transforms.,title="fig:",width=302 ] + ) with @xmath175 for different temperatures . the system has @xmath176 atoms and @xmath177 . lower panel : imaginary part of the corresponding fourier transforms.,title="fig:",width=302 ] after the excitation of the radial quadrupole mode , we can look at the time evolution of the quadrupole moment @xmath178 as a function of time . results for different temperatures are displayed in fig . [ fig : respmode ] . contrary to the sloshing and breathing modes , the quadrupole mode is damped and the system approaches equilibrium ( @xmath179 ) after a certain time . at high temperatures ( @xmath180 ) , the system gets so dilute that it is in the collisionless regime ( @xmath181 , @xmath182 being the mean time between collisions of one atom ) . in this case , it takes many oscillations before the system returns to equilibrium . for lower temperatures , the mode is damped because of the high collision rate ( @xmath183 ) , but the system is not yet in the hydrodynamic regime ( @xmath184 ) where the mode would become undamped again . for the analysis of the results , it is useful to apply a fourier transform @xmath185 the so - called response function is the imaginary part of @xmath186 and can easily be obtained from the numerical results for @xmath187 by using a fast fourier transform ( fft ) algorithm @xcite . as an example , the fourier transforms of the results discussed above are shown in the lower panel of fig . [ fig : respmode ] . from the fourier transform one can clearly see that the spectrum of the mode in the collisionless regime , i.e. , at high temperature , has a sharp maximum at @xmath188 , as it should be in an ideal fermi gas , whereas at lower temperature the spectrum is broadened and the centroid of the spectrum is shifted to lower frequencies . this can be understood since at lower temperature the system is closer to the hydrodynamic regime , where the frequency should be @xmath189 . of course , one would like to give numbers @xmath190 and @xmath191 corresponding to the frequency and damping rate of the quadrupole mode in order to quantify these effects . the simplest way to obtain such numbers would be to fit the response function @xmath187 with a damped oscillation of the form @xmath192 . however , in the case of strong damping , this ansatz fits very badly the numerical results for @xmath187 . this can be understood by looking at the fourier transforms : the fourier transform of this ansatz function is a lorentzian , which has a line shape quite different from that obtained in our numerical simulation for @xmath193 or @xmath194 , cf . lower panel of fig . [ fig : respmode ] . hence , in order to analyze our numerical results , we need some physically motivated ansatz for the fit . in most of the theoretical work on collective modes in normal - fluid fermi gases , the boltzmann equation was not solved numerically , but approximate analytical solutions were found with the help of the method of moments @xcite . for a detailed description of the method , see e.g. ref . @xcite . applying the method of moments to the case of a perturbation of the form ( [ eq : deltapulse ] ) with @xmath195 according to eq . ( [ eq : potexcquad ] ) , one obtains a theoretical prediction for the response function @xmath196 . a brief description of the derivation is given in appendix [ app : mmresp ] , the final result reads @xmath197 where @xmath198 is the mean energy per atom in equilibrium , and @xmath3 is the relaxation time as defined in refs . @xcite , and depends on the cross section ( i.e. , the interaction strength ) , and the equilibrium distributions , cf . ( [ eq : deftau ] ) . one can see from eq . ( [ eq : imqmoment ] ) that in the collisionless and hydrodynamic limits the quadrupole mode has the frequencies @xmath199 and @xmath200 , respectively . the shape of the response function is completely determined by a single parameter , @xmath3 . by looking for the poles of eq . ( [ eq : imqmoment ] ) , one can calculate the inverse fourier transform which gives @xmath187 . the result has the form @xmath201 i.e. , it is a superposition of a damped oscillation with frequency @xmath190 and damping @xmath191 , and a non - oscillating , exponentially decaying term . the explicit expressions for @xmath202 , @xmath191 , and @xmath190 as functions of @xmath3 as well as for the amplitudes @xmath203 and @xmath204 are given in appendix [ app : qmom ] . we will refer to @xmath190 and @xmath191 as the frequency and damping rate of the quadrupole mode . note that in experiments determining these quantities , the data are usually fitted with a function that is similar to eq . ( [ eq : qmoment_t ] ) @xcite . in fig . [ fig : fittf ] we compare the response function obtained from the ) where the parameter @xmath3 is obtained by the method of moments ( dots ) , or by a fit to the simulation ( long dashes ) . the short dashes represent the response obtained with the extended method of moments including fourth - order moments . the system is a gas of @xmath176 particles at @xmath205 and @xmath206.,width=302 ] numerical simulation ( solid line ) with the result obtained from the method of moments , eq . ( [ eq : imqmoment ] ) ( dotted line ) . as one can see , the height of the peak and its general shape are in good agreement , but the position of the maximum is at different frequencies . however , if we try to fit the numerical result with a function of the form of eq . ( [ eq : imqmoment ] ) , using @xmath3 as fitting parameter , we can very well reproduce the numerical response function ( long - dashed line ) . it is remarkable that by adjusting only one parameter , @xmath3 , one can simultaneously reproduce the position , the height and the width of the peak , and also the shape far away from the maximum . however , surprisingly , the fitted value of @xmath3 is larger by approximately @xmath207 than the one obtained by the method of moments . as a consequence , the frequency @xmath190 and damping rate @xmath191 obtained from the fit of the response function deviate significantly from those predicted by the method of moments . these results are summarized in table [ tab : omegamma ] . .[tab : omegamma ] relaxation time , frequency , and damping of the quadrupole mode as obtained from the method of moments and fitting the results of the numerical simulation with a function of the form ( [ eq : imqmoment ] ) , corresponding to the dotted and dashed curves in fig . [ fig : fittf ] . [ cols="^,^,^,^",options="header " , ] in the existing literature @xcite , the method of moments was limited to second - order moments , as described in appendix [ app : mmresp ] . however , as we have seen above , this implies that the system is characterized by a single relaxation time @xmath3 , whereas in the spirit of a local - density approximation one would expect that in a trapped system the relaxation time should be position - dependent , @xmath208 . for instance , one could imagine that the gas in the center of the trap is more or less hydrodynamic ( short relaxation time ) , whereas far away from the trap center it gets very dilute and hence collisionless ( long relaxation time ) . in the case of the quadrupole mode , this means that the fermi - surface deformation is stronger at larger radii than in the trap center . it seems therefore natural to include into the ansatz for the perturbed distribution function in addition to the standard term @xmath209 describing the fermi - surface deformation , a term @xmath210 . more generally speaking , we should go beyond the standard approximation to include only second - order moments , and include also fourth - order ( or perhaps even higher ) moments . the task of extending the method of moments to the next higher order is in principle straight - forward but in practice very tedious : in the case of the quadrupole mode , the number of moments is increased from three to twelve . some details are given in appendix [ app : fourthorder ] . the resulting response function is shown in fig . [ fig : fittf ] as the short - dashed line . surprisingly , its shape is still similar , but now the position of the maximum agrees rather well with the result of the numerical simulation ( solid line ) . the agreement is even better at higher temperature ( see fig . [ fig : compfourth0.7 ] ) . for a temperature @xmath211 . for clarity , the fit of the simulation is not shown . the system is a gas of @xmath176 particles at @xmath206.,width=302 ] this nicely confirms the correctness of our numerical simulation and shows explicitly that the method of moments , if truncated at the lowest order , is insufficient . by doing calculations for various interaction strengths and temperatures , we found that the relaxation time from the simulation is systematically longer than that from the method of moments ( without fourth - order moments ) , eq . ( [ eq : deftau ] ) . results for weaker and stronger interactions ( @xmath212 and @xmath124 ) are displayed in fig . [ fig : tauinv ] . , as obtained from the simulation ( crosses ) and from the method of moments eq . ( [ eq : deftau ] ) ( solid lines ) , as a function of temperature . the system consists of @xmath213 atoms with @xmath132 ( upper panel ) and @xmath124 ( lower panel).,title="fig:",width=264 ] , as obtained from the simulation ( crosses ) and from the method of moments eq . ( [ eq : deftau ] ) ( solid lines ) , as a function of temperature . the system consists of @xmath213 atoms with @xmath132 ( upper panel ) and @xmath124 ( lower panel).,title="fig:",width=264 ] we see that the general behavior of @xmath3 as a function of temperature is the same within the simulation and the method of moments , but quantitatively there is a discrepancy of the order of @xmath207 in the whole range of temperatures where our numerical simulation is very accurate ( @xmath214 , cf . [ fig : collrate ] showing the temperature dependence of the collision rate ) . note that at lower temperatures , the determination of the pauli - blocking factors in the simulation of the collisions is not completely accurate , as discussed below fig . [ fig : collrate ] , such that the collision rate below @xmath215 is slightly too high . nevertheless the inverse relaxation time is too small . from this one can conclude that if we could improve the pauli blocking in the simulation , the discrepancy between the simulation and the method of moments ( without fourth - order moments ) would be even worse . the fourth order is thus important for the determination of the relaxation of the system and particularly for the frequency and the damping of collective modes . as we have just seen , the numerical simulation gives systematically a longer relaxation time @xmath3 than the method of moments . as the frequency @xmath190 and damping rate @xmath191 of the quadrupole mode are parametrized in terms of @xmath3 ( see appendix sec . [ app : qmom ] ) , one can ask the question how strongly this difference in @xmath3 will affect the results for @xmath190 and @xmath191 . since we are mainly interested in the intermediate regime @xmath216 between the hydrodynamic and collisionless limits , a difference of @xmath207 in @xmath3 can completely change the temperature dependence of @xmath190 and @xmath191 . this is shown in fig . [ fig : quauni ] , where the crosses are the results of the simulation may be wrong by 15% . the system consists of @xmath176 particles close to unitarity ( @xmath217).,title="fig:",width=264 ] of the simulation may be wrong by 15% . the system consists of @xmath176 particles close to unitarity ( @xmath217).,title="fig:",width=264 ] obtained from the simulation , whereas the solid lines are the results from the method of moments . one can clearly see that the numerical results stay close to the collisionless limit to much lower temperatures than the results obtained by the method of moments . to estimate the resulting precision of our numerical result on @xmath190 and @xmath191 , we show in fig . [ fig : quauni ] the error bands ( short - dashed lines ) which we obtain if we assume that our simulation may give a @xmath3 which is wrong by at most @xmath218 . this error includes numerical uncertainties which can be estimated from the scattering of the points in fig . [ fig : tauinv ] and the systematic deviation of the collision rate shown in fig . [ fig : collrate ] . if we include the fourth order moments , a global relaxation time @xmath3 does not exist anymore but we could define an effective one by fitting the response function as we did for the simulation . this effective relaxation time agrees very well with the one of the simulation so that both results give very similar frequency and damping . however , from a theoretical point of view , the definition of these quantities should come from the zeroes of the determinant of the matrix @xmath219 defined in appendix [ app : fourthorder ] . such a discussion is postponed to a forthcoming publication @xcite . in this paper , we presented a test - particle method for solving numerically the boltzmann equation for trapped fermi gases . while such methods have been popular in other fields of physics for many years , there have been only a few applications to ultracold atomic gases @xcite . our method is similar to that of refs . @xcite with some differences in the treatment of the collision term . in order to compute the occupation numbers in the pauli - blocking factors in the collision term , we represent each test particle by a gaussian in @xmath98 and @xmath63 space . the minimum value of the width of the gaussian is dictated by the statistical fluctuations due to the finite number of test particles , limiting the applicability of the method to temperatures above @xmath220 . as a first application of the method we discussed some collective modes . for simplicity , we considered only toy systems consisting of @xmath221 atoms in a spherical harmonic trap and neglected the mean - field potential and medium modifications of the cross section . as expected , the sloshing and monopole modes are undamped and independent of the collisions . in contrast , the quadrupole mode is very sensitive to collisions . in the hydrodynamic limit , its frequency should approach @xmath222 , while it is @xmath147 in the collisionless limit . in our simulations , we never reach the hydrodynamic regime , but the collisionless regime can be realized at high temperature due to the diluteness of the gas . surprisingly , the frequency and damping rate of the quadrupole mode obtained within the numerical simulation are quite different from those obtained within the widely used method of moments including moments up to second order in @xmath98 and @xmath63 . the method of moments predicts a relaxation time @xmath3 which is significantly shorter than the one obtained within the simulation . the reason is that the @xmath98 dependence of the relaxation time is neglected if only the @xmath223 moment is taken into account for the description of the fermi - surface deformation . we have shown that if the method of moments is extended to moments up to fourth order in @xmath98 and @xmath63 , e.g. , the @xmath224 moment , the agreement with the simulation becomes very good . the focus of the present paper was mainly to explain the test - particle method and to show its usefulness . for instance , the deficiency of the method of moments up to second order would not have been detected without the comparison with the numerical result . in future studies , we plan to apply the method to more realistic cases . in particular , in order to reach the typical numbers of atoms in the experiments , we will have to increase @xmath14 by a factor of @xmath225 . however , this should not pose a big problem : according to eq . ( [ conditiont ] ) , if we increase @xmath14 but keep the ratio @xmath226 fixed , the widths @xmath38 and @xmath39 may be chosen larger ( @xmath227 ) . this means that eq . ( [ conditionn ] ) stays satisfied with the same number of test particles , @xmath30 , i.e. , with a reduced ratio @xmath228 . the computation time will only grow because of the increased collision rate ( due to the larger test - particle cross section @xmath229 ) . another point is the trap geometry . the traps in the experiments are usually not spherical , but elongated . concerning the propagation of the test particles , this does not cause any difficulty , but in the calculation of the occupation numbers , it will probably be necessary to replace the width @xmath38 of the gaussian in @xmath98 space by different widths @xmath230 , @xmath231 , and @xmath232 in the three space directions . another important advantage of the numerical method is that an anharmonicity of the trap potential , which is always present in real experiments , can easily be included . finally , the mean field @xcite and medium modifications of the cross section @xcite should be included . the mean field , which originally depends on the chemical potential @xmath96 and the temperature @xmath97 , can be expressed as a function of the local density and energy density , which are both obtainable in the simulation . however , as shown previously @xcite , the mean field is not just proportional to the density : this leads to a huge numerical effort which is beyond the scope of this paper . the in - medium cross section is also difficult to be included because it depends on too many variables to be tabulated : @xmath233 . one possible solution of this problem is to replace the full @xmath234 and @xmath235 dependence of the in - medium cross section by a simple parametrization which results in the same local relaxation time @xmath236 . work in this direction is already in progress . in refs . @xcite , within the method of moments up to second order , the use of the in - medium cross section spoiled the agreement with experimental data because the resulting relaxation times were too short . since the present work shows that the numerical simulation gives a longer relaxation time than the method of moments , we hope that this problem can be solved . further important extensions of the present work are the generalization to polarized fermi gases and to superfluid systems . these questions , however , require more fundamental theoretical studies before they can be tackled numerically . replacing the distribution functions in eq . ( [ eq : collrate ] ) by equilibrium distribution functions @xmath237 , one obtains after some algebra the equilibrium collision rate @xmath238 where @xmath239 , @xmath240 , @xmath241 and @xmath242 . the total rate of allowed and blocked collisions is , in turn , given by eq . ( [ eq : collrate ] ) but without the factor @xmath243 in the integrand , leading to @xmath244 if the trap potential is spherically symmetric or harmonic , the spatial integrals can be reduced to one - dimensional ones . the remaining three - dimensional integrals are evaluated numerically with a monte - carlo algorithm . in the case of a weak perturbation , we can write the deviation of the distribution function from the equilibrium one in the form @xmath245 inserting this expression into the boltzmann equation and keeping only terms linear in the perturbation , one obtains ( see eq . ( 36 ) of ref . @xcite , for the case without mean field but with an external perturbation ) : @xmath246 . \label{eq : boltzlin}\end{gathered}\ ] ] here , @xmath247 $ ] is the linearized collision term as defined in eq . ( 37 ) of ref . @xcite ( up to a factor @xmath248 since here we are using a different normalization of @xmath2 ) : @xmath249 = \int \frac{d^3p_1}{(2 \pi ) ^3}\int d\omega \frac{d\sigma}{d\omega } \|{\bm{\mathrm{v}}}-{\bm{\mathrm{v_1}}}\| f_{\mathit{eq}}f_{{\mathit{eq}}\,1}\\ \times ( 1-f_{\mathit{eq}}^\prime)(1-f_{{\mathit{eq}}\,1}^\prime ) ( \phi+\phi_1-\phi^\prime-\phi_1^\prime)\ , . \label{collisionterm}\end{gathered}\ ] ] the perturbation @xmath250 is given by eqs . ( [ eq : deltapulse ] ) and ( [ eq : potexcquad ] ) . the usual approximation consists in making the ansatz @xmath251 with time - dependent coefficients @xmath252 and @xmath253 , @xmath254 , and @xmath255 , i.e. , only quadratic moments are considered . evaluating the moments @xmath256 , one obtains a system of equations for the fourier transformed coefficients : @xmath257 with @xmath258 \bigg\rbrack\end{gathered}\ ] ] and @xmath259 where @xmath260 are the poisson brackets . using the virial theorem , we obtain explicitly : @xmath261 where the relaxation time @xmath3 is defined by @xcite @xmath262\ , . \label{eq : deftau}\ ] ] solving this system of equations , we find @xmath263 and similar expressions for @xmath264 and @xmath265 . however , only the coefficient @xmath266 contributes to @xmath267 : with eq . ( [ eq : f1phi ] ) , and using again the virial theorem , we obtain @xmath268 , or explicitly : @xmath269 taking the imaginary part , we obtain eq . ( [ eq : imqmoment ] ) . in order to compute the fourier transform of eq . ( [ eq : qomega ] ) , let us start by factorizing the denominator : @xmath270 the expressions for the roots @xmath271 can be given in closed form . defining @xmath272 and @xmath273 we can write the roots @xmath271 as @xmath274 with @xmath275 now it is straight - forward to evaluate the inverse fourier transform of the response ( [ eq : imqmoment ] ) using the residue theorem . the result is given by eq . ( [ eq : qmoment_t ] ) with @xmath276 taking fourth order moments into account , we extend the previous ansatz eq . ( [ eq : ansatzmom2 ] ) as follows : @xmath277 which can be written as @xmath278 with , for example , @xmath279 . following the same steps as explained in appendix [ app : mmresp ] , we obtain now a system of twelve equations . the matrix @xmath219 can be computed explicitly . contrary to the second order calculations , the virial theorem can no longer be used to reduce the number of unknown quantities so that the system now depends on @xmath280 , @xmath281 and @xmath282 . in the matrix elements of the collision term , more parameters appear , generalizing the single parameter @xmath3 of the second order method . * j. kinast , s.l . hemmer , m.e . gehm , a. turlapov , and j.e . thomas , phys . lett . * 92 * , 150402 ( 2004 ) . j. kinast , a. turlapov , and j.e . thomas , phys . a * 70 * , 051401(r ) ( 2004 ) . m. bartenstein , a. altmeyer , s. riedl , s. jochim , c. chin , j.h . denschlag , and r. grimm , phys . lett . * 92 * , 203201 ( 2004 ) . a. altmeyer , s. riedl , m.j . wright , c. kohstall , j.h . denschlag , and r. grimm , phys . a * 76 * , 033610 ( 2007 ) . a. altmeyer , s. riedl , c. kohstall , m.j . wright , r. geursen , m. bartenstein , c. chin , j.h . denschlag , and r. grimm , phys . lett . * 98 * , 040401 ( 2007 ) . s. riedl , e r. sanchez guajardo , c. kohstall , a. altmeyer , m.j . wright , j.h . denschlag , r. grimm , g m. bruun , and h. smith , phys . a * 78 * , 053609 ( 2008 ) . s. nascimbne , n. navon , k.j . jiang , l. tarruell , m. teichmann , j. mckeever , f. chevy , and c. salomon , phys . lett . * 103 * , 170402 ( 2009 ) . m. zwierlein , talk given at the conference bec 2009 , sant feliu de guixols ( spain ) , sept . 5 - 11 , 2009 . c. menotti , p. pedri , and s. stringari , phys . rev . lett . * 89 * , 250402 ( 2002 ) . m. cozzini and s. stringari , phys . lett . * 91 * , 070401 ( 2003 ) . p. pedri , d. gury - odelin , and s. stringari , phys . a * 68 * 043608 ( 2003 ) . e. taylor and a. griffin , phys . a * 72 * , 053630 ( 2005 ) . m. urban and p. schuck , phys . a * 73 * , 013621 ( 2006 ) . m. urban , phys . a * 75 * , 053607 ( 2007 ) , phys . a * 78 * , 053619 ( 2008 ) . f. toschi , p. vignolo , s. succi , and m.p . tosi , phys . a * 67 * , 041605(r ) ( 2003 ) . f. toschi , p. capuzzi , s. succi , p. vignolo , and m.p . tosi , j. phys . b * 37 * , s91 ( 2004 ) . p. massignan , g.m . bruun , and h. smith , phys . a * 71 * , 033607 ( 2005 ) bruun and h. smith , phys . rev . a * 76 * , 045602 ( 2007 ) s. chiacchiera , t. lepers , d. davesne and m. urban , phys . a * 79 * , 033613 ( 2009 ) . bertsch and s. das gupta , phys . rep . * 160 * , 189 ( 1988 ) . b. jackson and e. zaremba , phys . a * 66 * , 033606 ( 2002 ) . e.m . lifshitz and l.p . pitaevskii , _ physical kinetics _ , l.d . landau and e.m . lifshitz course of theoretical physics vol . 10 ( pergamon , oxford , 1980 ) . landau and e.m . lifshitz , _ quantum mechanics _ , course of theoretical physics vol . 3 ( pergamon , london , 1958 ) . w.c . swope , h.c . andersen , p.h . berens , and k.r . wilson , j. chem . phys . * 76 * , 637 ( 1982 ) . l. verlet , phys . rev . * 159 * , 98 ( 1967 ) . w.h . press , teukolsky , w.t . vetterling , and b.p . flannery , _ numerical recipes in fortran : the art of scientific computing _ , 2nd edition ( cambridge university press , 1992 ) . w. kohn , phys . rev . * 123 * , 1242 ( 1961 ) . l. brey , n.f . johnson , and b.i . halperin , phys . b * 40 * , 10647 ( 1989 ) . thomas lepers , institut de physique nuclaire , universit de lyon , phd thesis , to be published . t. lepers , s. chiacchiera , m. urban , and d. davesne in preparation .	we numerically solve the boltzmann equation for trapped fermions in the normal phase using the test - particle method . after discussing a couple of tests in order to estimate the reliability of the method , we apply it to the description of collective modes in a spherical harmonic trap . the numerical results are compared with those obtained previously by taking moments of the boltzmann equation . we find that the general shape of the response function is very similar in both methods , but the relaxation time obtained from the simulation is significantly longer than that predicted by the method of moments . it is shown that the result of the method of moments can be corrected by including fourth - order moments in addition to the usual second - order ones and that this method agrees very well with our numerical simulations .
a basic theorem in riemannian geometry inspires our work . recall two loops in a space @xmath0 are _ freely homotopic _ if one loop can be deformed into the other without leaving @xmath0 . the resulting equivalence classes of loops are the _ free homotopy classes_. this basic theorem asserts if @xmath0 is a compact riemannian manifold then every one of its free homotopy classes of loops is realized by a periodic geodesic . this theorem suggests an analogue for the planar newtonian three - body problem . replace the riemannian manifold above by the configuration space @xmath0 of the planar three - body problem : the product of @xmath1 copies of the plane , minus collisions . is every free homotopy class of this @xmath0 realized by a solution to the planar three - body problem ? by a _ reduced free homotopy class _ we mean a free homotopy class of loops for the quotient space @xmath2 of the configuration space @xmath0 by the group @xmath3 of rotations acting on @xmath0 by rigidly rotating the triangle formed by the three bodies . by a _ reduced periodic solution we mean a solution which is periodic modulo rotations , or , what is the same thing , a solution which is periodic in some rotating frame . [ th_1 ] consider the planar three - body problem with fixed negative energy and either equal or near equal masses . then , for all sufficiently small _ nonzero _ angular momenta , every reduced free homotopy class is realized by a reduced periodic solution . the case of zero or large angular momentum remains open . for further motivation and history regarding this problem please see @xcite . * syzygy sequences . * the reduced free homotopy class of a periodic motion of the three bodies can be read off of its _ syzygy sequence_. a ` syzygy ' is a collinear configuration occurring as the three bodies move . ( the word comes from astronomy . ) syzygies come in three types , type 1,2 , and 3 , depending on the label of the mass in the middle at collinearity . a generic curve in @xmath0 has a discrete set of syzygies . list its syzygyies in temporal order to obtain the syzygy sequence of that curve . if two or more of the same letters occur in a row , for example 11 , cancel them in pairs using a homotopy . we call such multiple sequential occurrences of the same letter a stutter " . continue canceling stutter pairs until there are no more . for example @xmath4 . we call the final result of this cancellation process the _ reduced syzygy sequence _ of the curve . periodic reduced syzygy sequences are in a 1:2 correspondence with reduced free homotopy classes . theorem 1 implies that all reduced syzygy sequences are realized by reduced periodic solutions . * shape space and topology . * after fixing the center of mass , the configuration space @xmath0 of the planar three - body problem is diffeomorphic to the product @xmath5 where the first factor represents the size of the triangle , the second an overall rotation and the third the shape of the triangle . the rotation group @xmath3 generates the circle factor @xmath6 and can be identified with it . we will forget the circle of rotations , and focus on homotopy classes and solutions _ modulo _ rotation . up to homotopy equivalence , the size factor is also irrelevant . the remaining @xmath7 is the _ shape sphere _ minus its three binary collision points . see figure [ fig_shapesphere1 ] . points of the shape sphere represent oriented similarity classes of triangles . there is a canonical quotient map @xmath8 which sends a configuration to its _ shape_. the equator of the shape sphere represents collinear configurations : all three masses in a line . the three deleted collision points , denoted @xmath9 lie on this equator and split it into three arcs , which could be labelled 1 , 2 , and 3 according to the syzygy type . we redo the construction of the syzygy sequence of a curve using shape language . take a closed curve in @xmath0 . project it to the shape sphere , perturbing it if necessary so that it crosses the equator transversally . make a list of the arcs encountered in temporal order . this list is the curve s syzygy sequence . refer to figure [ fig_stutter ] for the picture in the shape sphere of cancelling stutter pairs . we cancel stutter pairs until there are no more , arriving at the reduced syzygy sequence of the loop . the reduced syzygy sequence is a free homotopy invariant : two loops , freely homotopic in @xmath2 , have the same reduced syzygy sequence . this map from reduced free homotopy classes to reduced syzygy sequences is two - to - one , with a single exception . the two reduced free homotopy classes which represent a given nonempty reduced syzygy sequence are related by reflection , or , what is the same thing , by the choice of direction north - to - south " or south - to - north " with which we cross the equator when we list the first letter of our syzygy sequence . the single exception to this two - to - one rule is the empty reduced syzygy sequence having no letters . this empty class represents only the trivial reduced free homotopy class whose representatives are the contractible loops in @xmath2 . theorem [ th_1 ] is an immediate corollary of the following theorem . to state it , define a stutter block of size @xmath10 to be a syzygy sequence of the form @xmath11 where @xmath12 . the theorem will guarantee that we can realize any syzygy sequence which is a concatenation of such stutter blocks with sizes in a certain range @xmath13 . let s call such a syzygy sequence a _ stutter sequence with range @xmath14$]_. note that one could put several blocks with the same symbol together so while @xmath15 is a lower bound on the number of repetitions of a symbol , @xmath16 is not an upper bound . [ th_2 ] consider the planar three - body problem at fixed negative energy . there is a positive integer @xmath15 such that given any @xmath17 , all bi - infinite stutter sequences with range @xmath14 $ ] are realized by a collision - free solution , provided the masses are sufficiently close to equal and the angular momentum @xmath18 satisfies @xmath19 . if the syzygy sequence is periodic then it is realized by at least one reduced periodic solution . between each stutter block the realizing solutions pass near triple collision . their projections to the reduced configuration space lie near the black curves of figure [ fig_skeleton ] . proof of theorem 1 from theorem 2 . take @xmath20 in the theorem [ th_2 ] so that we are guaranteed that there is an odd integer @xmath21 in the interval @xmath22 $ ] . given a nontrivial reduced periodic syzygy sequence , replace each occurrence of letter @xmath23 by @xmath24 to get a corresponding periodic unreduced syzygy sequence satisyfing the conditions of theorem 2 . since @xmath21 is odd , this sequence reduced to the given reduced periodic sequence . the periodic solution guaranteed by theorem [ th_2 ] then realizes one of the two free homotopy classes having the given reduced syzygy sequence . to realize the other class apply a reflection to this solution . finally , to realize the empty sequence we can simply take the lagrange solution . or we could take all the exponents @xmath25 even : for example all @xmath26 s , or repeated @xmath27 . qed .3 cm * a history of failed variational attempts . * the proof of the realization of all free homotopy classes in the riemannian case uses the direct method of the calculus of variations . fix a free homotopy class . minimize the length over all loops realizing that class . since the manifold is assumed compact , the minimum is realized and is a geodesic . in mechanics the extremals of the _ action _ yield solutions to newton s equations . it is natural to try to proceed in the same way . fix a reduced syzygy sequence . minimize the action over all representatives of that class . the configuration space is noncompact which complicates the approach . there are two sources of noncompactness : spatial infinity @xmath28 and collision @xmath29 , where @xmath30 denotes the distance between any two of the three bodies . the first type is fairly easy to overcome . noncompactness due to the collision is the essential difficulty in applying the direct method . for the standard @xmath31-type potential of newtonian gravity , there are paths _ with collision _ which have finite action . through these paths we find , by explicit examples , that while trying to minimize the action we may leave the free homotopy class we started in , through a ` pinching off ' via collision , in which we leave @xmath0 and we lessen the action , entering a different free homotopy class . in other words : the infimum of the action over a fixed free homotopy class is typically not achieved , and minimizing sequences tend to collision . see @xcite for an assertion that this is probably a ubiquitous problem , suffered when we choose almost any free homotopy class . the fix to this collision problem , going back at least to poincar @xcite , is to cheat . change the newtonian potential to one which blows up like @xmath32 ( or even stronger ) as we approach collisions at @xmath33 . if the negative of the potential for the force is greater than @xmath34 whenever the distance @xmath30 between two of the bodies is sufficiently small , then the action of any path suffering a collision is infinite . such potentials are called strong force " . under the strong force assumption each free homotopy type is separated from every other by an infinite action wall " . the direct method works like a charm . poincare @xcite proved that almost every nontrivial homology class in @xmath0 is represented by a periodic orbit . about a century later one of us proved @xcite that almost every free homotopy class modulo rotations , i.e. , that almost every reduced syzygy sequence , is realized by a periodic solution to the planar strong force three - body problem . ( the only ones that are not realized are those that only involve 2 letters , so : 1212 .. , 1313 ... and 2323 ... . these spiral in to binary collision between the two involved bodies . ) for years one of us had tried to realize all reduced free homotopy classes for the honest newtonian @xmath31 potential by variational methods . these efforts led to interesting discoveries @xcite , @xcite , @xcite but did not seem to bring us any closer to solving the original realization problem . ( for more on the various attempts and theorems along the way also see @xcite . ) so it was with great joy that we realized that ideas developed by the other author in the 1980 s @xcite , summarized and extended as theorem 2 , would prove of theorem 1 . * the main idea behind the proof of theorem [ th_2 ] . * the solutions described by theorem 2 arise out of chaotic perturbations of the homothetic solutions of euler and lagrange . we recall that the solutions of euler and lagrange are precisely those solutions to newton s equations whose shape does not change as they evolve . when projected onto the shape sphere such a solution curve is a single constant point . there are five points in all . they are called central configurations and denoted here @xmath35 . the @xmath36 are euler s collinear configurations . @xmath36 lies on the collinear arc marked @xmath23 . the @xmath37 are lagrange s central configurations and are equilateral triangles , with one for each possible orientation of a labelled equilateral triangle in the plane . each central configuration gives rise to a family of solutions , the family being parameterized by an auxiliary kepler problem , i.e. , by conics . for each choice of conic the corresponding solution consists of the three masses travelling along homographic ( = scaled , rotated ) versions of this conic , with the conics placed so that one focus is the center of mass of the three bodies . the masses travel their individual conics in such a way that the triangle they form remains in a constant shape . at fixed energy we can think of the parameter as the angular momentum . the homothetic solutions correspond to angular momentum zero and are degenerate conics . for these solutions the three bodies move along collinear rays , exploding out of triple collision until they reach a maximum size ( dependent on the energy ) at which instant they are all instantaneously at rest . from that instant they reverse their paths , shrinking back homothetically to triple collision . when the angular momentum is turned on to be small but near zero , the masses move along highly eccentric nearly collinear ellipses . see figure [ fig_eccentriccentral ] . the solutions of theorem 2 feature many close approaches to triple collision . while they are away from collision they look much like the eulerian homothetic solutions the bodies are close to one of the three eulerian central configurations @xmath36 while the size expands and then contracts again to another close approach to triple collision . near collision , however , their behavior is quite different . the shape begins to oscillate around the collinear shape , producing a large syzygy block @xmath38 . very close to collision the shape approaches one of the two lagrange equilateral triangles the transitions from @xmath36 to @xmath40 or @xmath41 are through isosceles shapes ( the circular arcs in figure [ fig_skeleton ] ) . one can specify a transition to either @xmath40 or to @xmath41 at each close approach . after the shape is nearly equilateral , the triangle spins around ( similar to the behavior near collision of the lagragian homographic solution in figure [ fig_eccentriccentral ] ) . next , the shape makes a transition back to one of the eulerian central configuration @xmath42 , following one of the three isosceles circles . arriving near @xmath42 the shape oscillates and the size starts to increase again . away from collision , the behavior is like the euler homothetic solution with shape @xmath42 . the main point is that , during each approach to collision , we can choose which lagrange configuration to approach and then which eulerian homothetic orbit to mimic next . in this way we can concatenate the three types of syzygy blocks at will . note that the arrangement of the three isosceles half - circles in figure [ fig_skeleton ] has the same free homotopy type as the sphere minus three points . by following the dark lines of figure [ fig_skeleton ] we can realize every reduced free homotopy class . the results of @xcite allow us to make the necessary transitions near the poles from one kind of eulerian behavior to the other . what was not explicitly said there , but follows from the proof , is that when the masses are nearly equal these realizing solutions , when projected to the shape sphere , stay within a small neighborhood of the isosceles half - circles . this prevents the occurrence of additional , unintended syzygies and also allows us to avoid binary collisions . there are a number of reasons we need some angular momentum to achieve the gluing of the three half - circles . without angular momentum , the isosceles problems are invariant sub - problems and so we can not switch from one circle to another . sundman s theorem asserts that triple collision is impossible when the angular momentum is not zero , hence some angular momentum relieves us of the inconvenience of investigating whether or not solutions die " in triple collision . but the most important reason is that we need the lagrange - like spinning behavior near triple collision to connect up the orbit segments approaching collision with those receding from it . see figure [ fig_loop ] where this spinning behavior appears as a restpoint connection at @xmath33 . this mechanism is described in more detail in the next section . in this section , we will describe how theorem [ th_2 ] follows from previous results on constructing chaotic invariant sets for the planar three - body problem with small , nonzero angular momentum @xcite . these results , in turn , are based on a study of the isosceles three - body problem , so we begin there . when two of the three masses are equal , there is an invariant subsystem of the planar three - body problem consisting of solutions for which the shape of the triangle formed by the three bodies remains isosceles for all time . all of these solutions have zero angular momentum . after fixing the center of mass , the isosceles subsystem has two degrees of freedom so fixing the energy gives a three - dimensional flow . in the early eighties , there were several papers published about this interesting problem @xcite . we will describe a few features relevant to syzygy problem , referring to @xcite for more details . in order to be definite , let us suppose that we are thinking of @xmath43 so that mass 3 forms the vertex of the isosceles triangles . the orbits of interest pass near triple collision . using mcgehee s blow - up method , we replace the triple collision singularity by an invariant manifold forming a two - dimensional boundary to each three - dimensional energy surface . the blow - up method involves introducing size and shape variables @xmath30 and @xmath44 and corresponding momenta @xmath45 and @xmath46 . the size @xmath30 of the triangle is the square root of the moment of inertia with respect to the center of mass . thus @xmath33 represents triple collision of the three masses at their center of mass . a change of timescale by a factor of @xmath47 slows down the orbits near collision and gives rise to a well - defined limiting flow at @xmath33 . the blown - up equations read @xmath48 together with the energy constraint @xmath49 here @xmath50 $ ] parameterizes the isosceles great circle in shape space for the two equal masses , and so describes the shape of the isosceles triangle . the boundary of the @xmath44-interval , @xmath51 , represent binary collision between the two equal masses . we refer to @xcite for details regarding @xmath52 . as @xmath44 runs over @xmath53 $ ] , the triangle opens up , becoming equilateral at a certain angle @xmath54 and collinear at @xmath55 , then passing through another equilateral shape at @xmath56 before collapsing to binary collision again . the equilateral and collinear shapes are the lagrangian ( @xmath39 ) and eulerian @xmath57 central configurations . the triple collision orbits of the three - body problem , isosceles or not , converge to a central configuration . for each central configuration , there are two restpoints on the associated full collision manifold , one associated to solutions converging to that limiting shrinking shape , the other exploding out of it . the isosceles collision manifold forms an invariant submanifold of the full planar collision manifold , discussed in the next subsection . for the isosceles subproblem we have a total of six rest points denoted @xmath58 , @xmath59 and @xmath60 and @xmath57 . ( the notation is different in @xcite . ) isosceles solutions having triple collision in forward time converge to one of the restpoints @xmath61 as the rescaled time tends to infinity . in other words , they form the stable manifolds of these three restpoints . similarly , orbits colliding in backward time ( ejection orbits ) are represented by the unstable manifolds of @xmath62 . in the figure , there are connecting orbits running from the ejection restpoints to the corresponding collision restpoints . these are the well - known homothetic solutions where the shape remains constantly equal to one of the central configurations while the size expands from @xmath33 to some maximum and then contracts to zero again . the six restpoints in the collision manifold are all hyperbolic and the dimensions of their stable and unstable manifolds are apparent from figure [ fig_isosceles ] . within the two - dimensional collision manifold @xmath33 , the lagragian restpoints are saddles with one - dimensional stable and unstable manifolds . in the three - dimensional energy manifold the starred restpoints pick up an extra stable dimension while the unstarred ones pick up an extra unstable dimension . thus @xmath63 and @xmath64 are two - dimensional surfaces . the lagrangian homothetic connecting orbits represent transverse intersections of these surfaces . within the collision manifold , @xmath65 is a repeller and @xmath57 is an attractor . viewed within the three - dimesional energy manifold , @xmath57 and @xmath59 pick up an extra stable and unstable dimension , respectively . clearly the @xmath66 connecting orbit , passing as it does through the interior of the isosceles phase space , is not a transverse intersection of stable and unstable manifolds . this connecting orbit represents the euler homothety orbit . a crucial fact for us is that the eulerian restpoints have nonreal eigenvalues within the collision manifold for a large open set of masses , including those near equal masses . this spiraling , together with the presence of restpoint connections in the collision manifold , causes the surfaces @xmath63 and @xmath64 to wrap like scrolls around the @xmath66 homothetic orbit as indicated in the figure . to get a simpler picture of these scrolls , we use a piece of the plane @xmath67 as a cross - section to the flow near the homothetic orbit . the origin of @xmath68 represents the euler homothetic orbit . the surfaces @xmath63 and @xmath64 intersect @xmath68 in spiraling curves . more precisely ( see @xcite ) , in polar coordinates @xmath69 near the origin , each such curve is parametrized by the polar angle @xmath70 by setting @xmath71 where @xmath72 is some strictly monotonic function : the curves wind monotonically around the origin . figure [ fig_spirals ] shows the cross - section @xmath68 together with a schematic drawing of the four spiraling curves . the curves representing @xmath63 spiral in opposite directions from those representing @xmath64 . the four curves are related by reflection symmetries . namely , let @xmath73 and @xmath74 . then @xmath75 and @xmath76 it follows that there must be infinitely many intersection of each of the unstable manifolds with each of the stable ones . since the manifolds are real analytic , these intersections are either transverse or , at worst , finite - order tangencies . even a finite - order crossing is topologically transverse " and this is enough for the topological construction given later on . we have found infinitely many topologically transverse intersections of lagrangian ejection and collision orbits . these orbits behave qualitatively as follows . beginning near one of the lagrange ejection restpoints @xmath39 they follow a connecting orbit in the collision manifold to a neighborhood of the eulerian ejection rest point @xmath57 . then they follow the eulerian homothetic orbit until they are close to the eulerian collision restpoint @xmath65 . finally , they follow another connecting orbit in the collision manifold to end at collision at one of the restpoints @xmath77 so far , this subsection has summarized previous work on the isosceles problem . for the syzygy problem , we will need a few new observations . the claim is that this type of orbit has a syzygy sequence of the form @xmath78 for some large value of @xmath10 and , moreover , any sufficiently large @xmath10 can be achieved in this way . to see this , first note that the connecting orbits @xmath79 and @xmath80 in the collision manifold go directly from the equilateral to collinear shapes , staying far from double collision and far from any syzygy other than type @xmath1 . to study syzygies of type @xmath1 recall that the collinear shape is represented in our @xmath81 coordinates by the plane @xmath82 , which in the plane @xmath68 of figure [ fig_spirals ] is represented by the @xmath46-axis . the plane @xmath83 contains the eulerian restpoints and eulerian homothetic orbit . due to the spiraling at the restpoints , it is clear that our orbits will have a large number of type @xmath1 syzygies . it is also clear that the number is odd for connections @xmath84 and @xmath85 . it remains to show than any sufficiently large number of syzygies can be achieved . one of the differential equations for the blown - up isosceles problem is @xmath86 . thus , except for the euler homothetic orbit , orbits cross the syzygy plane transversely . in figure [ fig_isosceles ] orbits in the front half of the energy manifold ( @xmath87 ) are crossing from right to left while those at the back ( @xmath88 ) cross from left to right . it follows that the number of syzygies which occur on any orbit segment is a continuous function of the endpoints of the segment , as long as the endpoints are not on the syzygy plane . now consider one of the two stable spirals in figure [ fig_spirals ] . each point of the spiral determines a forward orbit segment ending , say , in some convenient cross - section near the corresponding lagrange restpoint . it follows that as we vary the point on the spiral , the total number of syzygies experienced in forward time can change only when the initial point crosses the line @xmath55 . since the spiraling is monotonic , the plane @xmath55 divides the spiral into curve segments in each of which the total number of forward syzygies is constant . moreover , points near @xmath55 immediately cross it once or have just crossed it . hence for nearby points of the spiral on opposite sides of @xmath55 , the total number of syzygies differs by one . since the number of syzygies tends to infinity as the spiral converge to the center , every sufficiently large number of forward syzygies is attained . there is a similar story for counting backward syzygies of points on the unstable spirals . consider a segment @xmath89 of @xmath90 with endpoints in @xmath91 whose interior points have exactly @xmath10 forward time syzygies . then @xmath92 is a segment of @xmath93 whose interior points have @xmath10 syzygies in backward times . these spirals intersect at a point of the line @xmath94 ( the zero velocity curve ) . this point represents one of our ejection - collision orbits having exactly @xmath95 syzygies of type @xmath1 . similarly there are @xmath96 ejection collision orbits with exactly @xmath95 syzygies . on the other hand , for the same segment @xmath89 of @xmath90 the other reflection @xmath97 is a segment of @xmath98 . these intersect at their endpoints , that is , along @xmath91 . these intersection points represent @xmath99 ejection collision orbits . one endpoint or the other has exactly @xmath10 syzygies in both forward and backward time including the initial syzygy ( if the segment lies in @xmath100 it s the endpoint with @xmath88 ; otherwise the one with @xmath87 ) . for this endpoint , the total number of forward and back ward syzygies will be exactly @xmath101 . thus any sufficiently large number of syzygies can be achieved . consider the planar three - body problem with equal masses . for each fixed angular momentum @xmath18 , we get a rotation - reduced hamiltonian system of three degrees of freedom . when the energy is further fixed at the value @xmath102 we get a flow on a 5-manifold denoted @xmath103 . if the angular momentum is zero then three separate isosceles problems appear as subsystems . for each of these we can construct ejection - collision orbits as in the last subsection . so we can realize syzygy sequences of types @xmath104 and @xmath78 for all sufficiently large @xmath10 , say @xmath105 . in this subsection we will show how to concatenate these blocks to get bi - infinite syzygy sequences of the form @xmath106 where @xmath107 and @xmath108 . this requires perturbing to nonzero angular momentum and for this we need to choose an upper bound @xmath109 for @xmath110 as in the statement of theorem [ th_2 ] . in addition we will see that nonequal masses can be handled and that every periodic sequence of this form is realized by at least one periodic orbit . the rest of this section is a summary of the results of @xcite which we refer for more details . see also @xcite for some similar arguments . our aim is to convey the spirit of the proof and not all the details , and to point out where a bit of new work is needed to keep track of stutter block lengths . the mcgehee blow - up for the planar problem has a lot in common with the blow - up for the isosceles problem . the energy manifolds @xmath111 are five - dimensional and all share the four - dimensional collision manifold @xmath33 as a common boundary . there are five central configurations which lead to a total of ten restpoints in the collision manifolds , namely five ejection restpoints @xmath112 and the corresponding collision restpoints with starred notation . the stable and unstable manifolds @xmath90 and @xmath113 now have dimension three and the lagrange homothetic solutions are still transverse intersections of these . moreover , all of their infinitely many other topologically transverse intersections viewed within the isosceles submanifolds remain topologically transverse when viewed from the larger planar energy manifolds . there are two extra dimensions to be understood in the planar problem . the cross - section @xmath67 ( figure [ fig_spirals ] ) to the euler homothetic orbit is now four - dimensional and instead of spiraling curves we have spiraling two - dimensional surfaces . it is shown in @xcite that these surfaces intersect topologically transversely in @xmath68 , the realization of arbitrary concatenations of stutter blocks is accomplished by the methods of symbolic dynamics analogous to those used for the smale horseshoe map . ( see @xcite . ) one version of the horseshoe is built by taking two rectangles each of which is stretched in one direction and contracted in the other and then mapped across both of the original rectangles . label the two rectangles by symbols @xmath114 . one can prescribe an arbitrary bi - infinite sequence of @xmath115 s and @xmath116 s and prove that the sequence is realized by a unique orbit . the sequence itself is the _ itinerary _ of the orbit : the list of rectangles visited in order of visit . the uniqueness comes from uniform hyperbolic stretching . in the topological approach used here , we will not be able to guarantee uniqueness . the topological approach based on windows " was pioneered by easton @xcite and further developed and used in @xcite for the planar three - body problem . we follow the discussion and notation of @xcite , p.56 - 60 , closely . instead of the rectangles in the horseshoe map , we have four - dimensional boxes called windows homeomorphic to the product @xmath117 of two two - dimensional disks . the analogous splitting for a two - dimensional window is shown in figure [ fig_windowpic ] . in that figure , horizontal line segments of one window are being stretched horizontally across the next box . for our four - dimensional window @xmath118 , the first disk represents two directions which will be stretched in forward time and will be called positive . the second disk represents two directions which will be compressed in forward time , or alternatively , stretched in backward time . these directions will be called negative . there is a splitting of the boundary of @xmath117 into two parts @xmath119 and @xmath120 , each homeomorphic to a solid torus . a positive disk has its boundary in @xmath121 and represents a generator of the relative homology group @xmath122 . such a window can be constructed near each of the transverse intersections of the surfaces @xmath64 and @xmath63 in the four - dimensional cross - section @xmath68 . in carrying out the perturbation to nonzero angular momentum later on , we will only be able to work with finitely many windows . this is where the choice of the number @xmath109 in the statement of theorem [ th_2 ] comes in . we already have @xmath15 as lower bound on the number of syzygies which are realized by the ejection collision orbits . given any @xmath123 we can restrict attention to the finitely many topologically transverse ejection - collision orbits which realize stutter blocks size in the range @xmath124 . then our perturbation will be constructed based on this finite collection of windows . the windows are chosen so that the positive directions are aligned with the unstable manifold involved in the intersection while the negative directions are aligned with the stable manifold ( thus a positive disk crosses the stable manifold transversely in some sense and , when followed forward near the restpoint , will get stretched ; the positive boundary @xmath121 is linked with the stable manifold ) . for definiteness , consider a point of intersection of @xmath98 and @xmath125 which realizes the stutter block @xmath126 for some odd number @xmath10 . using topological transversality , one can choose a @xmath127 local coordinate system which makes these manifolds coordinate planes and then choose the window to be a product of small discs in these planes . see figure [ fig_saddlewindow ] . if the window is sufficiently small , the orbits starting there will behave qualitatively like the ejection - collision orbit at its center . they will enter a small neighborhood of @xmath128 in forward time and a small neighborhood of @xmath40 in backward time . in between they will avoid collisions , realize the stutter block @xmath78 , and stay close to the isosceles manifold for @xmath1 . setting up one such window near each of a finite number of intersection points whose corresponding solutions have @xmath10 stutters with @xmath129 $ ] , we need to show that when we perturb to small nonzero angular momentum , each of these windows is stretched across each of the others by the perturbed flow . a topological analogue of the uniform stretching in the smale horseshoe is provided by a homological definition of correct alignment " . the version used in @xcite differs from that in easton s work and will be described briefly now . we ca nt expect the poincar mappings of the three - body flow to carry one window exactly onto another . rather , as in the horseshoe , a window is stretched through the other and overlaps it . for this reason it is convenient to set up some larger auxiliary windows which capture the homology of the original as shown schematically in figure [ fig_windowpic ] . to each window @xmath46 we associate an auxiliary window @xmath130 with a thickened boundary set @xmath131 such that there is a retraction map @xmath132 fixing @xmath46 and inducing an isomorphism on relative homology groups @xmath133 . similarly , each window will have an associated auxiliary pair @xmath134 . then two windows @xmath135 are said to be _ correctly aligned _ if there is a flow - induced poincar map taking @xmath136 and inducing an isomorphism on the second relative homology groups . we also require the inverse poincar map to take @xmath137 and to induce an isomorphism on homology . so if two windows @xmath138 are correctly aligned there is a kind of homological stretching instead of the usual hyperbolic stretching . the next step is to show that given any bi - infinite sequence of correctly aligned windows , there is a nonempty compact set of orbits which passes through their interiors using the given poincar maps . to get the idea of the proof , recall the proof of the analogous statement for the smale horseshoe . consider any bi - infinite itinerary specifying which of the two rectangles the orbit should hit at each iteration . the set of points in the initial rectangle which map into the next one is a negative subrectangle of the initial rectangle . in fact any finite forward time itinerary is realized by such a subrectangle . similarly , any finite backward time itinerary is realized by a positive subrectangle . by intersecting these rectangles we get a nonempty compact set realizing any finite itinerary . since an intersection of nested , nonempty compact sets is nonempty and compact , we can also realize a bi - infinite itinerary . from the topological perspective , we replace the idea of negative subrectangle with compact set " which intersects every positive segment or better compact set which intersects the support of every chain which is nontrivial in @xmath139 " . returning to our four - dimension case , define a _ positive chain _ to be a relative singular two - chain whose homology class in @xmath140 is nonzero and similarly for _ negative chains_. using the definition of correct alignment , one can show that the set of points in the initial box which realize any given finite forward time itinerary is a compact set which intersects the support of any positive chain . similarly the points realizing a chosen finite backward itinerary form a compact set intersecting the support of every negative chain . using some algebraic topology one can show that any two such sets of this type have nonempty intersection and the proof for bi - infinite sequences is completed as for the horseshoe . the proof that a periodic sequence is realized by at least one reduced periodic orbit is similar . here we have a composition of poincar maps giving a map @xmath141 from a subset of the initial box to itself . consider @xmath142 , that is , the set of points whose first two coordinates ( out of four ) are fixed . similarly let @xmath143 . then one can show that @xmath144 intersects the support of every positive chain and @xmath145 intersects the support of every negative chain , and it follows as above that @xmath146 , that is , there is at least one fixed point . it remains to explain how we can get each of the windows defined near our ejection - collision orbits to be correctly aligned with itself and all of the other windows when the angular momentum is nonzero and sufficiently small . we know that each window approaches one of the hyperbolic collision restpoints @xmath77 in forward time and one of the ejection restpoints @xmath39 in backward time . we need to find a way to go from a neighborhood of @xmath77 to a neighborhood of @xmath39 . it turns out that there are transverse connecting orbits @xmath147 and @xmath148 contained in the boundary flow at @xmath33 . we have already discussed the four - dimensional collision manifold at @xmath33 which serves as a boundary to the @xmath149 , fixed energy manifolds . but there is another five - dimensional flow at @xmath33 representing the limits as @xmath150 of nonzero angular momentum orbits ( @xmath151 in figure [ fig_loop ] ) . in particular , consider the highly eccentric lagrange homographic solutions . the equilateral triangle formed by the three bodies expands and contracts to near zero . instead of colliding , the triangle quickly spins by @xmath152 and the triangle expands again . in the limits as @xmath150 these solutions converge to cycles of restpoint connections . we get the lagrange homothetic orbits @xmath153 and also connections going in the other direction @xmath154 . the latter represent the limit of the spinning behavior . these new connections are also transverse and we set up another window along each of them . for @xmath149 the windows near our ejection - collision orbits approach the lagrangian restpoints and are stretched out by the hyperbolic dynamics there . similarly the new windows constructed along the connecting orbits in @xmath155 approach these same restpoints and are also stretched out . but the four - dimensional collision manifold @xmath156 stands between them within @xmath157 and prevents them from aligning under the flow . it turns out that when we perturb to nonzero angular momentum , the restpoints disappear , as does @xmath156 itself , allowing each incoming window to flow across the neighborhood of the former restpoints to meet each outgoing window . this is shown schematically in figure [ fig_perturb ] . thus there is a chain of alignments connecting each of our ejection - collision windows with each of the others via the new window near @xmath33 . since the qualitative behavior represented by the new window is just the spinning of a small equilateral triangle , no new syzygies or collisions are introduced . \1 . does theorem 1 hold for zero angular momentum ? this was our original question , suggested by wu - yi hsiang . solutions of theorem 2 converge to triple ejection - collision solutions when they are continued to angular momentum zero . another idea will be needed to yield such realizing solutions . how much can we increase angular momentum before our realizing classes disappear ? eventually all of them disappear for topological reasons except those realizing the tight binary " reduced free homotopy classes , meaning the syzygy sequences @xmath158 , @xmath159 , and @xmath160 repeated periodically . topologically speaking , all free homotopy classes are possible below the euler value " in the equal mass case . ( see @xcite . ) do realizations survive right up to this euler value ? \3 . does near - isosceles imply isosceles ? suppose that the angular momentum is zero , the energy is negative and the masses are all equal . is there a positive @xmath161 such that any solution which spends its whole existence within @xmath161 of the isosceles submanifold must be an isosceles solution ? here within @xmath161 " means as measured in the shape sphere , computing the distance of the projected solution to the isosceles great circles . this question is motivated by contemplating the likely fate of the solutions described in this paper as the angular momentum tends to zero , at fixed negative energy . * remark . * the analogous question with the collinear submanifold replacing the isosceles submanifold has the answer no " . the schubart collinear orbit is kam stable with respect to noncollinear perturbations . near to it we have persistent torii filled with noncollinear solutions . can we get rid of all the stutters in a solution realizing a given reduced free homotopy class ? the realizing solutions of theorem 2 have lots of stutters . for the equal mass , under the strong force assumption , and in the case of angular momentum zero , we have proved ( @xcite ) that realizing solutions have no stutters . ( this no - stuttering follows from the fact that the realizers in this case are variational minimizers and variational minimizers can not have stutters . ) more generally , and less precisely , how far can periodic solutions be pushed away from triple collision , away from the hill boundary , into the interior " of the three - body problem ? . we can use the homotopy method to follow any realizing solution from the equal mass , strong force , angular momentum case as discussed in the previous question , changing the potential , the energy , and the angular momentum continuously along some short arc . do any of these persist all the way to the case of the newtonian potential , and in so doing connect up with the realizing solutions described here by theorem 2 ? \6 . which reduced free homotopy classes admit dynamically kam stable realizing solutions ? all realizing solutions obtained by theorem 2 here are unstable . we do not know many such stable classes " . an initial perusal yields only the class @xmath162 of the figure eight , the classes @xmath158 , @xmath159 and @xmath160 as realized by planetary configurations ( depending on which is the dominant mass ) and earth - moon - sun type configurations , and the class @xmath163 realized by broucke - henon solutions we would like to thank wu yi hsiang for posing a version of the realization problem back in 1997 . thanks to carles simo for challenging us to look for a dynamical mechanism behind realization . r.moeckel and r.montgomery gratefully acknowledge nsf grants dms-1208908 and dms-1305844 , respectively , for essential support . a. albouy , h. cabral , and a. santos _ some problems on the classical n - body problem _ , celestial mechanics and dynamical astronomy , 113 ( 2012 ) , 369 - 375 http://link.springer.com/article/10.1007 or http://arxiv.org/abs/1305.3191 r. moeckel , _ some qualitative features of the three - body problem _ , in proceedings of the ams summer research conference on hamiltonian systems , k. meyer and d. saari , contemporary mathematics 81 , ams ( 1987 ) , 1 - 21 .	the configuration space of the planar three - body problem when collisions are excluded has a rich topology which supports a large set of free homotopy classes . most classes survive modding out by rotations . those that survive are called the reduced free homotopy classes and have a simple description when projected onto the shape sphere . they are coded by syzygy sequences . we prove that every reduced free homotopy class , and thus every reduced syzygy sequence , is realized by a reduced periodic solution to the newtonian planar three - body problem . the realizing solutions have nonzero angular momentum , repeatedly come very close to triple collision , and have lots of stutters" repeated syzygies of the same type . the heart of the proof is contained in the work by one of us on symbolic dynamics arising out of the central configurations after the triple collision is blown up using mcgehee s method .
this paper considers ( multivariate ) generating functions of the form : _ n=1^ _ g [ n ] ( g ) s^e(g ) , [ eq : genfn ] where @xmath2 indicates a subclass of graphs . @xmath3 is the number of edges a graph @xmath4 . @xmath5 $ ] indicates that the graph has vertex set @xmath6 $ ] and @xmath7 is a specified positive graph weight . the exponents of the variables @xmath8 and @xmath9 indicate the size of the vertex set , respectively , the number of edges . when evaluating at @xmath10 , there are some remarkable cancellations , leading , in some cases , to simple formul for the coefficients . this paper gives combinatorial explanations for the class of two - connected graphs in particular . two - connected graphs are those graphs for which we can remove any vertex and its incident edges and the resulting graph remains connected . there are four cases of used in mayer s theory of cluster and virial expansions , depending on the class of graphs considered and the weights . the sum is either over connected graphs , denoted by @xmath11 , or two - connected graphs , denoted by @xmath12 . the weights are either those for a discrete hard core gas , often referred to as the one - particle hard core gas , or for a continuum one - dimensional hard core gas , also named the tonks gas . for the discrete gas , the goal is to count the number of graphs ; for the continuum model , the coefficients are given by the volume of a polytope associated with the graph @xmath4 . we write a graph @xmath4 as the ordered pair of its vertex set and edge set as @xmath13 . we define the polytope corresponding to the graph @xmath4 as : _ g : = \{()_[2,n]^n-1 \|_i-_j\|<1 \{i , j } e(g ) } [ eq : graphpolytope ] , with @xmath14 . we use the notation @xmath15}:=(x_2 , \cdots , x_n)$ ] . mayer , in @xcite , established important connections between weighted graph generating functions and expansions in statistical mechanics . these connections are also presented in the framework of combinatorial species of structure in the work of ducharme , labelle and leroux @xcite , leroux and kaouche @xcite and faris @xcite . the results of mayer are that the weighted sum over connected graphs gives the pressure as a function of activity and the weighted sum over two - connected graphs is related to the virial expansion of pressure expanded in terms of density . the two formul are : @xmath16 } w(g ) \\ \beta p(\rho ) & = \rho - \sum\limits_{n=2}^{\infty}(n-1 ) \frac{\rho^n}{n ! } \sum\limits_{g \in \mathcal{b}[n]}w(g ) , \end{aligned}\ ] ] where @xmath17 is the graph weight specified by the particular model . the answers for the four cases are given by the formul for the connected graph discrete case : _ g [ n](-1)^e(g)= ( -1)^n-1(n-1 ) ! and for the connected graph continuum case : _ g [ n](-1)^e(g)(_g ) = ( -1)^n-1n^n-1 . there are also formul for the two - connected discrete case : _ g [ n](-1)^e(g)=-(n-2 ) ! [ eq : virial1phc ] and the two - connected graph continuum case : _ g [ n](-1)^e(g)(_g)=-n(n-2 ) ! [ eq : virialtg ] .for the discrete cases the results are straightforward computations . for the continuum case , derivations are given in @xcite . the statistical mechanical background is explained in full detail in section [ sec : models ] . it is tempting to try and find a simple combinatorial interpretation that explains the cancellations in a direct way . this was posed as a challenge in the paper of ducharme , labelle and leroux @xcite . in the connected graph cases , this was done by bernardi @xcite . the approach was to use an involution that exhibits the result of the almost perfect cancellation as a contribution from the fixed points of the involution . the fixed points were identified as increasing trees in the discrete case and rooted trees in the continuum case . the purpose of this paper is to present similar derivations for the two - connected graph cases . as always this is considerably more complicated . the concept of using an involution to understand the cancellations is natural . recall the formula that , for any finite non - empty set @xmath18 , we have : _ s x(-1)^\|s\|=0 . in order to prove this , we show we have the same number of sets with even cardinality as we do of odd cardinality . one approach is to pair sets differing by one element . this pairing idea is captured by the involution . in this example , the involution is defined by first fixing a singleton subset of @xmath18 , say @xmath19 , and taking the symmetric difference @xmath20 . if we consider a fixed vertex set @xmath5 $ ] for a graph , then a graph @xmath4 is determined precisely by its edge set @xmath21 , which are subsets of the collection of unordered pairs in @xmath5 $ ] , denoted @xmath5^{(2)}$ ] . we can also use this symmetric difference operation on the edge set for graphs . an important complication is that we consider particular subsets for which taking the symmetric difference with a fixed edge will not suffice , since the removal or addition of the edge may take us outside of the prescribed collection of subsets . we need to find an efficient way of choosing an edge based on the graph we are considering so that we obtain a pairing that will not take us outside of the prescribed collection . in section [ sec : results ] , we present the combinatorial structures that give the interpretations of the cancellations in the two - connected case . sections [ sec:1phcinv ] and [ sec : chcginv ] give the proofs of the one particle hard core and the tonks gas case respectively . in the latter , the decomposition of polytopes into unimodular simplices attributed to lass is given so that it may be proved as an extension of the previous case . we provide an interpretation why @xmath22 should appear as the number of edges in section [ sec : structure ] . from the perspective of statistical mechanics , the motivation for understanding such cancellations is to be able to adapt the understanding to models where more complicated weights are used . indeed , the key idea is to emulate what is done for the connections between connected graphs and trees and understand how to modify these in this context . the first parallel to draw is that the involution of bernardi fits within a general concept of externally and internally active elements of a set with a matroid structure as given by bjrner and sokal @xcite . the idea to emphasise here is that this allows the set of connected graphs to be partitioned into subsets , indexed by trees . when we consider graphs with the partial order defined by bond inclusion , the blocks in this partition are boolean . that is , each set has a tree @xmath23 as minimal graph and a corresponding maximal graph @xmath24 , all graphs with edge set @xmath25 such that @xmath26 are included in the set in the partition . this form of a partition lends itself well to performing estimates on the cluster coefficients . this was actually realised earlier by penrose @xcite in the specific case of connected graphs . understanding this partition into boolean subsets also gives rise to an alternative involution . it is intriguing to realise that the general construction does not include the penrose construction as a subcase . these ideas are addressed in section [ sec : extend ] . this combinatorial understanding is also closely linked to the tree - graph identities of brydges battle and federbush @xcite , for which a symmetric version is provided by abdesselam and rivasseau @xcite and a matroid generalisation by faris @xcite . these identities allow estimations to be made on these coefficients , since we may express the sum over connected graphs as a sum over trees with modified weights . a greater goal is to extend these to partially ordered sets where a matroid structure may not be present . interest in providing such bounds on the virial expansion coefficients has recently been renewed with the papers by pulvirenti and tsagkarogiannis @xcite and morais and procacci @xcite , which use the canonical ensemble as a method of achieving bounds . the paper by jansen @xcite suggests that at high temperatures the radius of convergence should be improved : actual improvements on the bounds of lebowitz and penrose @xcite have been proposed recently @xcite . in a classical gas system of @xmath0 indistinguishable interacting particles in a vessel @xmath27 with only two - body interactions and no external potential , we may write the hamiltonian as : h(,)=_i=1^n + _ 1 i < j n(q_i , q_j ) , where @xmath28 represents the generalised coordinates and @xmath29 the conjugate momenta . the canonical partition function of the gas model is : z(,,n)=_i=1^n ( _ ^dq_i _ ^d ^dp_i ) ( - h ) . integrating out the gaussian integrals for the momenta , we obtain a factor @xmath30 , where @xmath31 is the thermal wavelength . the partition function is therefore : z ( , , n)= _ i=1^n ( _ ^dq_i ) _ 1 i < j n(-(q_i , q_j ) ) [ eq : canonicalpfinvol ] . the mayer trick @xcite , allows us to rewrite the canonical partition function in terms of weighted graphs . the first stage is to define the mayer @xmath32-function : f(q_i , q_j):=(-(q_i , q_j))-1 .we realise that the product of exponentials in may be rewritten as : _ 1 i < j n ( -(q_i , q_j))=_1 i < j n(1+f(q_i , q_j))=_g [ n]_(i , j ) e(g)f(q_i , q_j ) , where @xmath33 $ ] is the set of simple graphs ( no multiple edges or loops ) on @xmath0 points . we write a graph @xmath34 , where @xmath35^{(2)}$ ] is the edge set and @xmath36 $ ] is the vertex set . this motivates the graph weight : w(g)=_i=1^n ( _ ^dq_i ) _ ( k , l ) e(g)f(q_k , q_l ) . we can therefore write the partition function as : z ( , , n)=_g [ n]w(g ) . in order to obtain the grand canonical partition function we sum : ( , , z)=_n=0^ z^n^n z ( , , n ) , where @xmath37 the activity and @xmath38 is the chemical potential . in terms of graphs , we write this as : ( , , z)=_n=0^_g [ n]w(g ) = : _ the pressure is defined to be : p= _ \|\| ( , , z ) .if we define the new weight @xmath39 , then the pressure function can be written in terms of connected graphs : p = _ w(z)=_n=1^ _ g [ n]w(g ) [ eq : pressureconnectedinv ] .this is the content of mayer s first theorem @xcite and is explained in the paper @xcite . the density @xmath40 is : = z p = ^_w(z ) , where @xmath41 denotes a rooted connected graph . from mayer s second theorem @xcite or by the dissymmetry theorem @xcite , we are able to obtain a series expansion for pressure in terms of density , in which the coefficients are , up to a prefactor , the @xmath42-weighted two - connected graphs . p = - _ n=2^ _ g [ n]w(g ) [ eq : pressure2connectedinv].one may also consult the book by mccoy @xcite for an explanation of the derivation of these two theorems . the potential for a one - particle hard core gas is : ( q_i , q_j)= , so that @xmath43 and @xmath44 . the grand canonical partition function is : ( z)=1+z . the statistical mechanical relationships give pressure and density as : @xmath45 we may invert , to obtain : z = and substitute for @xmath8 in , to obtain : p = - ( 1- ) . the two series expansions derived from statistical mechanics are : @xmath46 if we compare these two power series with and respectively , using the graph weight @xmath47 , where @xmath3 is the number of edges in graph @xmath4 , we obtain : @xmath48}(-1)^{e(g)}&=(-1)^{n-1}(n-1 ) ! \\ \sum\limits_{g \in \mathcal{b}[n]}(-1)^{e(g)}&=-(n-2 ) ! .\end{aligned}\ ] ] for a continuum hard core gas in one dimension with diameter @xmath49 , the potential is : ( q_i , q_j)= & \|q_i - q_j\|<1 + 0 & . the exponential and mayer @xmath32-functions are : @xmath50 we therefore have the graph weight : w(g)=(-1)^e(g)_^n-1 _ \{i , j } e(g)(\|x_i - x_j\|<1 ) x_2 x_n , where @xmath14 and @xmath51 is the indicator function . in @xcite , this is interpreted as a the volume of a convex polytope @xmath52 in @xmath53 . the polytope is defined by : _ g = \ { ( ) _ [ 2,n ] ^n-1 \|x_i - x_j\|<1 \{i , j } e(g ) x_1=0 } . we use the notation @xmath54=\{2 , 3 , \cdots , n\}$ ] and @xmath15}=(x_2 , \cdots x_n)$ ] . hence the graph weight may be written as : w(g)=(-1)^e(g)(_g ) . the derivation of the cluster and virial expansions , using statistical mechanics , are more difficult in this case , but they are done in @xcite and we achieve : @xmath55 where @xmath56 is the lambert @xmath57-function . if we compare these to the results of mayer s first and second theorems , and , we obtain the combinatorial relationships : @xmath48}(-1)^{e(g)}{\mathop{\mathrm{vol}}}(\pi_g ) & = ( -1)^{n-1}n^{n-1 } \\ \sum\limits_{g \in \mathcal{b}[n]}(-1)^{e(g)}{\mathop{\mathrm{vol}}}(\pi_g ) & = -n(n-2 ) ! .\end{aligned}\ ] ] the results of this article are the combinatorial interpretations of the cancellations in the alternating sums of weighted two - connected graphs . [ thm:1phc ] the difference of two - connected graphs with an even number of edges and an odd number of edges is given by the following formula : _ g [ n](-1)^e(g)=-(n-2)!. this is proved through an involution @xmath58 , given in section [ sec:1phcinv ] , which effectively pairs graphs differing by only one edge , leaving some small collection of graphs fixed , which give the @xmath59 factor . the fixed graphs are formed from an increasing tree on the vertex set @xmath60 $ ] with the vertex @xmath0 adjacent to every other vetex . the number of increasing trees on @xmath60 $ ] is @xmath59 . the tree has @xmath61 edges and we add @xmath62 edges from the vertex labelled @xmath0 to achieve @xmath22 edges . this gives the definite minus sign and the combinatorial factor . an increasing tree is a labelled tree on which the sequence of vertex labels along all paths from the vertex labelled @xmath49 to the leaves form increasing sequences . an example of such a graph is shown in figure [ fig : increasingtree ] . [ thm : chcg ] when we add the polytope weights to the alternating graph sum , we achieve the following identity : _ g [ n](-1)^e(g)(_g)=-n(n-2 ) ! . this is proved through a collection of involutions @xmath63 . the index @xmath64 is related to the partition of the polytopes into areas of equal volume attributed to lass in @xcite . the meaning of @xmath64 is explained in subsection [ subsec : lasspolytope ] . the fixed points of these involutions occur only when @xmath64 is of the form @xmath65 , meaning that any edge is possible . there are precisely @xmath0 possibilities of these sequences , which corresponds to the @xmath0 positions of the last zero . the particular @xmath64 provides a bijection @xmath66 \to [ n]$ ] on which the fixed graphs correspond to an increasing tree ( given by the order @xmath67 if and only if @xmath68 ) on the labels @xmath69 . this is paired with every edge from @xmath70 to the vertices @xmath69 . the number of these increasing trees on @xmath62 vertices is @xmath59 and hence we obtain the factor @xmath71 . we notice that these graphs are on @xmath22 edges as above , which provides the minus sign . the two - connected case is necessarily more complicated than the connected case . first of all , minimal two - connected graphs do not all have the same number of edges for a fixed number of vertices as trees ( minimal connected graphs ) do . simply removing edges appropriately down to a minimal graph can not provide a combinatorial understanding as there will still be sign differences to take care of . furthermore , the sign of the factor is constant - the number of edges must always be odd for whatever value of @xmath0 we take . as indicated in the introduction , the proof of theorem [ thm:1phc ] is done through an involution . to explain how the involution @xmath58 provides the combinatorial factor through the number of fixed points , we use the manipulations of bernardi @xcite , where we know that the involution either adds an edge , removes an edge or leaves the graph fixed . we have that : _ g [ n](-1)^e(g)=_g [ n](-1)^e((g ) ) , since @xmath58 is a bijection . the sum of these is therefore : @xmath72}(-1)^{e(g)}&=\sum\limits_{g \in \mathcal{b}[n]}((-1)^{e(g)}+(-1)^{e(\psi(g ) ) } ) \notag \\ & = 2\sum\limits_{g \in \mathcal{b}[n ] \vert \psi(g)=g}(-1)^{e(g ) } . \end{aligned}\ ] ] the fixed points of the involution thus give us the combinatorial factor . this section describes the involution and proves it does what is required . for graphs , the analogous operation to symmetric difference explained in the introduction is the operation @xmath73 . @xmath74 is the graph @xmath75 . the specific task of the proof of both identities is to identify for each graph a unique edge that we can add or remove . this has to be done in a consistent and efficient manner . consistent in the sense that if we identify @xmath76 as the unique edge in @xmath4 , then we want @xmath77 so that @xmath58 is an involution . it needs to be efficient in the sense that the only graphs it leaves fixed are those that provide the combinatorial factor relevant for the alternating sum . we do not want further cancellations to consider . in each graph @xmath4 , we consider the vertex labelled @xmath0 . when the vertex @xmath0 is adjacent to every vertex , we realise that the collection of two - connected graphs with this property may be identified with the collection of connected graphs on the vertex set @xmath60 $ ] . bernardi @xcite has already provided an involution on this set that we can use in this case to obtain cancellations , since they will all come with the same prefactor @xmath78 from the @xmath62 edges from the vertex labelled @xmath0 . we thus firstly introduce the involution of bernardi and make rigorous the connection between connected graphs on @xmath60 $ ] and the particular subset of two - connected graphs where @xmath0 is adjacent to every other vertex . for those graphs where the vertex labelled @xmath0 is not adjacent to every other vetex , we may use the two - connected property of the graph to find an edge suitable for the involution . this is done through using a corollary due to whitney of menger s theorem and introducing a definition of permissible edges . we emphasise how these combine to give a complete involution and that the only contributions arise from the bernardi involution . firstly , we define the neighbourhood of a vertex @xmath79 in a graph @xmath4 . for a graph @xmath4 and a vertex @xmath79 , we define the neighbourhood of @xmath79 in @xmath4 as @xmath80 . we define the _ lexicographic order _ on edges @xmath81^{(2)}$ ] by : \{i , j } < \{k , l } \{i , j } < \{k , l } + \{i , j } = \{k , l } \{i , j } < \{k , l } . for a subset @xmath82 of a totally ordered set , we define @xmath83 . for a graph @xmath84 and an edge @xmath85 , we define @xmath86 with respect to the lexicographic order above . we give here bernardi s involution on connected graphs , since it used for the two - connected graph version . we write it for the vertex set @xmath60 $ ] as this is the form in which it will be used . an edge @xmath87^{(2)}$ ] is externally active for the graph @xmath88 $ ] , if there is a path in @xmath89 between the endpoints of @xmath85 . if a connected graph @xmath4 has an externally active edge , we define @xmath90 to be the maximal such edge . the involution @xmath91 \to \mathcal{c}[n-1]$ ] defined by bernardi @xcite , is given by : _ b : g g _ g & g + g & . the result of the involution is the following lemma . under the involution @xmath92 , only increasing trees are kept fixed . we introduce the following notation to simplify the formulation of the connection between two - connected graphs with vertex set @xmath5 $ ] , where the vertex labelled @xmath0 is adjacent to every other vertex , and connected graphs with vertex set @xmath60 $ ] . * for a graph @xmath84 , we denote by @xmath93 , the graph @xmath94 . * we denote the subset of two - connected graphs on vertex set @xmath5 $ ] with vertex @xmath0 adjacent to all other vertices by @xmath95 $ ] . we define the mapping @xmath96 \to \mathcal{c}[n-1]$ ] , by @xmath97 . we emphasise that removing a vertex and its incident edges from a two - connected graph leaves a connected graph and so defining the codomain of @xmath98 as @xmath99 $ ] is fine . the map @xmath96 \to \mathcal{c}[n-1]$ ] is a bijection . firstly it is injective . if @xmath100 , this means @xmath101^{(2)}=e(h ) \cap [ n-1]^{(2)}$ ] and since @xmath4 and @xmath102 $ ] , the remaining elements of @xmath21 and @xmath103 , namely @xmath104\}$ ] , are the same and so @xmath105 . this is surjective , since for any connected graph on @xmath60 $ ] , if we add the vertex labelled @xmath0 and all edges @xmath106 such that @xmath107 $ ] , the resulting graph is two - connected . if we consider removing any vertex @xmath108 from this new graph we see that every vertex is connected to every other vertex via @xmath0 . if @xmath0 is removed then it is connected by definition and hence it is two - connected . we define the inverse map of @xmath98 to be @xmath38 . a path is an alternating sequence of vertices and edges in a graph @xmath109 , which begins and ends with a vertex . the edges are written in terms of the preceding and following vertices : @xmath110 . two paths @xmath109 and @xmath111 are internally disjoint if the only common vertices or edges are the endpoints @xmath112 and @xmath113 . for an edge @xmath114 the endpoints are defined as the vertices @xmath79 and @xmath115 . the following result in the case @xmath116 is used to find an edge in each graph where @xmath117 $ ] , by using the fact that we have two internally disjoint paths between @xmath0 and some @xmath118 \setminus n_g(n)$ ] . this is a classical theorem of whitney @xcite based upon menger s theorem . [ thm : whitney ] a graph @xmath4 is @xmath119-connected if and only if every pair of vertices is connected by @xmath119 internally disjoint paths . we introduce the notion of permissible edges as those edges which have both endpoints in the neighbourhood of a vertex @xmath79 and can easily be understood as a chord in the graph @xmath4 , when we neglect any edges in @xmath120 . we will focus on the case when @xmath121 . given a ( two - connected ) graph @xmath4 and a vertex @xmath122 , such that @xmath123 , we define an edge @xmath81^{(2)}$ ] to be @xmath124-permissible if the following condition holds : @xmath125 there exists an @xmath126 , such that we have two vertex disjoint paths @xmath127 , intersecting each once in @xmath128 . the intersection vertices are the endpoints of @xmath85 . if a two - connected graph @xmath4 with @xmath36 $ ] has a @xmath129-permissible edge , then we denote the largest such edge in lexicographical ordering by @xmath130 . for every @xmath131 \setminus \mathcal{b}^{\delta}[n]$ ] , we have a @xmath129-permissible edge . we know that @xmath132 \setminus n_g(n ) \neq \emptyset$ ] , because we are outside of @xmath95 $ ] . if we choose some @xmath133 , then we know by theorem [ thm : whitney ] we have two internally disjoint paths between the vertices labelled @xmath134 and @xmath0 . both paths must hit @xmath135 at some point . when they first hit @xmath135 , then they could go straight to @xmath0 and so each path need only intersect @xmath135 in one place . this provides us with a permissible edge and so @xmath130 is well defined for every @xmath131 \setminus \mathcal{b}^{\delta}[n]$ ] . we define the involution @xmath136 \to \mathcal{b}[n]$ ] through bernardi s involution @xmath92 and the permissible edge concept . * if @xmath137 $ ] , we consider the graph @xmath138 . this is a connected graph and we may apply bernardi s involution to this subgraph and retain the vertex @xmath0 and its incident edges . + this can be written as @xmath139}:=\mu \circ \psi_b \circ \zeta$ ] . * if @xmath117 $ ] , then we define the involution @xmath140 . the first point to emphasise is that due to the bijection between @xmath95 $ ] and @xmath99 $ ] , we are able to obtain cancellations for these graphs in the same way as bernardi . we are left with increasing trees on the set @xmath60 $ ] and the vertex @xmath0 adjacent to every other vertex . we still need to prove that @xmath58 is indeed an involution . @xmath58 is an involution and its image is contained within @xmath141 $ ] . the fact this is true for @xmath142}$ ] follows from the proof of bernardi . if an edge is permissible , we note that it is a chord in a cycle within the graph @xmath4 . if we add an edge to a two - connected graph it remains two - connected . we prove below that if we remove a chord from a two - connected graph , then it remains two - connected . we denote the chord we are considering by @xmath143 , the original graph by @xmath4 and the graph @xmath144 by @xmath145 . we prove @xmath145 is two - connected by considering the effect of removing a vertex from @xmath145 . there are two cases : * @xmath146 and @xmath147 are connected as they are the same graphs as @xmath148 and @xmath149 respectively , which are connected since @xmath4 is two - connected . * if we consider another vertex @xmath119 . we assume for contradiction that the graph @xmath150 is not connected . we know @xmath151 is connected and the only difference is that we have the additional edge @xmath152 . this would then imply that @xmath79 and @xmath115 are in different connected components in @xmath150 . we know that @xmath79 and @xmath115 appear in a cycle in @xmath145 . this means if we remove one vertex then we still have a path between @xmath79 and @xmath115 , hence we obtain a contradiction unless @xmath150 is connected . the collection of permissible edges depends only on edges within @xmath132 \setminus n_g(n)$ ] , between @xmath82 and @xmath135 and edges involving @xmath0 . adding or removing a permissible edge does not change the available edges on which one can make the two internally disjoint paths . hence the collection of permissible edges for @xmath4 and @xmath153 are the same . this means that the largest elements in each set are the same i.e. @xmath154 . therefore it is an involution . hence , @xmath58 is an involution and has only fixed points in the set @xmath95 $ ] . the fixed points are those given by bernardi as increasing trees on the vertex set @xmath60 $ ] with @xmath0 adjacent to all vertices in @xmath60 $ ] . in order to deal with the polytope volume weights , we decompose the polytopes into simplices . this first appeared in @xcite and is used in @xcite to prove the connected graph case . this splitting of polytopes into unimodular simplices is attributed to lass . this subsection explains how this splitting of the polytopes into simplices is used to construct the involution for the continuum case . these ideas are important in reducing the case of the tonks gas to the one particle hard core model . the key idea is to split @xmath53 into @xmath155-simplices of equal volume . we then realise that a polytope either fully contains a simplex , intersects only on the boundary of the simplex or is disjoint from the simplex . the sum is then reorganised so that we may sum over each simplex on the outside and then undertake the alternating sum on the restricted set of graphs whose associated polytopes contain the simplex considered . consider @xmath15 } \in \mathbb{r}^{n-1}$ ] and let @xmath156 be the integer part of @xmath157 and @xmath158 be the fractional part such that @xmath159 . let @xmath160 \to [ 2,n]$ ] be a bijection . we may define the simplex @xmath161 , by the set of @xmath162 with integer part @xmath64 and whose fractional parts satisfy : @xmath163 . this simplex has volume @xmath164 . the condition @xmath165 is equivalent to @xmath166 . we therefore have that @xmath167 if and only if for all @xmath168 , we have that @xmath169 with @xmath170 and @xmath171 . for any graph @xmath172 $ ] , the value @xmath173 counts the pairs @xmath174 and @xmath175 such that @xmath176 is a subpolytope of @xmath52 . we may rearrange the sums over connected or two - connected graphs of the graph weights by first casting the sum as a sum over the pairs @xmath177 and symmetrising the weight over isomorphic graphs . the symmetrisation procedure can be understood by considering a permutation @xmath178 of @xmath54 $ ] and defining for any vector @xmath179 , @xmath180 . for any graph @xmath4 with labels in @xmath5 $ ] , the graph @xmath181 is the graph , with the same vertex set and satisfies @xmath182 . @xmath167 if and only if @xmath183 for any permutation @xmath178 of @xmath54 $ ] . this equivalence can be elucidated by rewriting @xmath184 and @xmath185 . this allows us to rewrite the latter statement as : @xmath186 . this implies , for the entries in vector @xmath187 , that @xmath188 , @xmath189 . since @xmath190 , we may rewrite this as : @xmath191 @xmath192 . we make the identification that @xmath193 to see that we get precisely the statement that @xmath194 . we let @xmath2 denote either @xmath11 or @xmath12 and then we rewrite : @xmath195 \\ \pi(\mathbf{h } , \sigma ) \subset \pi_g}}(-1)^{e(g ) } & = \sum\limits_{\substack{\mathbf{h } \in \mathbb{z}^{n-1 } \ , g \in \mathcal{h}[n ] \\ \pi(\sigma^{-1}(\mathbf{h}),\text{id } ) \subset \pi_{\sigma(g)}}}(-1)^{e(g ) } \notag \\ & = \sum\limits_{\substack { \mathbf{h } \in \mathbb{z}^{n-1 } \ , g \in \mathcal{h}[n ] \\ \pi(\mathbf{h } , \text{id } ) \subset \pi_g}}(-1)^{e(\sigma^{-1}(g ) ) } \notag \\ & = \sum\limits_{\substack{\mathbf{h } \in \mathbb{z}^{n-1 } \ , g \in \mathcal{h}[n ] \\ \pi(\mathbf{h } , \text{id } ) \subset \pi_g}}(-1)^{e(g ) } \end{aligned}\ ] ] we may therefore , understand the weight as : @xmath196}w(g)&=\sum\limits_{g \in \mathcal{h}[n]}(-1)^{e(g)}{\mathop{\mathrm{vol}}}(\pi_g ) = \frac{1}{(n-1)!}\sum\limits_{\substack { \mathbf{h } \in \mathbb{z}^{n-1 } \sigma \in s_{n-1 } \\ \text{such that } \pi(\mathbf{h},\sigma ) \subset \pi_g}}(-1)^{e(g ) } \notag \\ & = \sum\limits_{\substack{\mathbf{h } \in \mathbb{z}^{n-1 } \ , g \in \mathcal{h}[n ] \\ \pi(\mathbf{h},\text{id } ) \subset \pi_g}}(-1)^{e(g ) } \end{aligned}\ ] ] we define the _ centroid _ of the vector @xmath64 , by @xmath197 , where @xmath198 and @xmath199 . we define @xmath200 as the graph on @xmath5 $ ] where the edges are all pairs @xmath114 such that @xmath201 . we define @xmath202:=\{g \in \mathcal{h}[n ] \vert e(g ) \cap e(k_{\mathbf{h}})=e(g ) \}$ ] where @xmath2 can be replaced by @xmath11 or @xmath203 . the final sum indicates that we need to count pairs @xmath64 and @xmath4 such that @xmath204 . that is that the centroid @xmath205 , since @xmath206 is in the interior of @xmath207 . this can be recast as : for @xmath205 , we require that : \{i , j } e(g ) _ [ n](-1)^e(g ) we can thus consider the total sum as first a sum over the subset of graphs @xmath202 $ ] for each @xmath64 and add the results . this leads to considering separate @xmath208 \to \mathcal{b}_{\mathbf{h}}[n]$ ] which are involutions and finding their fixed points . we define an involution @xmath209 for each @xmath174 on the set @xmath210 $ ] of two connected graphs , which are compatible with the vector @xmath64 . we note that , by the definition of @xmath210 $ ] , edges with @xmath211 are forbidden . we call an edge @xmath212 such that @xmath213 allowed . in order to make the connection with the proof in the discrete case , we indicate a bijection @xmath214 related to the particular @xmath64 that provides a suitable relabelling of the vertices to allow for an efficient application of the lemmas of section [ sec : chcginv ] to a relabelled graph . we reframe the consequences of these lemmas in the context of the allowed edges . it is important to check that an edge we may want to add or remove by the prescription in section [ sec : chcginv ] is allowed within the specific collection of graphs @xmath210 $ ] . it is then proved that when we have a non empty set of forbidden edges , all terms cancel . in the case when the set of forbidden edges is empty , we obtain the exact values taken by @xmath64 and everything reduces to the discrete gas case with a relabelling . we have a definite order on the entries of @xmath206 , since each entry has a different fractional part . we define a re - ordering of the set @xmath5 $ ] , through a bijection @xmath215 \to [ n]$ ] . this re - ordering is defined through the order for the entries of @xmath206 : @xmath216 . the re - ordered lexicographic order on edges is given by : \{_(i),_(j ) } < \{_(k),_(l ) } \{i , j } < \{k , l } + \{i , j } = \{k , l } \{i , j } < \{k , l } . instead of considering @xmath129-permissible edges , we consider @xmath217-permissible edges since it makes the formulation of the involution easier . [ lem : permisallowed ] all edges @xmath218 are allowed . we realise that @xmath219 we have that @xmath220 and so for every pair @xmath221 , @xmath213 and hence the edge is allowed in @xmath210 $ ] . all @xmath217-permissible edges are allowed . if @xmath222)$ ] , then @xmath210=\mathcal{b}[n]$ ] and @xmath64 is of the form of an initial sequence of zeroes with remaining entries @xmath223 . by lemma [ lem : permisallowed ] , all edges in @xmath224)^{(2)}$ ] are allowed . the edges @xmath225 for all @xmath226)$ ] are already in the graph and so can not be forbidden . hence every edge is allowed and so @xmath210=\mathcal{b}[n]$ ] . since @xmath199 , this means @xmath227 for all @xmath79 . we also note that if @xmath228 , then @xmath229 for all @xmath230 . this arises from the fact that the entries of @xmath64 are restricted to @xmath231 . for a negative entry we will have @xmath232 , which is not within distance @xmath49 of the value @xmath233 for any @xmath230 this means that @xmath64 is of the special form of an initial sequence of zeroes with the remaining entries @xmath223 . we define the set @xmath234 $ ] as the collection of two - connected graphs where @xmath235 is adjacent to all other vertices . we have the corresponding maps @xmath236 and @xmath237 between @xmath234 $ ] and @xmath238 \setminus \{\xi_{\mathbf{h}}(n)\}]$ ] , which are the same as in section [ sec : chcginv ] , except we are removing the vertex @xmath235 instead of @xmath0 . formally , we can write these bijections as a conjugation with @xmath214 , when interpreted as its action on graphs . in this case : @xmath239 we define the modified bernardi involution @xmath240 as in section [ sec : chcginv ] , except @xmath89 is interpreted in the sense of the re - ordered lexicographic ordering and for @xmath90 to be maximal externally active edge we use this ordering too . this can also be simply written using the graphical label conjugation : _ b , : = _ _ b _ ^-1 the largest ( using the re - ordered lexicographic order ) @xmath241-permissible edge is denoted by @xmath242 . we define @xmath209 as the involution on @xmath210 $ ] , defined by : * if @xmath222)$ ] , then we may use a modified version of bernardi , since all edges are possible in @xmath243 . ^_[n ] : = _ _ b , _ * otherwise , we have a permissible edge and can perform the involution @xmath244 . @xmath209 retains the property of being an involution on two - connected graphs as in section [ sec : chcginv ] . we are thus left with only those graphs that have @xmath222)$ ] and are increasing with respect to the re - ordered lexicographic order . the only @xmath64 vectors that contribute are those with an initial sequence of zeros followed by @xmath223s . there are @xmath0 possibilities of these sequences , since the final @xmath245 can appear in any of the entries @xmath246 . the permutation @xmath214 related to the @xmath64-vector with @xmath247 for @xmath248 and @xmath249 for @xmath250 takes the special form : _ : s n. we observe that the entry @xmath251 has the smallest value so @xmath252 . we then note that the following entries are negative and are in increasing order . the preceding entries are also in increasing order but are positive . hence we have @xmath253 for @xmath254 and @xmath255 for @xmath256 , which can be written in the form in the lemma for brevity . hence we have a precise collection of two - connected graphs . we have the examples from section [ sec : chcginv ] with these linear relabellings . for @xmath257 , the fixed graphs are all distinct . we indicate that there are no labelled graph automorphisms of the form of @xmath214 described above for the increasing trees on @xmath60 $ ] with @xmath0 adjacent to every vertex . the first observation is that @xmath214 has no fixed vertex labels . we know the degree of the vertex labelled @xmath0 is @xmath62 . if we were to have an automorphism with no fixed labels , then we require another vertex of the same degree to send @xmath0 to . this means we need a vertex in the increasing tree adjacent to all other vertices in the increasing tree . when a tree has at least three vertices , only one vertex can be adjacent to the rest , since if we have two vertices adjacent to all vertices we have them adjacent to each other and some third vertex . this creates a @xmath258-cycle contradicting the fact a tree is acyclic . furthermore , in this increasing tree , this vertex can only be the vertex labelled @xmath49 or @xmath259 . for any @xmath260 $ ] , @xmath119 can not be attached to both @xmath49 and @xmath259 , or else we will have a @xmath258-cycle , as we always have the edge @xmath261 . we therefore require that the graph automorphism exchanges the labels of the two vertices . the automorphisms are translations and since @xmath262 or @xmath263 , we have to translate by @xmath49 or @xmath259 , but then the vertex labelled @xmath49 or @xmath259 would not be relabelled as @xmath0 as we would require . hence @xmath214 is not an automorphism for any of the prescribed graphs and so the collection of these graphs for @xmath257 are all distinct . in this section , we indicate how the structure of two - connected graphs indicates the importance of graphs with @xmath22 edges . firstly , we explain some preliminary concepts about block cutpoint trees and then use these to explain why minimal two - connected graphs , that is a two - connected graph , such that the removal of an edge renders the graph no longer two - connected , on @xmath0 vertices have at most @xmath264 edges . in this section , we introduce the notion of a block cut - point tree and state a result relating the number of vertices in the individual blocks to the number of vertices in the whole graph . we use the notation @xmath265 to denote the collection of trees . * an _ articulation point _ in a connected graph is a vertex , which when it and its incident edges are removed , renders the graph disconnected . a synonym that is frequently used is a _ cutpoint_. * a _ two - connected graph _ is a connected graph without articulation points . * a _ block _ is a maximal two - connected subgraph of a connected graph . maximal in terms of edges and vertices it includes . the block cutpoint tree ( bc - tree ) associated to a connected graph @xmath4 is a ( bipartite ) graph where the vertices represent the articulation points and the blocks in a connected graph . an edge , between an articulation point and a block , is present in this graph , when an articulation point is contained in a block . it is a tree , since if there were a cycle in this graph then the cycle itself would have been a block . an example of a block cutpoint tree is shown in figure [ fig : bctreethesis ] . to define the centre of a tree formally , we define first the _ eccentricity _ @xmath266 of a vertex @xmath267 as the minimal graph distance of @xmath267 to a leaf . this may be formally written as @xmath268 , where @xmath269 indicates the hamming or graph distance in the tree . the centre of a tree is the collection of vertices at which the maximum eccentricity is attained . this can either be two neighbouring vertices or a single vertex . in the former case , we often call the edge between the vertices the centre of the tree . one can apply the function @xmath270 , which for any given tree , removes all leaves and the edges incident to the leaves . formally , we can write this as : f:(v(),e ( ) ) ( v ( ) l , e ( ) ( l v ( ) ) ) , where @xmath271 , the collection of leaves . repeated application of @xmath32 , gives a sequence of trees , @xmath272 which becomes constant either when we have a single vertex or the empty graph . in the case of the single vertex , this is the centre of the tree . for the empty graph , the penultimate step will have been two vertices and an edge . this edge or the pair of vertices is defined as the centre . a bc - tree is bipartite with all leaves in one set ( the blocks ) . it therefore has a unique centre , since the eccentricity of the articulation points will be odd and the eccentricity of the blocks will be even so two neighbours can not have the same maximum eccentricity . since we have a unique vertex at the centre of the bc - tree , we may define a digraph arising from the bc - tree , where the edge is oriented to point away from the centre . an example is displayed in figure [ fig : digraph ] . [ lem : blockdecomp ] if we decompose a connected graph on @xmath0 vertices into its block structure and let @xmath273 index the collection of blocks and @xmath274 be the sequence of block sizes , then we have the following equality : _ i i(k_i-1)=n-1 [ eq : bctreeidentity ] . the key idea is to indicate what vertex we omit inside each block on the left hand side of . the digraph gives an ( essentially ) unique prescription of the missing vertex in each block and in which block an articulation point is counted . the digraph comprises of two types of directed edge @xmath275 and @xmath276 , where @xmath134 indicates an articulation point and @xmath277 a block . the arrow points from the first entry to the second entry . since there is a unique path from the centre to every other vertex , every vertex has precisely one edge in which they are the second entry . there are two key cases : * * the centre is an articulation point * + for a block , @xmath277 , the unique vertex we neglect on the left hand side of is the articulation point , @xmath134 , where @xmath276 is the directed edge in the digraph . + every articulation point , @xmath278 , except the centre appears in an edge @xmath279 , for which it is the second entry , meaning it is enumerated in the left hand side of in precisely one block . the central articulation point is the only neglected vertex , which gives the right hand side of . * * the centre is a block * + in this case every block , except the centre , can be given the prescription as for the first case . for the central block , we can choose precisely one of its neighbours to neglect . all articulation points in this case have an edge in which they are the second entry and so are counted , excepting the articulation point identified by the central block . therefore , we have . to understand why the two - connected graphs on @xmath0 vertices with @xmath22 edges play a special role , we first indicate that two - connected graphs with at least this number of edges can not be minimal . given a graph @xmath4 on the vertex set @xmath5 $ ] , we denote by @xmath280 , the degree of the vertex labelled @xmath49 . two - connected graphs on @xmath0 vertices with @xmath281 edges are not minimal , that is they necessarily have a chord . this is done by induction on the number of vertices @xmath0 . the cases @xmath282 are vacuous and one can see from the examples in figure [ fig:5edges4vertices ] that this holds when @xmath283 . the connected graph @xmath284 may be decomposed into its bc - tree . each block with @xmath285 vertices in the tree has to have @xmath286 edges or else we have a smaller graph which has a chord by induction . we note here that blocks of size @xmath259 or @xmath258 need to be treated separately . we let @xmath287 denote the size of the @xmath79th block not of size @xmath259 or @xmath258 and @xmath288 and @xmath289 denote the number of blocks of size @xmath259 and @xmath258 respectively . we have from lemma [ lem : blockdecomp ] : _ i(l_i-1)+b_2 + 2b_3=n-2 the total number of edges in @xmath290 must then not exceed : _ i2(l_i-1)-2b_4+b_2 + 3b_3 2n-4 - b_2-b_3 - 2b_4 where @xmath291 indicates the number of blocks with more than four vertices . we know that @xmath292 and so we obtain the inequality : d_1 1+b_2+b_3 + 2b_4 1 + if we have only one block , then we either have two neighbours of @xmath49 and can apply induction to this block , as it will be a two - connected graph on @xmath62 vertices and at least @xmath293 edges . if we have at least three neighbours of @xmath49 in a block , say @xmath278 , @xmath294 and @xmath295 , then we may find a path @xmath296 . this follows from theorem [ thm : whitney ] , since we have two internally disjoint paths between @xmath278 and @xmath294 and between @xmath294 and @xmath295 . if we go along one of the paths between @xmath278 and @xmath294 until we first hit one of the two paths between @xmath294 and @xmath295 , from here we follow the path towards @xmath294 and then take the disjoint path to @xmath295 , this is then a path between @xmath278 and @xmath295 that goes via @xmath294 and does not self - intersect . in this case @xmath297 forms a chord . the final case is if we have at least two blocks and at most two neighbours of @xmath49 in a block . then we have a block with two neighbours of @xmath49 call them @xmath278 and @xmath294 and we have a third neighbour of @xmath49 , @xmath295 in some other block . let @xmath298 be the articulation point of the block containing @xmath278 and @xmath294 closest to @xmath295 . we have a path from @xmath298 to @xmath295 outside of this block since it is a connected graph . we are also able to construct a path @xmath299 since they are all in one block . concatenating these paths gives again a path @xmath296 from which we determine @xmath297 is a chord . it is also possible to construct a graph with @xmath0 vertices and @xmath264 edges that is minimally two - connected , as shown in figure [ fig : minimally2conn ] . the number of edges being @xmath22 marks some transition in the possibility of being minimal . edges and @xmath0 vertices which is minimally two - connected ] edges and @xmath0 vertices which is minimally two - connected ] in this section , we convey the connection between involutions and partition schemes for connected graphs and how the latter is used to give estimations of the coefficients in the expansions . this is used as motivation to consider whether the two - connected graph involution may have such a connection . the paper @xcite presents the notion of the partition in the sense of penrose and gives the general idea of a partition . we define a partial order of @xmath300 $ ] by bond inclusion : @xmath301 @xmath302 @xmath303 . for @xmath304 , we define the set @xmath305=\{k\vert \ , g \leq k \leq h\}$ ] the penrose construction partitions the set of connected graphs into subsets of the form @xmath306 $ ] , where @xmath307 \to \mathcal{c}[n]$ ] . many different constructions can be used to achieve an @xmath308 . penrose gave one explicit example in @xcite . a partition scheme for the set of connected graphs @xmath300 $ ] is any map @xmath309 \to \mathcal{c}[n]$ ] @xmath310 , such that : * @xmath311 and * @xmath300 $ ] is the disjoint union of the sets @xmath312 $ ] for @xmath313 $ ] . the penrose scheme is as follows : for any vertex @xmath79 of @xmath313 $ ] , we denote by @xmath314 the tree distance between the vertices @xmath79 and @xmath49 . we let @xmath315 be the predecessor of @xmath79 i.e. @xmath316 and @xmath317 . we associate to @xmath23 , the graph @xmath318 found by adding ( only once ) to @xmath23 all edges @xmath319^{(2)}$ ] such that either : * @xmath320 edges between vertices at same generation * @xmath321 and @xmath322 edges between vertices one generation away for a partition scheme @xmath308 , denote by @xmath323 \vert \ , r(\tau)=\tau \}$ ] the set of @xmath308-trees . in particular , @xmath324 is the set of penrose trees . the following proposition emphasises where the boolean partition offers advantages to providing estimations . in models where we have soft repulsion ( a positive potential ) , the mayer @xmath32-function satisfies @xmath325 . using a partition scheme , we have the bound : _ g [ n]_e e(g)f_e _ e e()\|f_e\| for any numbers @xmath326^{(2)}}$ ] , we have : @xmath48 } \prod\limits_{e \in e(g ) } f_e & = \sum\limits_{\tau \in \mathfrak{a}[n ] } \prod\limits_{e \in e(\tau)}f_e \sum\limits_{f \subset e(r(\tau ) ) \setminus e(\tau ) } \prod\limits_{e \in f } f_e \notag \\ & = \sum\limits_{\tau \in \mathfrak{a}[n]}\prod\limits_{e \in e(\tau)}f_e \prod\limits_{e \in e(r(\tau))\setminus e(\tau)}(1+f_e ) \label{eq : penroserewrite } \end{aligned}\ ] ] when we take the absolute value of the right hand side , we may use the triangle inequality and bound the second product in by @xmath49 . in the hardcore case , the second product in is zero unless @xmath327 , giving that the fixed points of this @xmath308 function also give a combinatorial interpretation of the cancellations.the alternative combinatorial interpretation of fixed points provided by penrose trees is that , considering the tree as being rooted at @xmath49 , we are required to have precisely one vertex in each generation . this necessarily gives a linear tree . we have to determine the positions of @xmath328 $ ] , which are defined uniquely by their distance from @xmath49 , which corresponds to a bijection , @xmath329 \to [ n-1]$ ] , giving the @xmath330 factor . to define the penrose involution arising from the penrose construction , we make the following definition of a penrose active edge . for a graph @xmath4 , we define the hamming distance between vertices labelled @xmath79 and @xmath115 as @xmath331 which is the length of the shortest path between @xmath79 and @xmath115 . an edge @xmath114 is called _ penrose active _ for @xmath4 if , either : * @xmath332 or * @xmath333 and @xmath334 such that @xmath335 with @xmath336 . we let @xmath337 be the greatest penrose active edge for @xmath4 in lexicographic order . the mapping : _ : g g e_,g^ & g + g & is an involution on connected graphs . we first prove that @xmath338 . the two graphs @xmath4 and @xmath339 differ only on an edge @xmath340 , where @xmath341 . throughout this proof in the case where we have equality , we assume without loss of generality that @xmath342 . for any @xmath119 , we consider the distance from the vertex labelled @xmath119 to @xmath49 in both graphs . this is defined through the shortest path from @xmath49 to @xmath119 . we indicate that for any path between @xmath49 and @xmath119 containing the edge @xmath114 we can find a path of the same or shorter length that does not contain this edge . if @xmath343 , then considering a path from @xmath49 to @xmath119 up to this edge , we realise that the shortest length the path up to this edge can be is @xmath344 , but we know that there is a shorter path to this endpoint because @xmath343 and so we can replace this initial path with a shorter path . we are left with the case @xmath345 . we know from property @xmath346 that there is some @xmath315 such that @xmath336 and @xmath347 is an edge in both graphs . therefore if the initial segment of a path includes the edge @xmath114 , then the shortest this can be is @xmath348 . if the initial segment ends at @xmath115 rather than @xmath79 then we know we have a shorter path to @xmath115 that we can replace this initial segment by . otherwise it ends at @xmath79 . we know that we have a path of length @xmath349 to @xmath315 on which we can attach the edge @xmath350 to construct a new path of the same length but not using this edge . we now have that condition @xmath351 for penrose active edges is the same in both graphs , since the graph distance is the same . we now indicate that an edge satisfies condition @xmath346 independent of the presence of @xmath114 . we realise if @xmath114 was added or removed satisfying @xmath351 then it has no effect on an edge satisfying @xmath346 , since @xmath346 depends on edges between generations . therefore , we consider that @xmath114 satisfies @xmath346 . since @xmath333 , we have an @xmath322 such that @xmath347 is an edge in both graphs and @xmath336 . this means that if we use @xmath115 to invoke applying condition @xmath346 for an edge to be penrose active , then we can invoke it in both cases by using @xmath315 . we can also go the other way and find a bernardi construction to provide an appropriate partition . the map @xmath309 \to \mathcal{c}[n]$ ] , which adds to @xmath23 all externally active edges for the given tree graph @xmath23 is the appropriate partition scheme . this is explained in the context of matroids below . in the work of bjrner and sokal @xcite , it is explained that for a matroid @xmath352 , where we give a total order to the underlying set @xmath353 , we may find a partition of the collection of subsets of @xmath354 according to the matroid structure . we introduce below some key definitions for matroids to introduce this connection , which can be found in the book of oxley @xcite and the work of faris @xcite . a matroid @xmath352 on the ground set @xmath355 is defined by a collection of independent subsets , denoted @xmath356 . these subsets must satisfy the following three axioms : 1 . @xmath357 ( non empty ) 2 . if @xmath358 and @xmath359 then @xmath360 ( downward closed ) 3 . if @xmath358 and @xmath361 and @xmath362 , then there exists @xmath363 with @xmath364 ( augmentation property ) for a graphical matroid , the ground set is @xmath5^{(2)}$ ] . we define the independent sets as forests or acyclic graphs . a maximal independent set @xmath365 is called a basis . the set of bases is denoted @xmath366 . the maximal independent sets for a graphical matroid are therefore trees . the rank of a matroid @xmath352 , @xmath367 is the cardinality of a basis element . all bases have the same cardinality and so the rank is well defined . a matroid can be defined by its set of bases , since @xmath368 if and only if @xmath369 , for some @xmath370 . given a matroid , @xmath352 , consider @xmath371 . there is a matroid @xmath372 , which is the restriction of @xmath352 to @xmath18 . it has ground set @xmath18 and @xmath373 . for @xmath371 , the rank of @xmath18 , @xmath374 is the rank of the matroid @xmath372 or alternatively the cardinality of the largest independent subset of @xmath18 . we note that @xmath375 if and only if @xmath18 is independent , so the rank function completely determines the matroid . the dual of a matroid is defined on the same ground set , but has a dual rank function rk@xmath376 , defined by : ^(a):=\|a\|-(e)+(e a ) let @xmath12 be the set of bases for @xmath25 . the dual basis set is then @xmath377 . we fix a total order on @xmath25 in the following . let @xmath378 . an element @xmath379 is _ externally active _ on @xmath277 if @xmath85 is dependent on the list of elements of @xmath277 larger than it . we let @xmath380 be the set of externally active elements . an element @xmath381 is _ internally active _ on @xmath277 , if in the dual matroid @xmath85 is externally active on the complement @xmath382 . we denote by @xmath383 the set of internally active elements . for @xmath384 , we define @xmath385=\{a \vert r \subseteq a \subseteq s\}$ ] . @xmath386 can be written as the disjoint union : 2^e=_b [ b(b),b ( b ) ] [ eq : c1 ] . for the case of the graphical matroid , we recall that the bases are the collection of trees . if we use the lexicographical order on the edges , then an edge is externally active for a tree @xmath23 in this sense , if and only if it is externally active in the sense of bernardi @xcite . this is due to the fact that all independent sets are forests and so a set of edges is dependent if it creates a cycle . we emphasise that for connected graphs , internally active edges play no role , since trees are minimally connected graphs . this therefore gives , when we intersect each set with connected graphs : [ n]=_[n ] [ , r ( ) ] , where @xmath24 has edge set @xmath387 . we note that the penrose construction does not fit in the construction given above . in figure [ fig : penrosenoorder ] , we see that we would add the dashed edge in each case . in order to do this , we can not have a consistent ordering on the edges @xmath388 , @xmath389 and @xmath390 . the motivation of emphasising this connection is to understand if a similar connection may be drawn for two - connected graphs as the important context of the result . the main conclusion is that we are able to identify combinatorially the cancellations in the alternating sums of weighted two - connected graphs . the combinatorial factor arises from increasing trees on the subset @xmath60 $ ] of the vertices , with the vertex @xmath0 adjacent to every other vertex . there are modified versions for this in the case of the polytope , where we have the isomorphic graph structures , differing only through a relabelling in the form @xmath391 for all @xmath392 $ ] . the key outlook for the work contained in this paper is to modify the set up explained in section [ sec : extend ] towards two - connected graphs so that we obtain a helpful resummation of the graphs amenable to suitable estimation , which is important for the virial expansion . the parallel that is useful to draw here is that for the cluster expansion , we have the increasing and cayley trees as the combinatorial objects representing the two cases above . it has been shown by groeneveld @xcite that these examples provide the extreme cases for positive potentials and an adaptation is available for stable potentials . * acknowledgements . * the author would like to thank the anonymous referees for constructive comments in improving this article . the work for this paper has been funded by epsrc grant ep / g056390/1 and sfb tr12 . the author acknowledges helpful discussions with d. brydges , r. koteck , d. ueltschi and s. jansen for particular discussions relating to this paper . a. bjrner , homology and shellability of matroids and geometric lattices in _ matroid applications _ , n. white ( ed . ) encyclopdia of mathematics and its applications , vol . cambridge university press , ( 1992 )	mayer s second theorem in the context of a classical gas model allows us to write the coefficients of the virial expansion of pressure in terms of weighted two - connected graphs . labelle , leroux and ducharme studied the graph weights arising from the one - dimensional hardcore gas model and noticed that the sum of these weights over all two - connected graphs with @xmath0 vertices is @xmath1 . this paper addresses the question of achieving a purely combinatorial proof of this observation .
looking at star formation across the milky way , a large variety of different conditions is encountered , in many cases leading to observed different modalities . one of the most relevant differences is found in the amount of star formation activity between the inner and the outer galaxy . the smaller content of star formation in the outer galaxy is mainly related to the lower density of atomic and molecular hydrogen compared with inner galaxy @xcite . nevertheless the outer galaxy offers the chance to study the interstellar clouds and possible ongoing processes of star formation with a lower degree of confusion , also close to the galactic plane . furthermore , reconstructing the structure of the velocity field from line observations is easier in this case , compared to the inner galaxy , thanks to the lack of distance ambiguities at any observed radial velocity . in this paper we present the first photometric study of a portion of the outer galaxy surveyed by the herschel space observatory @xcite as part of the hi - gal key - project @xcite , selecting a range of longitudes ( @xmath6 ) in the third galactic quadrant ( hereafter tgq ) , covered also by a co(1 - 0 ) line map obtained with the nanten radiotelescope . this combination of fir / sub - mm dust continuum maps at unprecedented resolution and sensitivity and the information about kinematics provided by the line observations allow us to investigate the first stages of the star formation in this portion of the milky way , by estimating and discussing in statistical sense the physical properties of the compact sources ( cores and clumps ) detected in the herschel maps . with this analysis we aim at shedding more light on the star formation in the tgq , to which a relatively little amount of literature has been devoted so far . the first extensive surveys of the the tgq were carried out by @xcite , at an angular resolution of 0.5@xmath7 and @xmath8 , respectively . they confirmed the warping of the disk at these longitudes @xcite and the presence of weak emission from molecular cloud located at galactocentric distances larger than @xmath9 kpc . the latter survey was then drawn on by @xcite to derive the properties of the clouds located beyond 2 kpc from the sun . a relevant finding of this analysis is that no grand design spiral pattern is seen in the distribution of the clouds on the galactic disk . the warp of the galaxy in the tgq has been observed also by @xcite and @xcite through the distribution of the iras sources and of the visual extinction maps derived from the digitized sky survey i , respectively . to assess the spiral structure of the galaxy in the tgq other techniques have been also adopted . @xcite and @xcite have studied the distribution of young star clusters , tracing the local , the perseus and the norma - cygnus ( outer ) arms . this latter is not detected at these longitudes by @xcite based on the @xcite co survey ( which is taken , for most of the tgq , from that of * ? ? ? * ) . instead , the @xcite results were confirmed by @xcite , who attribute to the outer arm a grand design feature extending from @xmath10 to @xmath11 ( reaching heliocentric distances @xmath12 and 12 kpc at the two extreme longitudes , respectively ) , expanding the sample of clusters of @xcite and combining this information with kinematics extracted from new co observations ( still unpublished at present ) . focusing on the vicinity of the region analyzed in this work , the most important star - forming region is cma r1 , dominated by the presence of the cma ob1 association @xcite located at @xmath13 pc , i.e. in the local arm , and with an age of @xmath14 yr @xcite and of the sh2 - 296 ( s296 in the simplified notation we adopt hereafter ) hii region @xcite . the population of young stars in cma r1 has been studied by @xcite based on x - ray data , and its typical age has been estimated as older than @xmath15 yr . however , only a very small fraction of the investigated proto - stars lie in the area surveyed in the present paper . this area was fully covered , instead , by the survey of molecular clouds in the monoceros and canis major regions carried out with the nanten telescope in the @xmath16co(1 - 0 ) line @xcite , spanning an area of 560 square degrees . the two most massive giant molecular clouds found are the aforementioned cma ob1 and g220.8 - 1.7 @xcite , this latter located at @xmath17 pc . these authors also studied the relation between the gas emission morphology and the star formation signatures have been observed in this region , as several iras young stellar object candidates and hii regions @xcite . the low level of extinction ( @xmath18 ) found by @xcite in the entire region , however , seems to exclude the possibility of encountering massive star formation ( msf hereafter , cf , e.g. , * ? ? ? the western part of our survey corresponds , instead , to the north - eastern corner of g216 - 2.5 ( or maddalena s cloud , @xmath19 pc ) , mapped in both co(1 - 0 ) and @xmath16co(1 - 0 ) by @xcite . these authors claim the non - star - forming nature for this cloud , partially controverted by more recent spitzer observations of @xcite , while recognize the star forming character of a nearby filamentary cloud associated with the s287 hii region , fully covered by our observations . in this framework , the new hi - gal observations in the crucial range @xmath20 m will permit to obtain a more exhaustive description of the star forming activities or capabilities in this region . in particular , both the early stages of the proto - stellar core evolution and the properties of their pre - stellar progenitors can be studied , being in both cases the peak of the cold dust emission lying in this range of wavelengths ( e.g. , * ? ? ? * ; * ? ? ? * ) . thanks to nanten co(1 - 0 ) data , we are also able to estimate kinematic distances for those sources , needed to derive masses and luminosities and then to depict a possible evolutionary scenario for the region . this allows us to carry out a complete study of cold dust and gas properties at unprecedented resolution and sensitivity in a significant portion of the outer galaxy . this analysis is structured and presented as follows . in section [ obs ] the details about herschel and nanten observations are reported , together with data reduction strategy . in section [ overall ] , and also in appendix [ allmaps ] , the obtained maps are presented , and used to provide an overall description of the column density and temperature distributions in the surveyed area . starting from section [ cores ] the discussion revolves about the compact sources that are extracted from the hi - gal maps and from the ancillary co(1 - 0 ) observations through cloud decomposition techniques . in section [ seds ] the main properties of the hi - gal compact sources are derived from photometric data by means of a modified black body fit and then discussed in the framework of the differences between pre- and proto - stellar sources . in section [ evol ] , in particular , these quantities are exploited to situate the hi - gal sources into an evolutionary picture . finally , in section [ concl ] a summary of the obtained results is reported . the _ herschel infrared galactic plane survey _ ( hi - gal , * ? ? ? * ) is a herschel open - time key - project that initially aimed at covering the inner galaxy ( @xmath21 ) , and was subsequently extended to the entire galactic disc . in this paper we publish the first four `` tiles '' observed in the tgq as part of this survey , spanning the range of galactic longitudes @xmath0 and the range of latitudes @xmath1 , because in this portion of the galactic plane the hi - gal observations do not straddle the latitude @xmath22 , but follow the warped midplane . the @xmath23 fields presented in this paper were observed between may 5 and 9 , 2011 . as in the case of all the hi - gal survey , the maps were obtained with the herschel fast scan parallel mode , i.e. , using simultaneously pacs at 70 and 160 @xmath2 m @xcite and spire at 250 , 350 , and 500 @xmath2 m @xcite at the scan speed of 60@xmath24s , and performing two scan maps of each field , one with nominal coverage and the other with orthogonal coverage , in order to increase data redundancy . due to the 21@xmath25 separation of the pacs and spire focal plane footprints on the sky , the areas covered by these two instruments result slightly different , namely shifted each other of this angular quantity . the diffraction limits at the five bands are 5.0 , 11.4 , 17.6 , 23.9 , and 35.1 , respectively . however , due to the pacs on - board data averaging and the use of the fast speed mode a degradation of the pacs psf is produced along the scan direction , up to 12.2 and 15.7at 70 and 160 @xmath2 m , respectively ( pacs manual , available from the esa herschel science centre ) . the hi - gal observation name convention consists in naming the tiles using the approximated value of the central longitude , then in the following we will designate these four tiles as @xmath26 , would be closer to the center of the map , but we use the conventional name already reported in the herschel data archive . ] @xmath27 , @xmath28 , @xmath29 . hereafter , we will call @xmath30 the entire region composed by these four tiles and analyzed in this paper . the data reduction strategy followed the hi - gal pipeline described in @xcite . briefly , we used scripts written in the herschel interactive processing environment ( hipe , * ? ? ? * ) to reduce archival data to the level 1 . the obtained time - ordered data ( tods ) of each bolometer were then processed further by means of dedicated idl routines to remove spurious instrumental effects and glitches due to cosmic rays , and maps were obtained using the fortran code romagal based on generalized least square ( gls : e.g. , * ? ? ? * ) approach , able to reduce the influence of the @xmath31 noise in the tods . since the gls technique is known to introduce artifacts in the maps , namely crosses and stripes in correspondence of the brightest sources , a weighted post - processing of the gls maps ( wgls , * ? ? ? * ) has been applied to finally obtain images in which artifacts are removed or heavily attenuated . the pixel sizes of the maps are 3.2 , 4.5 , 6.0 , 8.0 , and 11.5at 70 , 160 , 250 , 350 , 500 @xmath2 m , respectively . the astrometry of the maps has been checked comparing the positions of several compact sources in the 70 @xmath2 m maps with those of their possible wise @xcite counterparts at 22 @xmath2 m , and the calculated rigid offset applied for correction . finally , the absolute calibration has been definitively performed by applying a linear transform of the maps adopting the coefficients determined by comparing herschel with iras and planck , following @xcite . all the single tile images obtained at the five hi - gal wavelengths are reported in appendix [ allmaps ] . in figure [ rgbmosaic ] , instead , we show a composite rgb image of the region , assembled combining the mosaics of the four tiles at 160 , 250 and 500 @xmath2 m . the 70 @xmath2 m observations have not been used because the diffuse emission they contain appears dominated by the noise and only the brightest regions are well recognizable . the mosaics have been produced in the same way as the single tiles , combining together the tods of the four tiles to obtain an overall map . the common science area , i.e. the area covered by pacs and spire , highlighted with a cyan contour in figure [ rgbmosaic ] , amounts to @xmath32 deg@xmath33 . the nanten co(j=1 - 0 ) galactic plane survey data set @xcite was used in the present analysis . at the time of the observations , carried out in a period from 1996 to 2003 , 4-m nanten millimeter - wave telescope of nagoya university was installed at las camapans observatory in chile . the half - power beamwidth was @xmath34 . the data were gathered using the position switching mode . the pointing accuracy was better than @xmath35 . the observations covered a region of @xmath36 ( the coordinates of the bottom - left and top - right pixels are @xmath37=[225.60^{\circ},-2.53^{\circ}]$ ] and @xmath37=[216.53^{\circ},0^{\circ}]$ ] , respectively ) , and the observed grid consists of @xmath38 points located every @xmath39 . the spectrometer used was acousto - optical with 2048 channels providing a frequency resolution of 250 khz . the spectral intensities were calibrated by employing the standard room - temperature chopper wheel technique @xcite . an absolute intensity calibration was achieved by taking the absolute peak antenna temperature , @xmath40 , of orion kl ( r.a . = @xmath41 , dec . = @xmath42 for the equinox of 1950.0 ) to be 65 k. the rms noise fluctuations were typically 0.2 k with about 5 seconds integration for an on - position . a first order degree polynomial was subtracted from the line spectra to account for instrumental baseline effects . finally , antenna temperatures were divided by the main beam efficiency factor @xmath43 @xcite to obtain main beam temperatures . the integrated intensity maps obtained from these line observations are presented in section [ comaps_par ] . a first look at the spectra in the reduced co(1 - 0 ) data cube reveals that the structure of the velocity field is relatively clear due to the small number of spectral components detected and to the infrequent presence of more than one along the same line of sight . a global view of the velocity field is provided in figure [ totspec ] in which we show the sum of all the spectra composing the cube , where at least three main components are found , peaking at the 17 , 28 , and 40 km s@xmath45 velocity channels , respectively . a secondary peak of the second component also appears around 33 km s@xmath45 . finally , a further weak component , the most red - shifted , is present around 54 km s@xmath45 . these indications can be exploited to determine opportune velocity ranges @xmath46 $ ] to build meaningful channel maps , rather than blindly choosing equally spaced ranges of channels . the maps have been obtained by integrating the line emission ( main beam temperature ) in those lines of sight where a feature with a signal - to - noise better than 3 has been found ( the average baseline rms noise of the data cube is @xmath47 k : @xmath48 in figure [ comaps ] we show both the map of the @xmath44co(1 - 0 ) intensity integrated over the whole range from -0.5 to 60.5 km k s@xmath45 ( panel @xmath49 ) , and the set of the maps corresponding to the ranges highlighted in figure [ totspec ] ( panels @xmath50-@xmath51 ) . lcccccccc i & 17 & 29.5 & 51.5 & 223 & 1.1 & 69.2 & 3.2 & 1.1 + ii & 28 & 51.5 & 66.5 & 218 & 2.2 & 26.2 & 3.3 & 1.9 + iii & 40 & 66.5 & 75.5 & 222 & 3.3 & 4.3 & 3.5 & 0.7 + iv & 54 & 75.5 & 94.5 & 217 & 5.8 & 0.4 & 4.0 & 0.2 + the first and the second velocity components ( sorted from the bluest to the reddest ) are evidently predominant in the total budget of the emission . to quantify this , in table [ co_stats ] the contribution of each component to the integrated intensity map in panel @xmath49 is quoted , together with the average distance calculated at a characteristic longitude for each component . a gross estimate of the masses of these four components can be obtained , given the information on the global distance . first , the local column density map is derived directly from the maps in figure [ comaps ] through the empirical relation @xmath52 where the conversion coefficient value adopted in this case is calculated as a function of the galactocentric radius @xmath53 through the relation of @xcite : @xmath54=1.4 \times 10^{20}\exp(r/\;11 \mathrm{kpc})\;.\ ] ] then , for the each component the mass is obtained integrating the column density over the area occupied in the sky : @xmath55 , where @xmath56 indicates the pixel scale of the co(1 - 0 ) map ( in radiants ) . nevertheless , the value of the @xmath57 conversion coefficient is matter of debate and strong variations of the obtained mass values are expected if another estimate of @xmath57 is adopted . for instance , @xcite make a twofold analysis using both the approach described by the equation [ nh2co ] and the constant value @xmath58 of @xcite . in our case , this latter value would lead to underestimate of the masses table [ co_stats ] of a factor from 1.8 to 2.2 , depending on the distance . the overall distribution of the column density derived from these maps is presented in section [ nh2_sect ] , while the correspondence with hi - gal maps and with previous co(1 - 0 ) surveys is discussed in section [ conh2 ] . here we just highlight the low degree of complexity of the velocity field , as it emerges from these line observations . in figure [ vfield ] the co(1 - 0 ) contours of figure [ comaps ] , @xmath59 , are overlapped on the spire 250 @xmath2 m mosaic . the degree of overlap among components appears relatively low : the eastern part of the region is dominated by the first ( bluest ) component , the western one is dominated by the second component , whereas the third and fourth component are basically associated to few small and bright features of the spire map . both the herschel and nanten observations give the possibility of deriving maps of the column density in the region . possible differences between the resulting maps are expected , due to the assumptions made on the considered component of the ism . therefore , any comparison requires care and must be accompanied with some caveats . we obtained the maps of column density and temperature for @xmath30 from the mosaics of the four hi - gal tiles , using the wavelengths from 160 to 500 @xmath2 m , being the emission at 70 @xmath2 m generally contaminated by warmer dust components as circumstellar matter and hii regions ( cf . this implies , of course , that these maps basically represent the properties of the cold component of the dust ( large grains ) . from the practical point of view , considering only the common science area , all the maps have been rebinned onto the grid of the 500 @xmath2 m one , and a pixel - to - pixel grey - body fit has been performed , according to the usual expression @xmath60 where @xmath61 represents the spectral energy distribution ( sed ) of the pixel , @xmath62 is the map pixel scale at 500 @xmath2 m in sr , @xmath63 cm@xmath33 g@xmath45 at @xmath65 ghz ( @xmath66 m ) already accounting for a gast - to - dust ratio of 100 @xcite , @xmath67 is the atomic hydrogen mass , and @xmath2 the mean molecular weight , assumed to be equal to 2.8 to take into account a relative helium abundance of 25% in mass . we imposed @xmath68 in order to reduce the number of free parameters in the fit ( see also * ? ? ? the obtained column density and temperature maps are shown in figures [ hiiregions ] and [ htemp ] . combining the information contained in these maps and in that of figure [ vfield ] with the global impression provided by figure [ rgbmosaic ] , we are now able to provide an overall description of the @xmath30 region . the eastmost tile , namely @xmath29 , is by far the richest in bright features and , apparently , in star formation activity . all these regions belong to the cma ob1 giant molecular cloud , located at @xmath13 pc ( * ? ? ? * and references therein ) , and are encircled by the hii region s296 @xcite , as illustrated in figure [ hiiregions ] . a large cavity centered at about @xmath37=[224.8^{\circ},-0.9^{\circ}]$ ] is well recognizable , with a filamentary ridge that turns out to be particularly massive and cold ( @xmath69 k ) in the western part . another bright filament aligned with the east - west direction seems to connect , in perspective , the center with the ridge of the cavity . this corresponds to the massive ( @xmath70 ) cloud 4 of the paper of @xcite ( [ kky2004]4 according to the simbad nomenclature ) . the southern part of @xmath29 is dominated by another extended , bright and warmer region corresponding to s292 and [ kky2004]3 ; this latter is the most massive cloud ( @xmath71 ) of those identified by @xcite in the area considered in this paper . cold filaments coexist with warm ( @xmath72 k ) diffuse emission or star forming sites , as s292 . the young cluster ngc2335 , located north of this region ( @xmath37=[223.62^{\circ},-1.26^{\circ}]$ ] ) and with distance and age estimated in 1.79 pc and 79 myr , respectively @xcite is not detected at any of the herschel wavelengths . in the @xmath28 tile another small cloud is found , [ kky2004]1 , still belonging to cma ob1 . in the southern part , the bfs64 hii region @xcite is associated with the velocity component iii . the emission in @xmath27 , again , is mostly associated with the velocity component i. two hii regions , bfs61 and bfs62 , represent the signature of recent high - mass star formation activity , albeit much less dramatic than in @xmath29 . in @xmath26 a filamentary cloud already mentioned in section [ intro ] is prominent in the northern part . it belongs to the velocity component ii and is associated with the s287 hii region and other minor ones , and with four molecular clouds of @xcite . the angular size of the whole filament is approximately 1.6@xmath7 , corresponding to an extent of 59 pc at a distance of @xmath73 kpc . it is considered an active site of star formation by @xcite , who justify its filamentary shape with the action of stellar winds of the o - b stars driving the hii regions . a peak of dust temperature at @xmath74 k is found right by the position of the bfs57 hii region . the l1649 cloud @xcite , indicated as [ kky2004]113 , is well visible below the filament , but physically unrelated with it , being instead the only prominent feature of the velocity component i in the @xmath26 field . finally , two dense and relatively compact regions are seen in the south . the first one corresponds to s286 and belongs to the velocity component iv ( @xmath75 kpc , @xmath76 kpc ) , representing the farthest unconfused region detected in this portion of the hi - gal survey ; in the dust temperature map it appears warmer ( @xmath77 k ) than the portion of sky it is projected on . the second one ( also iras06522 - 0350 ) is quoted as a satellite star - forming cloud @xcite of the g216 - 2.5 cloud @xcite . as a global remark , we notice that , invoking the theoretical result of @xcite about a column density threshold for having massive star formation ( msf hereafter ) , namely @xmath78 g @xmath79 which assuming a mean molecular mass of 2.8 corresponds to @xmath80 @xmath79 , there are no pixels exceeding this value in our column density map . however , this result can not be considered resolutive against the presence of msf in @xmath30 , for at least three main reasons : @xmath81 ) mainly for distant components , possible column density peaks might be diluted in the solid angle subtended by a pixel , because of the original instrumental angular resolution and/or the subsequent regridding of the maps for calculating the column density ; @xmath82 ) these column densities and temperatures are derived using the total emission measured along the whole line of sight ; however , as it is described below in the paper , if the compact source emission is separated from the background , lower temperatures and higher masses are found for those cores , making it possible to find higher values of the column density ; @xmath83 ) as shown by @xcite , and also suggested by hi - gal observations @xcite , the @xcite critical value probably constitutes a too severe threshold for msf ; this point will be further discussed in section [ higalprops ] . the availability of both the co(1 - 0 ) and herschel column density estimates allow us to compare the performance of an optically thick spectral line and of the dust as density tracers of the clouds , and to check the validity of the empirical relation [ nh2co ] in this portion of the galactic plane . we consider the herschel - based map ( figure [ hiiregions ] ) the most reliable of the two , being the cold dust a truly optically thin tracer of the ism density . instead , the empirical co - based method may fail in single cases , mainly in the densest parts of the clouds , as recently further demonstrated by @xcite by means of radiative transfer models . also , it is sensitive to the excitation conditions of the gas . a systematic characterization of these factors affecting the co - derived column densities will be reported in the forthcoming paper of @xcite . in figure [ histonh2 ] , panel @xmath49 , the histogram of this map is reported . it is characterized by a peak well described by a log - normal behavior ( the log - normal fit is also shown ) , followed by a power - law - like extension at higher column densities . this situation is commonly found in regions forming low mass @xcite and high mass @xcite stars , being the log - normal shape attributed to turbulence , and the deviation at large densities due to strongly self - gravitating systems , a typical signature of star formation activity ( e.g. , * ? ? ? * ; * ? ? ? a more quantitative analysis of the power - law tail of this requires separating the different components along the line of sight . however , this is beyond the scope of this present paper and will be presented in another study focusing on the probability density functions . the co - based column density has been derived for the four distance components through the equation [ nh2co ] , adopting the corresponding @xmath57 factor values reported in table [ co_stats ] ; subsequently , these four maps have been co - added to obtain the total column density map . to better compare the two maps , we considered only the portion corresponding to the common coverage area ( figure [ histonh2 ] , panel @xmath50 , green dotted line ) , and then reprojected the herschel - based map onto the grid of the co(1 - 0 ) map , considering only the pixels with meaningful values in both maps ( zeroes can be found , especially in the co - based map ) . the histograms obtained in this way ( figure [ histonh2 ] , panel @xmath50 , solid lines ) become directly comparable , being derived from the same amounts of pixels . it is possible to appreciate that the histogram of the co - based column density appears quite different from those both the original herschel - based map and its regridded version . in the first case , the largest values co - based density values are more than half order of magnitude lower than those revealed in the herschel - based map . this can be due to the combination of different concomitant effects . first , the co(1 - 0 ) is not ideal tracer of the high - density gas distribution . this mainly affects the histogram shape , as a lack of the tail that , for instance , we observe in the herschel - based one ( e.g. , * ? ? ? second , the the significant pixel size difference between the spire and nanten beams ( @xmath84 against @xmath39 ) produces in the latter case a much higher degree of dilution of the local bright peaks in the final pixels , removing from the histogram the high - density tail that is seen instead in the former . this effect can be assessed by considering the histogram of the reprojected herschel - based map : in this case the two histograms have a more similar peak position and ranges of variability . however , the regridding of the herschel - based column density map produces a strong attenuation of both the low- and the high - density tail of its histogram , where now the co - based one appears to be more populated . this can be seen also in the plot of the herschel - based column density ( reprojected ) vs the co - based one , built pixel - to - pixel and shown figure [ histonh2 ] , panel @xmath85 . in the majority of cases ( @xmath86 ) the former is smaller than the latter and , the ratio of the sums of the involved pixels is @xmath87 . this is just a global indication that can not be simply taken as a sort of further correction of the adopted @xmath57 factors , since the relation between the two distributions does not appear linear . the @xmath57 factor is also known to depend strongly on local variations of the physical conditions , because it is expected to be lower in zones of star formation activity @xcite . furthermore , the co line intensity depends on the temperature of the gas . however , we can at least evaluate how the choice of the @xmath57 factor might affect the conclusions achievable by the comparison of the two column density maps . adopting @xmath88 ( * ? ? ? * see section [ comaps_par ] ) , the original co map histogram is shifted towards lower densities , and its shape gets sligthly modified because of different coefficients initially adopted for the components i - iv ( magenta dashed histogram in panel @xmath50 ) . in this case , the distribution peaks at a value smaller than that of the herschel - based one . on the other hand , a different choice of the @xmath89 constant in the equation [ greybody ] would produce a rigid shift of the herschel - based column density histogram . for example , the constant of @xcite ( @xmath90 cm@xmath33 g@xmath45 at 230 ghz ) can be translated , for @xmath68 as in our case , to 0.14 cm@xmath33 g@xmath45 at 1200 ghz , producing @xmath91 lower column densities . these possible corrections are represented in panel @xmath85 as arrows , showing that the choice of a smaller @xmath57 coefficient would generally produce a better agreement between the two distributions , while a larger dust opacity would obtain the opposite effect . in this section we start to focus on the distribution and the properties of the compact sources detected in the hi - gal observations of @xmath30 . the distance estimate for these sources is performed exploiting the kinematic distances of co clumps . the compact source detection and photometry have been performed at each of the five available bands following the procedure described in previous hi - gal articles @xcite . we used the curvature threshold extractor package ( cutex , * ? ? ? * ) , which detects the sources as local maxima in the second - derivative images and then fits an elliptical gaussian to the source brightness profile to estimate the integrated flux of the source . the parameters of the best fit ( the peak position and strength , the minimum and maximum fwhm , and the integral of the gaussian ) are used to estimate the source position , the peak flux , the angular size and the total flux . the flux uncertainties have been derived from the best fit as well . using this approach , five independent lists of photometry have been obtained at the five hi - gal wavebands . the completeness limits have been estimated evaluating the cutex ability of recovering synthetic compact sources randomly dispersed on the original maps . both the size and the flux of the generated sources mirror the distributions contained in the original cutex lists . the completeness limit at each waveband is then estimated as the flux value at which the histogram of the fraction of the recovered sources exceeds the threshold of 90% of the total number of injected synthetic sources having fluxes in that bin . the procedure is illustrated in figure [ completeness ] , and the obtained values are 1.75 , 1.06 , 0.60 , 0.53 , and 0.56 jy at 70 , 160 , 250 , 350 , 500 @xmath2 m , respectively . subsequently the sources extracted in each waveband have been associated by means of a simple spatial criterium @xcite in order to obtain a band - merged catalog . the total number of entries in this catalog is 3476 , composed by sources having at least a detection in one band . in section [ seds ] the further selection of a sub - sample of reliable seds eligible for the fit through a model is reported . similarly to the source extraction performed on the herschel maps , a search for spatially and kinematically coherent overdense structures has been carried out in the co(1 - 0 ) @xmath92 cube , using cloud decomposition techniques . doing this we do not aim at finding exactly the counterparts of the herschel sources and at establishing a some biunivocal correspondence between the two populations . rather , due to the coarser spatial resolution of the co(1 - 0 ) observations and to the different source extraction techniques adopted , the co(1 - 0 ) coherent sources ( hereafter `` co clumps '' , although in some cases they could not match exactly the definition of clump having typical cloud sizes ) are expected to be generally more extended than the herschel ones , and thus to be found in association with more than one of them in general . for the cloud decomposition we used the cloudprops package @xcite , which is also able to derive a set of the basic morphological and physical properties of found the clumps . we tested both the native clump finding algorithm present in the package , and the widely used cloud decomposition algorithm clumpfind @xcite , that is also implemented in the package as alternative option . although clumpfind is known to be strongly depending on the initial parameter setup , as discussed in detail , e.g. , by @xcite , @xcite , @xcite , and @xcite , in this case it turned out to be more able than the native cloudprops search engine to detect structures with the same efficiency across the whole map , probably due to different rms noise conditions throughout the data cube . , see * ? ? ? * ) ] the list of the detected clumps is reported in table [ coclumps ] in appendix [ clumplist ] , complemented with the following quantities provided by cloudprops and calculated as described in @xcite , with some further tuning of the default settings : * _ velocity dispersion_. it is derived as the second moment of the intensity distribution along the velocity dimension . * _ distance_. to derive the heliocentric distances , the galactic rotation curve of @xcite ( @xmath93 kpc , @xmath94 km s@xmath45 ) has been used instead of the default one of @xcite implemented in cloudprops . * _ radius_. the deconvolved radius @xmath95 is estimated as the geometric mean of the deconvolved major and minor axes of the clump , given in units of pc once the clump distance is known . * _ mass_. the mass of the clump @xmath96 is obtained as in section [ comaps_par ] , i.e. integrating the column density over the solid angle of the clump . the column density is derived by cloudprops as in equation [ nh2co ] with a @xmath57 factor constant for all clumps . we then rescaled the mass considering a @xmath57 factor varying with the galactocentric distance as in equation [ xco ] . * _ virial mass_. the virial mass is calculated as @xmath97 = 1040 \sigma^2_v[\mathrm{km~s^{-1}}]~r[\mathrm{pc } ] $ ] , under the assumption that each clump is spherical and virialized with a density profile of the form @xmath98 @xcite . density power laws with exponent 0 and -2 would require , instead , a constant of 1150 and 690 m@xmath99 ( km s@xmath45)@xmath45 pc@xmath45 , respectively . * _ luminosity_. the clump luminosity in the co(1 - 0 ) line is provided in units of k km s@xmath45 pc@xmath33 . the conversion factor to solar luminosities for this transition is @xmath100^{-1}$ ] @xcite . after ruling oututput sources having meaningless physical parameters ( as not - a - number masses , luminosities , etc . ) the final list is composed by 321 clumps . in this section we derive the basic physical properties of the hi - gal compact sources from their spectral energy distributions ( seds ) . however , first of all a selection has to be performed in order to obtain a reliable sample of seds eligible for the modified black body fit . where present , the flux at 70 @xmath2 m is not considered in this fit , being generally affected by the direct contribution of a proto - stellar component ; for this reason , its presence can be used as a signature to distinguish a proto - stellar from a starless core / clump ( e.g. * ? ? ? * ; * ? ? ? we searched also for possible wise 22 @xmath2 m counterparts ( detected at @xmath101 ) within a searching radius of 12.6(corresponding to the wise resolution at that wavelength and to the wise and herschel positional uncertainties combined in quadrature ) around the hi - gal source coordinates . the low signal - to - noise ratio in our pacs 70 @xmath2 m maps , responsible of the large flux completeness limit at this band , can result in an association between a wise source and a herschel one lacking emission at 70 @xmath2 m . in these cases , the association is considered genuine and the source is classified as proto - stellar if at least the flux at 160 @xmath2 m is available . in this way , 236 out of 943 hi - gal sources with a selected sed ( see below ) are found to be associated with a wise counterpart . to estimate which fraction of the wise - herschel associations could be spurious , we generated a random distribution of coordinate pairs having the same number and spatial coverage of the wise sources in our field , and run the same procedure of matching with the hi - gal sources . only 18 of the aforementioned 943 hi - gal sources are associated with a source generated in such a way , i.e. less than 8% if compared with the number of the wise counterparts found . this percentage represents the occurrence of spurious matches that should be expected to affect our wise - herschel association . at the longer wavelengths , we deal with the frequent case of the increase of the beam - deconvolved diameter , as estimated by cutex ( @xmath102 ) , at increasing wavelengths , already discussed by @xcite and @xcite . due to this increase of the area over which the emission is integrated , many seds flatten or even rise longwards @xmath103 m . here we adopt the strategy of @xcite , consisting in assuming a reference wavelength to estimate the source angular size , and scaling the fluxes at a larger wavelength @xmath104 by the ratio between the deconvolved sizes at the reference wavelength and at @xmath104 , respectively . due to the further selection to identify seds eligible for the fit , described at the end of this section , the shortest wavelength always available is 250 @xmath2 m , that we adopt as reference wavelength @xmath105 this is based on the assumptions that @xmath81 ) the source is optically thin at @xmath106 m ; @xmath82 ) the temperature gradient is weak @xcite ; @xmath83 ) the radial density profile of the source is described by @xmath107 , then @xmath108 . it is noteworthy that scaling fluxes in this way corresponds to consider fluxes emitted by the same volume of dust , that is more compatible with fitting a modified black body to a sed . on the other hand , hereafter it must be kept in mind that the physical properties derived from the fit of the scaled seds describe the average conditions of the source inside a volume delimited by the size detected at the reference wavelength . finally , considering a smaller volume , we reduce the influence of possible multiplicities ( i.e. shorter wavelength counterparts that appear confused in an unresolved source at 350 - 500 @xmath2 m ) on the fluxes composing the final sed . a further selection has been performed on the seds built as described above , accepting only those : @xmath81 ) being composed by consecutive fluxes ( at least three ) ; @xmath82 ) showing no dips ( e.g. concave shape ) ; @xmath83 ) not peaking at 500 @xmath2 m ; @xmath109 ) having a distance estimate , as described in section [ hdistance ] . after applying these constraints , we finally obtained a sample of 943 seds eligible for the fit , 255 of which are proto - stellar ( 184 of which with detected at both 22 @xmath2 m and 70 @xmath2 m , 52 only at 22 @xmath2 m , and 19 only at 70 @xmath2 m ) , and 688 starless . the positions of these sources in the surveyed region are reported in figure [ sourcepos ] , where we introduce the color convention blue for proto - stellar objects , and red for the starless ones . we remark that some bright sources might have been ruled out from the present statistical analysis by the constraints we applied to deal with regular seds . furthermore we draw attention on the fact that many sources appear to lie on the most prominent filaments present in the maps . although these filaments appear uniform in figures like those shown in appendix [ allmaps ] due to the choice of the intensity scale , in actual fact they possess a complex internal structure revealed by herschel and characterized by the presence of numerous compact sources ( see , e.g. , * ? ? ? this aspect is well exemplified by the figure [ extract250 ] , where the positions of the sources extracted at 250 @xmath2 m in the neighborhoods of the [ kky2004]4 cloud are superimposed on the map of this wavelength rendered by means of two different intensity scales . the first one is appropriate to show the appearance of the diffuse emission , while the second one is needed to reveal the internal structure of the bright filaments , where it is possible to find compact sources whose genuineness is confimred , for example , by the comparison with the 70 @xmath2 m map . * the occurrence of false positives in our extraction can not be excluded a priori . however , since the detection is performed independently at each waveband , it is unlikely to systematically find counterparts of them at the other wavelengths . in any case the source selection described in this section , based on the regularity of the collected sed , ensures the reliability of the sources considered for the subsequent analysis . * the problem of the distance estimation of the hi - gal sources is a crucial point that has to be faced to derive a complete and meaningful set of physical properties for these objects . fortunately in the portion of the galactic plane hosting the @xmath30 field this issue turns out to be less problematic , thanks to both the low degree of confusion along the line of sight and the absence of the near / far ambiguity on the kinematic distance , as pointed out in section [ comaps_par ] . a possible approach might consist in examining the line of sight in the co(1 - 0 ) data cube corresponding to an hi - gal source , and to take the @xmath110 of the brightest component encountered , as done for example in @xcite . here however we want to exploit the additional information offered by the co(1 - 0 ) data cube decomposition in clumps ( section [ cdecomp ] ) : in the co(1 - 0 ) line of sight containing the source , we consider the spectral channels assigned by cloudprops / clumpfind to one or more clumps , and associate to the source the distance of the clump , if any , having the closest centroid . there is a double advantage in adopting this strategy : first , the distance assigned to all the sources associated to a given clump is the same , without finding a different distance at each spatial / spectral position in the co(1 - 0 ) cube , perhaps affected by the clump internal kinematics unrelated with galactic rotation . second , this strategy acts quite efficiently in cases of presence , along the same line of sight , of a peripheral co(1 - 0 ) feature of a large and bright clump and of a weaker but central feature of a smaller clump : in this case we believe that the herschel source is more likely associated with the small clump being centered on it , rather than with the outer part of a large clump . taking into account only the brightest line in this example , instead , would have lead to a completely different association . the distance information can be exploited to generate a view of the spatial distribution of the sources in the galactic plane ( figure [ galpos ] , upper panel ) . in this figure we notice a certain degree of continuity in the distribution of the heliocentric distances , without significant gaps between the components i and ii . to better illustrate how this classification works , the isotachs corresponding to the velocities conventionally separating the four components are also plotted . according to the spiral arm loci from @xcite kpc ) from the one we adopt here . ] , in the longitude range we consider in this paper the local arm should be encountered at @xmath111 kpc , the perseus arm at @xmath112 kpc , and the outer arm at @xmath113 kpc , respectively . our results , in the limit of the reliability of the adopted rotation curve in this part of the galaxy , partially confirm this view . looking at the lower panel of figure [ galpos ] , the highest concentration of component i sources is found in the bins centered at 0.8 and 1 kpc . instead , the distance distribution of the component ii sources peaks around 2 , 2.6 , and 3.2 kpc , suggesting instead the presence of an inter - arm bridge of matter similar to those found by @xcite in other locations of the tgq . the 0.7 - 1.1 , 2.0 , 2.6 and and 3.2 kpc distances also correspond with the global distances ( photometric or kinematic ) assigned by @xcite to their clouds present in our field . the outer arm is not detected in this portion of the tgq , in agreement with @xcite and @xcite . finally , we notice that none of the sources we found at the farthest galactocentric distances can be classified as belonging to the far outer galaxy ( @xmath114 kpc * ? ? ? as said above , a modified black body ( described by the equation [ greybody ] , but with the terms @xmath115 replaced by @xmath116 ) has been fitted to the seds , to obtain concomitant estimates of the mass and the temperature for each individual source . in table [ higaltab ] the astrometric , photometric and physical properties of the analyzed sources are resumed . here we start discussing the physical diameter and the temperature ( figure [ sizetemp ] ) . for sources provided with a distance estimate , the diameter @xmath117 is easily obtained from the source deconvolved fwhm estimated at 250 @xmath2 m . the histogram in panel @xmath49 shows that our sample is composed by a mixture of cores ( @xmath118 pc ) and clumps ( @xmath119 pc ) , according to the typical definitions of these object classes based on their size ( e.g. , * ? ? ? it is important , therefore , to keep in mind that a relevant fraction of the compact sources analyzed in this work do not correspond to progenitors of single star sytems but , rather , to distant complex and large structures we can not resolve ( in particular , 153 out of 255 proto - stellar sources must be classified as clumps ) . the two size distributions do not appear much different each other ( @xmath120 in both cases ) , unlike in @xcite , where proto - stellar sources were found to be on average smaller that the starless ones . to better discuss this point , we notice that sources spread on a wide range of distances are used all together to build the histograms for both the samples , so that some large proto - stellar clumps located at long distances are obviously larger than closer pre - stellar cores . for this reason we show also the distributions of the diameters as they would appear if placed at a distance of 1000 pc ( panel @xmath50 ) . also in this case , there is no clear indication of significant size differences . in fact , @xcite adopted a more flexible criterium to decide the reference wavelength at which the source size is estimated : 160 @xmath2 m when a detection at this wavelength were available , and 250 @xmath2 m otherwise . since many starless sources are not detected at 160 @xmath2 m , their size was estimated at 250 @xmath2 m , where it appears generally larger ( section [ sedbuild ] ) , thus favoring the observed segregation in size , while in our case the reference wavelength is the same in all cases . concerning the temperature distributions of the two classes , also in this case the dichotomy proto - stellar - warmer against starless - colder is not evident as in @xcite , again due to different reference size adopted in that case implying a more consistent flux scaling at the spire wavelengths for the proto - stellar sources , finally leading to higher temperature and lower mass estimates . in our case , the uniform choice of the size at 250 @xmath2 m as reference reduces the differences of the two distributions , that can be found in the presence of a significant high - temperature tail for the proto - stellar sources , and in the average temperatures ( @xmath121 ) . it is worthy of notice that the temperature is estimated on the range 160 - 500 @xmath2 m : this implies that the difference between proto - stellar and starless sources is independent from the presence of a flux at 70 @xmath2 m , i.e. the constraint we impose to classify the proto - stellar sources . finally note that , unlike the size distribution , in this case there is no distance effect biasing the observed trend of the temperature . the combination of hi - gal and nanten data allow us to provide an estimate of the clump and star formation efficiency ( cfe and sfe , respectively ) for the four distance components seen towards @xmath30 . indicating the total mass of the compact sources ( both proto - stellar and starless ) with @xmath122 and the total cloud mass with @xmath123 , the cfe is expressed by @xmath124 . the cloud masses can be easily derived from the column density map of figure [ hiiregions ] , resolving the infrequent cases of overlap of two distance components along a co line of sight by adopting the distance corresponding to the brightest feature in the spectrum . due to the different spatial coverage of the hi - gal and nanten observations , for each component we derive again the cloud mass only in the lines of sight falling inside the common area . for the same reason , to compute @xmath122 we must consider only the hi - gal sources lying in the area surveyed in the co(1 - 0 ) line , but all the sources considered for the present analysis fulfill this requirement because they are required to have an associated kinematic distance . since the column density and the compact sources masses have been computed using the same opacity @xmath89 , our cfe will not be affected by the large uncertainties associated to the choice of this constant . the obtained cfes for the four components are @xmath125 , @xmath126 , @xmath127 , and @xmath128 , respectively . the uncertainties have been estimated assuming a conservative error of 10% on both the source masses and pixel - to - pixel column density , combined in quadrature . in the same way , if @xmath129 is the total mass of the formed stars , the sfe is expressed by @xmath130 . to compute @xmath129 we consider only the hi - gal proto - stellar cores lying in the area surveyed in the co(1 - 0 ) line , obtaining core total masses of 2880 , 2540 , 2140 , and 2550 @xmath131 for the four components , respectively . finally , these values have been converted in stellar masses assuming a core - to - star conversion efficiency @xmath132 @xcite , and a average sfe of @xmath133 , @xmath134 , @xmath135 , @xmath136 has been obtained in the four cases . here the relative uncertainties are not simply the same associated to the cfe , since one must take into account also the assumptions on the @xmath137 . following @xcite , we adopt @xmath138 as upper limit for this coefficient . further considerations , however , lead us to take these estimates as lower limits for the real sfe : @xmath81 ) this derivation of the sfe is based only on fir / sub - mm observations , neglecting the mass of more evolved young stellar objects detected only in the medium infrared ; @xmath82 ) as it will be discussed in section [ evol ] , it is possible that a small fraction of sources classified as starless cores , not taken into account for this calculation , may already have a strongly embedded early - stage proto - stellar content . on the other hand , at increasing distances the so - called compact sources correspond to structures having increasing physical sizes , departing from the definition of core . therefore , unresolved distant regions are considered as a single source , although containing in most cases a complex internal substructure consisting of dense cores and diffuse matter . in particular , for such sources classified as proto - stellar clumps , only a fraction of the mass is taking part in the ongoing star formation processes , nevertheless the total mass is entirely taken into account in the sfe calculation . in any case , these results are comparable with typical sfe values . in the 40 star forming complexes analyzed by @xcite , the sfe ranges from 0.002 to 0.2 , with an average of 0.08 . in @xcite , who analyzed 11 nearby molecular clouds , it ranges from @xmath139 ( pipe nebula ) to @xmath140 ( rcra ) . compared with these data , the sfe of our four components can be classified as relatively low . it is noteworthy , in this sense , that we find the lowest sfe in what we call component ii , that is likely spatially related to the maddalena s cloud , i.e. one of the most striking cases of giant molecular with an unusual combination of high gas mass and little evidence for star formation ( e.g. , * ? ? ? * ; * ? ? ? * ) . in the following discussion we focus on the starless objects that are also gravitationally bound , which we call pre - stellar cores , because they may give rise to the formation of proto - stellar clusters . in the recent herschel literature , to study the stability of such sources , the lack of kinematic information has forced the adoption of the critical bonnor - ebert ( be ) mass as a surrogate for the virial mass @xcite . we use it also in the present case : although in principle we could exploit the velocity dispersion obtainable the co(1 - 0 ) data to estimate the virial mass , in practice the co(1 - 0 ) line spectral resolution is too coarse for determining reliable line widths . the be mass is given by : @xmath141 where @xmath49 is the sound speed at the source temperature , @xmath142 is the gravitational constant , and @xmath143 is the be radius ( in pc ) , that in this case is assumed to coincide with the source radius observed at the reference wavelength . we consider gravitationally bound , and therefore pre - stellar , all the starless sources with @xmath144 ( cf . * ; * ? ? ? according to this definition , we find 398 , 131 , 42 and 19 pre - stellar sources associated to the gas velocity component i , ii , iii , and iv , corresponding to the 82% , 94% , 98% and 100% of the starless core populations , respectively . in our case the availability of four different distance components gives us the chance to discuss a trend emerging from the recent herschel literature . the studies of the relatively nearby star - forming regions of aquila @xcite , orion - a ( polychroni et al . 2013 , in prep . ) , and vela - c @xcite , located at 260 , 414 and 700 pc away , respectively , provided an estimate of the pre - stellar fraction over the total starless core number of 69% , 87% , and 94% , respectively . these percentages have been estimated using a less severe threshold for identifying the gravitationally bound sources , namely @xmath145 . if we adopt the same criterium for the components i - iv of the present work , the fraction of pre - stellar sources increases to 94% , 100% , 100% , 100% , respectively . we notice a general trend of this fraction to increase with the distance ( see figure [ prefrac ] ) , highlighting a clear selection effect against gravitationally unbound structures located at large distances ( at @xmath146 pc the saturation is practically achieved ) . this is implicit in fact in equation [ mbe ] : given a certain sed and changing its distance @xmath147 , @xmath96 varies as @xmath148 , while @xmath143 varies as @xmath147 , thus making it easier for the same sed to fulfil the be criterium when @xmath147 increases . on one hand , this means that to exploit this fraction of gravitationally bound sources in a region as an evolutionary indication about its star formation capability is affected by a strong and practically unsolvable bias related to the distance . on the other hand , the slight deviation of the behavior of the component i from the general trend seen in figure [ prefrac ] can be considered , from the qualitative point of view , indicative of an intrinsically lower occurrence of pre - stellar sources compared , for example , with the vela - c region . a mass vs radius plot , shown in figure [ masssize ] , can be used to illustrate and quantify the discussion on the gravitational stability of the starless cores ( neglecting the turbulence and magnetic field supports against gravity ) . the lines corresponding to 1 @xmath149 mass behavior at two different temperatures , @xmath150 k and 20 k , respectively , are displayed ( orange dashed lines ) . it can be seen that all the unbound sources ( red open symbols ) lie below the curve at @xmath151 k , while , at the same time , most of the bound ones ( red filled symbols ) show a remarkable degree of boundedness ( i.e. @xmath152 ) . although this diagram is meaningless in the case of the proto - stellar sources , since part of the mass has already formed the central protostar(s ) or could have been ejected in form of proto - stellar jets , we report also these sources in the plot ( blue filled symbols ) for completeness . in figure [ masssize ] `` larson s third law '' is shown as a green dotted - dashed line . originally formulated as @xmath153 @xcite it has been revised later : @xcite found a value 2 for the power - law exponent , i.e. a constant column density as a function of the radius . since the two values are very similar , we plot this relation in its original form . in any case , it can be seen that the sources of our sample do not obey a precise power law . therefore we do not try here to fit a power law to the mass - size distribution of the pre - stellar cores in @xmath30 . the mass vs radius diagram is also a useful tool for checking the presence of the conditions for msf , based on the intuitive concept that to form massive stars a conspicuous mass reservoir concentrated in a relatively small volume is required . in the recent literature a theoretical column density threshold of 1 g @xmath79 @xcite is largely invoked and used to identify candidate sites of msf ( see also section [ nh2_sect ] ) . it can be represented as a locus in the bi - logarithmic plot of figure [ masssize ] , down to a minimum core mass value that , for a 10 @xmath131 star and a core - to - star conversion efficiency @xmath154 , amounts to 30 @xmath131 . a small but significant fraction of sources ( both pre- and proto - stellar ) exceed this threshold and can be considered therefore potential progenitors of massive stars . also , an empirical and less demanding limit for msf has been estimated by @xcite , based on observations of infrared dark clouds ( irdcs * ? ? ? * ) , as @xmath155 and reported in figure [ masssize ] . further sources , belonging to all the four distance components of @xmath30 , fulfil this threshold , testifying the occurrence of the conditions for having msf also in this portion of the tgq and , in particular , at remarkably large galactocentric distances ( @xmath156 kpc for the component iv ) . moving to the analysis of the mass distribution , it becomes necessary to deal with homogeneous samples , from both the spatial and the evolutionary point of view . thus , we consider only the pre - stellar sources of the closest velocity component ( i ) , a choice that also ensures enough statistics to obtain a reliable mass distribution . again , it is correct to calculate this kind of distribution only for the pre - stellar sources , as the masses of the proto - stellar ones are not representative of the initial core mass . the core mass distribution is reported in figure [ cmf ] . the mass completeness limit of @xmath157 has been estimated from the flux completeness limit at the reference wavelength 250 @xmath2 m ( section [ cores ] ) using the modified black body equation , combined with the average temperature and distance of the sample , and @xmath68 . the turn over of the mass function falls around @xmath158 , too close to the completeness limit to claim that we are well tracing the log - normal behavior of the mass distribution below the turnover point ( as , e.g. , in * ? ? ? we then fit a power law @xmath159 to the high - mass part of the distribution using an approach based on the method of maximum likelihood and independent from the histogram binning ( see * ? ? ? * and references therein ) . it allows us to determine the lower limit of range of validity of the power - law , @xmath160 , and its slope @xmath161 . whereas the core mass functions derived for nearby regions through spectral lines ( orion a , @xmath162 , * ? ? ? * ) , spitzer ( perseus , serpens and ophiucus , @xmath163 , * ? ? ? * ) , and herschel observations ( aquila rift , @xmath164 , * ? ? ? * ) have slope values consistent with those of the stellar initial mass function ( @xmath165 for single stars , @xcite ; @xmath166 for multiple systems , @xcite ) , when the considered sample is contaminated by the presence of larger structures ( clumps ) as in our case ( see figure [ masssize ] ) , the slope generally tends to be shallower , towards the values found for the clump mass functions ( e.g. , @xmath167 for various clouds , @xcite ; @xmath168 for the galactic center region , @xcite ; @xmath169 for the vela - d cloud , @xcite ) . a very similar behavior was observed also in the case of the herschel - based source mass distribution of the vela - c cloud ( @xmath170 , * ? ? ? * ) , affected by the same kind of contamination . in our case , we can check on the same region this gradual steepening of the mass function from clumps to cores , by considering the masses of the co clumps extracted as described in section [ cdecomp ] , and belonging to the component i ( purple histogram in figure [ cmf ] , @xmath171 ) . this steepening is quite usual and is explained , in the general case , as a consequence of the fragmentation of the clumps in cores that constitute the last fragment generating the single star ( or multiple system ) . however , in our case , notice that the error bars associated to the two slopes ( obtained with the same method ) partially overlap . furthermore , the disagreement between herschel- and co - derived column densities highlighted in figure [ nh2co ] surely introduces an uncertainty on the slope of the co clump mass function larger than that derived simply from the power - law fit . it is possible to make a direct comparison with the mass function slope derived by @xcite for their `` group i '' clumps ( namely sources associated with cma ob1 and g220.8 - 1.7 ) , @xmath172 , that is consistent with our result . to further investigate the differences between the proto- and pre - stellar sources considered so far , we discuss here a plot of the bolometric luminosities vs the envelope masses . in the previous literature @xcite such a diagram has been already used as a meaningful method to infer the evolutionary status of the compact sources , through the comparison between their positions in the plot and the theoretical tracks obtained from models of accreting cores . such tracks start from a given initial mass and initially follow an almost vertical path corresponding to the accretion phase . subsequently , the proto - stellar outflow activity produces mass loss and disperses the residual envelope ( horizontal portion of the tracks ) . recently , this diagnostics has been applied to the clump populations of star forming regions observed by herschel @xcite , revealing some degree of segregation between proto- and pre - stellar sources . a general caveat about the use of this diagram for distant star forming regions consists in the fact that , rigorously speaking , the evolutionary tracks are calculated for a single protostar . instead , in cases like ours , many sources are in fact clumps having an unresolved internal structure , with the possible presence of few proto - stars , and/or smaller pre - stellar cores . therefore , a precise star formation timeline can not be deduced from such tracks . of course also in these cases to find a higher luminosity at the same envelope mass is indicative of a more evolved evolutionary stage , but it is difficult to understand whether this is a general property of the entire ( unresolved ) population hosted by the clump , or it is essentially due to only the brightmost contained source(s ) . to build this diagram we used the mass values discussed in the previous sections . they represent , in fact , the envelope mass in the case of a proto - stellar source , and of course the whole mass of a starless one . the luminosities have been derived integrating the whole observed sed , including @xmath81 ) the flux of the best - fitting modified black body at 1 mm , to take into account also the millimetric portion of the sed we do not observe directly , and @xmath82 ) the wise flux at 22 @xmath2 m , where available , to ensure a better estimate in the case of the sources most evolved and then bright in the mid - infrared unlike , for example , the similar plot in @xcite . by the way , in @xcite and @xcite , since their targets were constituted by limited amounts of proto - stellar sources ( this explains the lack of the pre - stellar ones in their analysis ) , more extended seds were collected , down to 8.3 and 3.6 @xmath2 m , respectively . on one hand , this increases and improves the bolometric luminosity estimation for more evoled sources , on the other hand this operation requires special care in checking the reliability of the short - wavelength counterpart assignment to each single fir source , and therefore it lies outside the goals of the present work . the @xmath173 vs @xmath174 plot is reported in figure [ lm_temp ] , whith the same symbol convention of figure [ masssize ] . furthermore , the symbol size linearly scales with the source temperature . the theoretical evolutionary tracks for the low and high mass regimes adopted by @xcite are also plotted . also in our case , the proto - stellar sources and the pre - stellar ones populate quite different regions of the diagram , corresponding to the accreting core phase in the first case , and to a quiescent or collapsing core in the second , respectively . obviously , there is a transition region with some overlap between the two populations , especially at high masses ( discussed below ) , but a global look suggests a certain degree of segregation . in the low - mass regime , where the distinction of well - defined evolutionary classes makes more sense ( * ? ? ? * and references therein ) , the region corresponding to the class 0 to i transition appears populated , as in @xcite . however , we remind the reader that the @xmath173 vs @xmath174 plot is not a tool * suitable * for obtaining a precise classification , essentially because @xmath81 ) @xcite showed that sources classified as class i from their nir - to - mir sed are found to be compatible instead with a class 0 status based on their fir fluxes , and there is not a sharp transition from class 0 to i to ii in their @xmath173 vs @xmath174 diagram ; @xmath82 ) this evolutionary classification can not be easily extended to the high - mass regime , a fortiori when the considered source is a large clump hosting multiple star formation . furthermore , comparing our diagram with those of @xcite and @xcite , we can explain the lack of evolved sources in ours with the aforementioned different ranges used for calculating @xmath173 . in summary , due to both the investigated spectral range and mass regime , this @xmath173 vs @xmath174 plot can not be used to obtain a classical single - yso classification . from the point of view of the distance , it is remarkable that proto - stellar activity is detected also in the most distant sources , confirming the occurrence of the star formation also at @xmath156 kpc ( see , e.g. , * ? ? ? * ; * ? ? ? * and references therein ) . focusing on the source temperatures , the first general indication is that the most luminous proto - stellar sources , i. e. the most evolved , are also the warmest , and of course the proto - stellar are on average warmer than the pre - stellar ones as already indicated by figure [ sizetemp ] . furthermore , the luminous pre - stellar sources occupying regions of the diagram where early - stage proto - stellar cores are instead expected ( beginning of the evolutionary tracks ) are generally characterized by a higher temperature . we interpret this fact as the presence of a strongly embedded early - phase protostar(s ) not detected at both the 22 and 70 @xmath2 m wavebands . this might lead to misclassify early proto - stellar sources as pre - stellar , a fact that should be taken into account in the future analyses of core / clump populations extracted from the herschel maps . finally , we notice that while the proto - stellar sources are spread over a wide range of possible evolutionary stages , the displacement of the bulk of the pre - stellar sources follows a power - law in a region of the diagram corresponding to a stage in which no mass accretion has started yet . a similar behavior was already discussed in @xcite for the vela - c cloud but here it looks more evident and , more interestingly , it appears the same over several spatially unrelated regions . a power - law fit gives a dependence @xmath175 . this slope is very similar to the 0.85 value found by @xcite for their sample of proto - stellar sources , while it is very different from the 0.42 value found by @xcite for pre - stellar sources in vela - c . however , we need to increase the statistics of this slope across the galactic plane , exploiting the whole hi - gal data base , before stating a possible link between the evolutionary stage of a star forming region and the slope of the luminosity vs mass relation of the pre - stellar cores . the @xmath173 vs @xmath174 diagram can be exploited to quantify the star formation rate ( sfr ) of the four distance components we identify in @xmath176 . for the high mass regime ( @xmath177 we followed the method of @xcite . briefly , it consists in associating each proto - stellar source mass to a nearby evolutionary track , and to consider the mass @xmath178 of the central object at the end of the track ( corresponding to a time @xmath179 ) , when the envelope clean - up phase is completed . the sfr is calculated by adding the @xmath180 ratios of all sources . in the low - mass regime , instead , we use the equation 1 of @xcite that assumes a constant final mass and time , @xmath181 and @xmath182 yr , respectively . anyway , the contribution of the low - mass part to the total sfr amounts to 8% for the component i and less than 1% for the remaining components . we find , for the components i - iv , @xmath183 yr@xmath45 , respectively . in the portion of the galaxy corresponding to @xmath176 , thus , @xmath184 yr@xmath45 . the error bars have been estimated assuming a 20% average uncertainty on both mass and luminosity . of course , larger uncertainties come from the choice of the yso model and from the distance estimate and are not easily quantifiable ( see * ? ? ? * ) . in two single hi - gal tiles of the first galactic quadrant , namely centered close to @xmath37=[30^{\circ},0^{\circ}]$ ] and @xmath185 $ ] , @xcite obtained @xmath186 and @xmath187 yr@xmath45 , respectively , considering all the distance components together . taking into account that the solid angle spanned by the @xmath176 maps is about five times larger than in the case of these two inner galaxy fields , we conclude that in our case the sfr per solid angle is significantly lower , as expected for the outer galaxy . however , the forthcoming studies based on the hi - gal survey will provide a more exhaustive description of the sfr as a function of the galactic position , and an extrapolation to the sfr of the entire galaxy . if one wants to to apply a diagnostics based on a luminosity - mass relation also to the clumps extracted from the co(1 - 0 ) data , the most meaningful way is to use the virial mass . indeed , the way the clump mass is calculated in our case , namely as proportional to the intensity integrated over all the channels assigned by the cloud decomposing algorithm to the clump ( equation [ xco ] ) , makes it directly proportional to the luminosity , which strictly speaking is a `` luminous mass '' : @xmath188 @xcite , who used the same decomposition algorithm we used , report a value for the constant @xmath189 with @xmath190 in @xmath131 and @xmath191 in k km s@xmath45 pc@xmath33 . instead , the estimate of the virial mass is independent from the clump luminosity . we show the plot of these two quantities ( derived as in section [ cdecomp ] ) in figure [ lm_co ] , panel @xmath49 , where we adopt the same symbol convention of figure [ lm_temp ] to separate sources associated to the different distance components . furthermore , blue and red colors are used for clumps associated with at least one hi - gal proto - stellar source and clumps associated only with hi - gal starless ones , respectively . finally , smaller green symbols represent the co clumps that do not contain hi - gal sources , or that lie outside the area surveyed by herschel . two lines corresponding to equation [ mlum ] for two different slopes have been plotted ( grey dotted lines ) . they are obtained correcting the @xmath85 value of @xcite ( corresponding to @xmath192 ( k km s@xmath45)@xmath45 @xmath79 ) , by the minimum ( upper line ) and maximum ( lower line ) values of @xmath57 used for the clumps of this paper , respectively , according to equation [ xco ] . this is equivalent to saying that all the values of the luminosity - luminous mass pairs lie in the stripe of the plot delimited by these two lines ( and in particular the majority of them lies close to the upper one , namely that corresponding to the smaller value of @xmath57 ) . the less rigorous classification between `` proto - stellar '' and `` starless '' co clumps probably explains the reason of the absence of a clear segregation like that found in figure [ lm_temp ] for dust cores and clumps . it is noteworthy , however , that for @xmath193 m@xmath99 the most luminous clumps @xmath81 ) all have a proto - stellar content ; @xmath82 ) all are supervirial . if we consider all the supervirial clumps in our sample , namely 22 objects corresponding to @xmath194 of the total , 17 of them have a proto - stellar content , 3 are starless and 2 are not associable with herschel sources . this indication is surely relevant , but not resolutive or sufficient to identify evolved clumps in such a diagram . on one hand , the majority of the supervirial structures are found to host star formation ; on the other hand , many `` proto - stellar '' clumps can appear as sub - virial , since their global physical properties plotted here can not describe local conditions that could allow star formation in limited parts of them . finally , unlike the case of the hi - gal @xmath173 vs @xmath96 diagram , we do not recognize any power - law trend either in the whole sample of clumps or in the single identified classes . the first hi - gal observations of the tgq consist of a `` chunk '' of @xmath195 deg@xmath33 corresponding to the longitude range @xmath196 . here we presented a first study based on both the overall displacement of the far - infrared emission in this region and on the statistical analysis of the properties of the compact sources derived from photometric measurements . the nanten co(1 - 0 ) observations have added the fundamental information about gas kinematics . this study confirms the view of the outer galaxy as an interesting laboratory to investigate star formation processes , benefitting of the generally simple structure of the velocity field ( compared with the inner galaxy ) . summarizing , we found that : * in the hi - gal images it is possible to recognize the dust emission from all the velocity components in this sky region detected in the line observations presented in this and previous papers @xcite . in particular we identify four components ( i - iv ) located at average heliocentric distances of 1.1 , 2.2 , 3.3 , and 5.8 kpc respectively . the overlap among these components is quite infrequent , making clearer their identification . * the large - scale column density and temperature distribution for the cold component of the dust reveal a variety of conditions , with star formation sites preferentially distributed along filaments . interestingly , two of these are likely associated with the s296 hii region , which might be responsible of triggering star formation in these filaments . * we obtained a five - band hi - gal compact source catalog of this region , complemented with the wise 22 @xmath2 m photometry and with the distance information derived from the spatial association with the clumps extracted from the co(1 - 0 ) data set . the classification between proto - stellar ( provided with a detection at 70 @xmath2 m and/or at 22 @xmath2 m and 160 @xmath2 m ) and starless sources has been made . * a modified black body has been fitted to the regular and well sampled seds ( 255 proto - stellar , 688 starless ) , to derive temperatures and masses . on average , the proto - stellar sources are found to be warmer than the starless ones . concerning the source size , moreover , we find that our sample is composed by a mixture of clumps and cores , due to the wide involved distance range . looking at their spatial distribution in the galactic plane , the local and perseus spiral arm are clearly seen , although a remarkable amount of sources with assigned inter - arm distances are also found . the limited range of investigated galactic longitudes does not allow a better recognition of a grand design spiral pattern . * we derived the sfe for the four velocity components ( 0.008 , 0.004 , 0.012 , 0.006 , respectively ) . these values lead to classify the star formation in these regions as not very efficient . on one hand , being only based on herschel far - infrared photometry , they should be taken as lower limits . on the other hand , a systematic overestimate of the sfe with increasing distance has to be taken into account . * applying a virial analysis to the starless hi - gal sources , large fractions of them ( ranging from 82% for the component i to 100% for the component iv ) are found to be gravitationally bound , i. e. pre - stellar . we show that these percentages must be considered as upper limits , because they are biased by a selection effect related to the distance and implicit in the adopted criterium . * in the mass vs radius plot for the pre - stellar sources we do not recognize any clear power - law behavior . we use this tool to check the potential ability of clump / cores to form massive stars , based on the different thresholds of @xcite and @xcite . in both cases we find condensations compatible with high - mass star formation in all the distance components . * a power - law @xmath159 has been fitted to high - mass portion of the core / clump mass distribution of the pre - stellar sources belonging to the component i. similarly to the vela - c case @xcite , its slope @xmath161 is shallower ( although still consistent within the errors ) than the stellar imf slope and the similar slopes also found for cores in nearby star forming regions , and steeper than the slopes found for the gas clump distributions , as in this case ( @xmath171 ) . * the @xmath173 vs @xmath174 plot of the hi - gal sources reveals an evident separation between proto- and pre - stellar sources . for both classes the warmest sources are , in general , the most evolved of their class . few proto - stellar sources populate the region of the diagram where objects in transition between class 0 and class i are expected to lie , whereas the majority populate the class 0 region . this classification obviously weakens when distant sources are considered , namely those corresponding to large and massive ( @xmath197 ) clumps . using this diagram as in @xcite we are able to derive also an estimate of the star formation rate of @xmath198 yr@xmath45 . * most of the co clumps ( 93% ) are subvirial . line observations at better angular resolution would probably reveal denser supervirial substructures . the remaining 7% fraction of supervirial clumps ( 22 objects ) is in most cases associated with hi - gal proto - stellar sources , suggesting that these clumps fulfill globally and not only locally the conditions for the gravitational collapse to occurr . + besides the characterization of star forming regions in the @xmath176 , this article gives also a first contribution to the larger statistics that will be built exploiting the entire hi - gal survey archive . prescriptions and caveats on using the photometric data , on combining them with kinematics derived from co observations , and on addressing effects of distance - resolution have been explicitly discussed in several points of the text . pacs has been developed by a consortium of institutes led by mpe ( germany ) and including uvie ( austria ) ; ku leuven , csl , imec ( belgium ) ; cea , lam ( france ) ; mpia ( germany ) ; inaf - ifsi / oaa / oap / oat , lens , sissa ( italy ) ; iac ( spain ) . this development has been supported by the funding agencies bmvit ( austria ) , esaprodex ( belgium ) , cea / cnes ( france ) , dlr ( germany ) , asi / inaf ( italy),and cicyt / mcyt ( spain ) . spire has been developed by a consortium of institutes led by cardiff univ . ( uk ) and including : univ . lethbridge ( canada ) ; naoc ( china ) ; cea , lam ( france ) ; inaf - ifsi , univ . padua ( italy ) ; iac ( spain ) ; stockholm observatory ( sweden ) ; imperial college london , ral , ucl - mssl , ukatc , univ . sussex ( uk ) ; and caltech , jpl , nhsc , univ . colorado ( usa ) . this development has been supported by national funding agencies : csa ( canada ) ; naoc ( china ) ; cea , cnes , cnrs ( france ) ; asi ( italy ) ; mcinn ( spain ) ; snsb ( sweden ) ; stfc , uksa ( uk ) ; and nasa ( usa ) . de , mp , and kjlr are funded by an asi fellowship under contract numbers i/005/11/0 and i/038/08/0 . dp is funded through the operational program `` education and lifelong learning '' and is co - financed by the european union ( european social fund ) and greek national funds . _ facilities : _ , . lccccccccc 1 & 216.61829 & -1.82868 & 44.0 & 1.8 & 4.3 & 5.7 & 5770 . & 19154 . & 704.6 + 2 & 216.62572 & -1.64908 & 30.6 & 0.9 & 2.6 & 6.5 & 2707 . & 5009 . & 366.3 + 3 & 216.64683 & -2.24570 & 23.8 & 1.9 & 1.9 & 3.6 & 3943 . & 13551 . & 554.5 + 4 & 216.66404 & -2.08390 & 25.9 & 2.0 & 2.1 & 5.0 & 6811 . & 20697 . & 947.4 + 5 & 216.66706 & -1.70928 & 32.8 & 0.8 & 2.8 & 8.9 & 5936 . & 6110 . & 791.9 + 6 & 216.67371 & -1.42988 & 30.9 & 1.1 & 2.6 & 5.5 & 3611 . & 7170 . & 487.8 + 7 & 216.68618 & -0.48407 & 55.2 & 1.7 & 6.1 & 8.7 & 6204 . & 25071 . & 671.9 + 8 & 216.69400 & -1.88198 & 29.7 & 1.6 & 2.5 & 7.8 & 8458 . & 19509 . & 1151.3 + 9 & 216.70952 & -0.44180 & 26.5 & 1.2 & 2.1 & 2.8 & 401 . & 3964 . & 55.6 + 10 & 216.72395 & -2.36709 & 24.1 & 1.5 & 1.9 & 4.0 & 3712 . & 9794 . & 521.4 + 11 & 216.74965 & -1.87533 & 51.6 & 0.9 & 5.4 & 8.2 & 4526 . & 6321 . & 512.5 + 12 & 216.81554 & -1.10761 & 30.7 & 1.6 & 2.6 & 5.0 & 4327 . & 13447 . & 585.9 + 13 & 216.81786 & -1.35122 & 31.9 & 1.1 & 2.7 & 5.8 & 4182 . & 7258 . & 561.7 + 14 & 216.84120 & -1.98124 & 24.9 & 1.7 & 2.0 & 6.5 & 7738 . & 20583 . & 1082.9 + 15 & 216.87085 & -2.05228 & 31.6 & 2.2 & 2.7 & 10.4 & 1980 . & 52479 . & 266.4 + 16 & 216.87131 & -2.20133 & 20.4 & 1.6 & 1.5 & 5.3 & 5724 . & 13453 . & 819.4 + 17 & 216.89132 & -1.53862 & 30.5 & 1.6 & 2.6 & 6.0 & 2776 . & 15040 . & 376.3 + 18 & 216.91345 & -2.21344 & 27.6 & 1.7 & 2.3 & 6.8 & 11262 . & 20623 . & 1552.9 + 19 & 216.95635 & -0.11574 & 20.1 & 1.1 & 1.5 & 2.1 & 105 . & 2814 . & 15.0 + 20 & 216.96588 & -1.31297 & 29.3 & 1.1 & 2.4 & 6.1 & 1111 . & 7907 . & 151.7 + 21 & 216.97759 & -1.82817 & 30.4 & 1.9 & 2.5 & 7.1 & 1296 . & 26417 . & 175.8 + 22 & 216.98623 & -0.36119 & 58.7 & 1.6 & 6.7 & 12.9 & 4969 . & 35366 . & 516.3 + 23 & 217.01355 & -1.80752 & 24.7 & 1.9 & 1.9 & 5.3 & 1398 . & 19528 . & 195.9 + 24 & 217.07652 & -1.55184 & 50.0 & 1.0 & 5.1 & 11.5 & 13848 . & 12248 . & 1602.2 + 25 & 217.07859 & -0.71760 & 25.3 & 1.4 & 2.0 & 2.4 & 505 . & 4893 . & 70.6 + 26 & 217.13268 & -1.83133 & 54.8 & 2.2 & 5.9 & 15.1 & 18877 . & 72960 . & 2068.5 + 27 & 217.13893 & -1.40723 & 51.3 & 1.0 & 5.3 & 10.7 & 8402 . & 11192 . & 958.8 + 28 & 217.14853 & -2.16251 & 22.4 & 1.1 & 1.7 & 6.6 & 1262 . & 8803 . & 179.1 + 29 & 217.16301 & -1.83244 & 32.8 & 1.0 & 2.8 & 9.7 & 5215 . & 10575 . & 697.5 + 30 & 217.17056 & -1.45563 & 32.0 & 1.2 & 2.7 & 3.2 & 697 . & 4431 . & 93.8 + 31 & 217.17616 & -2.09673 & 41.9 & 1.3 & 3.9 & 4.8 & 5176 . & 8574 . & 647.0 + 32 & 217.18633 & -1.21478 & 52.3 & 1.1 & 5.5 & 9.4 & 2024 . & 11649 . & 228.6 + 33 & 217.19393 & -1.31640 & 54.1 & 0.9 & 5.8 & 19.7 & 8476 . & 16360 . & 938.0 + 34 & 217.21256 & -1.16125 & 30.2 & 1.5 & 2.5 & 5.5 & 1463 . & 11980 . & 198.9 + 35 & 217.24703 & -0.56298 & 26.5 & 0.9 & 2.1 & 4.6 & 1241 . & 3551 . & 172.4 + 36 & 217.26359 & -1.45943 & 47.4 & 0.9 & 4.7 & 5.0 & 4689 . & 4539 . & 558.4 + 37 & 217.27331 & -2.03231 & 28.2 & 1.6 & 2.3 & 9.2 & 8295 . & 24668 . & 1141.5 + 38 & 217.27597 & -0.25348 & 16.1 & 1.1 & 1.1 & 1.9 & 1194 . & 2327 . & 174.3 + 39 & 217.30057 & -2.25778 & 26.0 & 1.5 & 2.1 & 7.9 & 5359 . & 18446 . & 746.5 + 40 & 217.32318 & -2.12037 & 33.3 & 1.4 & 2.9 & 9.7 & 3466 . & 18649 . & 462.3 + 41 & 217.33114 & -1.33540 & 50.2 & 1.3 & 5.1 & 6.5 & 18263 . & 11112 . & 2114.3 + 42 & 217.35038 & -0.07286 & 26.6 & 1.5 & 2.1 & 2.8 & 8211 . & 6348 . & 1140.5 + 43 & 217.43133 & -2.20728 & 41.8 & 1.2 & 3.9 & 5.1 & 4524 . & 7506 . & 566.8 + 44 & 217.43689 & -1.62192 & 28.3 & 1.3 & 2.3 & 4.8 & 1908 . & 8826 . & 262.7 + 45 & 217.48365 & -2.23507 & 33.3 & 1.6 & 2.8 & 9.5 & 7115 . & 26795 . & 950.1 + 46 & 217.48940 & -1.73221 & 26.8 & 1.9 & 2.1 & 3.3 & 1479 . & 12223 . & 205.2 + 47 & 217.52488 & -0.07782 & 23.6 & 0.9 & 1.8 & 2.4 & 1631 . & 2197 . & 230.3 + 48 & 217.52520 & -2.05917 & 41.2 & 1.4 & 3.8 & 6.6 & 2650 . & 14100 . & 334.2 + 49 & 217.58998 & -0.65801 & 23.4 & 2.7 & 1.8 & 1.8 & 409 . & 13394 . & 57.8 + 50 & 217.59529 & -2.17524 & 42.1 & 1.4 & 3.9 & 8.1 & 6917 . & 15589 . & 866.2 + 51 & 217.59816 & -2.33317 & 30.8 & 2.1 & 2.6 & 7.2 & 13550 . & 33112 . & 1838.6 + 52 & 217.62439 & -0.83164 & 26.2 & 2.0 & 2.1 & 3.8 & 1473 . & 15375 . & 205.2 + 53 & 217.64804 & -2.13506 & 30.8 & 1.5 & 2.6 & 6.4 & 5097 . & 15993 . & 691.8 + 54 & 217.67658 & -0.27373 & 21.9 & 0.8 & 1.6 & 3.7 & 502 . & 2716 . & 71.5 + 55 & 217.67767 & -1.74748 & 26.9 & 1.3 & 2.1 & 2.9 & 523 . & 5515 . & 72.6 + 56 & 217.67987 & -0.18635 & 26.4 & 1.6 & 2.1 & 4.5 & 4979 . & 12423 . & 693.3 + 57 & 217.69956 & -1.09217 & 26.6 & 1.5 & 2.1 & 4.1 & 540 . & 10151 . & 75.1 + 58 & 217.71058 & -0.68278 & 28.7 & 1.5 & 2.3 & 7.4 & 1276 . & 16562 . & 175.4 + 59 & 217.71504 & -2.28461 & 51.8 & 1.4 & 5.3 & 6.5 & 10590 . & 14067 . & 1209.7 + 60 & 217.79114 & -1.87444 & 33.6 & 1.8 & 2.9 & 6.5 & 4509 . & 21059 . & 601.5 + 61 & 217.81175 & -2.07470 & 25.4 & 1.2 & 2.0 & 4.1 & 4330 . & 6365 . & 606.0 + 62 & 217.82549 & -0.75672 & 24.1 & 1.6 & 1.8 & 4.8 & 835 . & 12811 . & 117.7 + 63 & 217.83223 & -2.04447 & 29.8 & 1.2 & 2.4 & 7.3 & 4193 . & 10394 . & 572.8 + 64 & 217.85049 & -1.31174 & 52.5 & 0.9 & 5.4 & 8.4 & 8898 . & 7358 . & 1011.8 + 65 & 217.85109 & -1.86782 & 26.2 & 1.0 & 2.1 & 5.9 & 1294 . & 6493 . & 180.4 + 66 & 217.85510 & -0.83449 & 28.9 & 1.0 & 2.3 & 5.4 & 1700 . & 6074 . & 233.6 + 67 & 217.91541 & -2.26242 & 32.8 & 1.4 & 2.8 & 8.7 & 9466 . & 16527 . & 1269.9 + 68 & 217.93003 & -2.06501 & 34.7 & 1.2 & 3.0 & 8.6 & 5723 . & 13022 . & 758.8 + 69 & 217.93015 & -0.65582 & 26.7 & 1.6 & 2.1 & 5.2 & 2446 . & 13292 . & 340.3 + 70 & 217.97411 & -2.33651 & 29.9 & 1.0 & 2.4 & 6.5 & 2935 . & 7020 . & 400.9 + 71 & 218.00000 & -2.02904 & 25.3 & 1.1 & 2.0 & 3.8 & 2982 . & 4643 . & 417.7 + 72 & 218.03485 & -0.12363 & 27.0 & 1.3 & 2.1 & 5.4 & 2548 . & 9998 . & 354.0 + 73 & 218.03869 & -2.08777 & 9.6 & 1.0 & 0.5 & 1.0 & 115 . & 1039 . & 17.2 + 74 & 218.04493 & -1.82313 & 32.3 & 1.5 & 2.7 & 9.3 & 8882 . & 21335 . & 1196.4 + 75 & 218.05294 & -0.30104 & 28.4 & 1.3 & 2.3 & 6.7 & 24699 . & 11965 . & 3405.1 + 76 & 218.11081 & -0.36260 & 21.1 & 1.5 & 1.5 & 3.7 & 780 . & 8942 . & 111.7 + 77 & 218.11564 & -1.82006 & 28.0 & 1.4 & 2.2 & 6.6 & 5698 . & 13715 . & 787.3 + 78 & 218.14830 & -2.20643 & 28.4 & 1.3 & 2.3 & 6.5 & 2948 . & 11655 . & 406.4 + 79 & 218.16229 & -0.54526 & 25.5 & 1.3 & 2.0 & 5.4 & 10267 . & 9427 . & 1438.1 + 80 & 218.17237 & -2.31574 & 31.8 & 0.8 & 2.6 & 6.3 & 2378 . & 3966 . & 321.4 + 81 & 218.17886 & -1.45503 & 63.8 & 1.0 & 7.4 & 13.9 & 17026 . & 15169 . & 1680.9 + 82 & 218.20702 & -2.26182 & 34.3 & 0.9 & 2.9 & 7.8 & 4813 . & 6200 . & 640.6 + 83 & 218.22667 & -2.04647 & 33.7 & 1.3 & 2.8 & 8.2 & 6292 . & 15240 . & 840.7 + 84 & 218.24427 & -1.99777 & 26.0 & 1.5 & 2.0 & 3.8 & 3589 . & 9506 . & 501.4 + 85 & 218.31216 & -0.73223 & 29.0 & 1.1 & 2.3 & 5.1 & 1712 . & 6136 . & 235.5 + 86 & 218.34052 & -2.34400 & 31.3 & 1.2 & 2.6 & 5.4 & 2203 . & 8751 . & 298.9 + 87 & 218.35491 & -0.12208 & 28.0 & 1.0 & 2.2 & 3.7 & 1876 . & 3835 . & 259.5 + 88 & 218.35827 & -0.33033 & 27.2 & 1.6 & 2.1 & 5.8 & 4653 . & 16313 . & 646.4 + 89 & 218.36033 & -2.19445 & 29.4 & 1.2 & 2.4 & 8.7 & 2001 . & 12997 . & 274.4 + 90 & 218.38635 & -2.13527 & 26.6 & 1.3 & 2.1 & 8.7 & 1213 . & 14362 . & 169.1 + 91 & 218.39177 & -0.31332 & 32.9 & 1.6 & 2.7 & 9.6 & 3952 . & 24602 . & 531.2 + 92 & 218.44969 & -1.88503 & 26.6 & 2.2 & 2.1 & 6.2 & 1264 . & 30638 . & 176.1 + 93 & 218.45210 & -1.68007 & 28.4 & 2.2 & 2.3 & 6.7 & 2408 . & 32962 . & 332.2 + 94 & 218.45495 & -1.43652 & 64.6 & 1.3 & 7.5 & 14.2 & 20753 . & 24681 . & 2035.8 + 95 & 218.45760 & -1.22120 & 32.8 & 1.3 & 2.7 & 7.0 & 2805 . & 12414 . & 377.4 + 96 & 218.45861 & -1.87242 & 32.8 & 1.2 & 2.7 & 10.3 & 5586 . & 15164 . & 751.5 + 97 & 218.49399 & -0.46073 & 32.2 & 1.6 & 2.7 & 3.5 & 950 . & 9555 . & 128.3 + 98 & 218.49497 & -0.08676 & 59.4 & 1.4 & 6.5 & 6.9 & 9305 . & 13403 . & 983.7 + 99 & 218.50857 & -1.84000 & 29.9 & 1.0 & 2.4 & 14.3 & 964 . & 13625 . & 132.0 + 100 & 218.50865 & -0.44554 & 28.3 & 1.7 & 2.2 & 6.2 & 2341 . & 17734 . & 323.5 + 101 & 218.53027 & -2.36531 & 32.0 & 2.0 & 2.6 & 5.7 & 4153 . & 23123 . & 561.4 + 102 & 218.58402 & -0.90594 & 22.0 & 1.5 & 1.6 & 2.4 & 277 . & 5592 . & 39.5 + 103 & 218.59721 & -2.08729 & 34.7 & 1.1 & 2.9 & 17.3 & 4920 . & 20718 . & 654.5 + 104 & 218.61812 & -0.61193 & 26.8 & 2.3 & 2.1 & 2.9 & 906 . & 16318 . & 126.2 + 105 & 218.65378 & -1.30765 & 34.7 & 1.3 & 2.9 & 9.2 & 2583 . & 15113 . & 343.6 + 106 & 218.65763 & -1.39615 & 29.7 & 1.0 & 2.4 & 4.5 & 1385 . & 5068 . & 189.8 + 107 & 218.66472 & -2.18965 & 31.7 & 1.6 & 2.6 & 8.2 & 4358 . & 22846 . & 590.4 + 108 & 218.66841 & -1.63922 & 34.7 & 2.3 & 2.9 & 6.8 & 3245 . & 35693 . & 431.7 + 109 & 218.69304 & -0.76835 & 30.7 & 0.9 & 2.5 & 4.1 & 715 . & 3150 . & 97.5 + 110 & 218.75213 & -0.44043 & 55.9 & 2.2 & 5.8 & 10.0 & 22483 . & 49378 . & 2488.3 + 111 & 218.77892 & -1.84992 & 31.8 & 1.1 & 2.6 & 7.8 & 4688 . & 9030 . & 635.3 + 112 & 218.79081 & -1.77784 & 34.6 & 1.5 & 2.9 & 8.3 & 3365 . & 18212 . & 448.1 + 113 & 218.80046 & -1.36580 & 36.1 & 2.1 & 3.1 & 7.4 & 3952 . & 35247 . & 521.5 + 114 & 218.80047 & -0.40866 & 22.4 & 2.4 & 1.6 & 2.2 & 385 . & 12918 . & 54.8 + 115 & 218.82371 & -0.17535 & 60.4 & 1.0 & 6.6 & 12.6 & 4323 . & 12250 . & 453.5 + 116 & 218.84698 & -2.32736 & 30.7 & 1.6 & 2.5 & 5.7 & 2865 . & 15311 . & 390.9 + 117 & 218.88310 & -1.51551 & 45.5 & 0.6 & 4.3 & 7.0 & 1620 . & 2226 . & 199.1 + 118 & 218.90849 & -2.23681 & 27.5 & 1.7 & 2.2 & 8.2 & 3131 . & 25332 . & 434.7 + 119 & 218.94575 & -1.95912 & 32.1 & 1.5 & 2.6 & 8.1 & 3619 . & 18335 . & 489.9 + 120 & 218.95846 & -2.35521 & 0.0 & 1.1 & 0.2 & 0.3 & 4 . & 346 . & 0.6 + 121 & 218.96655 & -1.23197 & 34.2 & 1.7 & 2.9 & 8.4 & 2671 . & 24761 . & 357.1 + 122 & 219.05289 & -2.20551 & 30.6 & 1.7 & 2.5 & 6.9 & 6944 . & 19697 . & 948.4 + 123 & 219.06029 & -1.84722 & 30.1 & 1.5 & 2.4 & 8.7 & 4807 . & 20610 . & 658.6 + 124 & 219.16395 & -1.64909 & 28.3 & 2.4 & 2.2 & 9.5 & 1150 . & 56849 . & 159.1 + 125 & 219.21546 & -1.18276 & 35.1 & 1.6 & 2.9 & 7.3 & 2609 . & 20157 . & 347.2 + 126 & 219.23624 & -0.35023 & 31.5 & 0.6 & 2.5 & 4.1 & 477 . & 1509 . & 64.9 + 127 & 219.24433 & -2.31049 & 28.2 & 1.5 & 2.2 & 6.6 & 981 . & 15808 . & 135.9 + 128 & 219.29442 & -1.56392 & 30.6 & 1.9 & 2.5 & 7.5 & 2261 . & 27567 . & 309.1 + 129 & 219.29619 & -1.11911 & 25.4 & 1.2 & 1.9 & 2.9 & 1226 . & 4182 . & 172.3 + 130 & 219.30648 & -2.12974 & 30.0 & 1.7 & 2.4 & 3.1 & 1889 . & 9450 . & 259.1 + 131 & 219.31675 & -1.09479 & 17.4 & 1.3 & 1.2 & 2.2 & 157 . & 4018 . & 22.9 + 132 & 219.35489 & -1.24640 & 32.1 & 1.2 & 2.6 & 7.9 & 1343 . & 12180 . & 182.1 + 133 & 219.39198 & -0.98446 & 16.3 & 1.1 & 1.1 & 2.0 & 214 . & 2320 . & 31.4 + 134 & 219.42825 & -1.16293 & 35.7 & 1.1 & 3.0 & 7.4 & 2342 . & 9987 . & 310.7 + 135 & 219.43898 & -1.41559 & 27.7 & 2.5 & 2.1 & 2.7 & 711 . & 17937 . & 98.8 + 136 & 219.45225 & -1.70908 & 32.4 & 1.6 & 2.6 & 4.9 & 2188 . & 12845 . & 296.2 + 137 & 219.47299 & -1.52531 & 34.6 & 0.5 & 2.9 & 14.6 & 1174 . & 4257 . & 156.8 + 138 & 219.47882 & -0.90028 & 32.3 & 1.3 & 2.6 & 6.7 & 1525 . & 12275 . & 206.6 + 139 & 219.49497 & -1.51006 & 51.1 & 2.0 & 5.0 & 8.2 & 2497 . & 33324 . & 293.2 + 140 & 219.51974 & -1.60112 & 30.9 & 1.5 & 2.5 & 6.4 & 5074 . & 13947 . & 693.1 + 141 & 219.56282 & -0.99047 & 35.0 & 1.3 & 2.9 & 6.7 & 1744 . & 11647 . & 232.4 + 142 & 219.59557 & -1.39016 & 31.8 & 1.1 & 2.6 & 7.3 & 3631 . & 9816 . & 493.4 + 143 & 219.64914 & -0.91668 & 34.8 & 1.1 & 2.9 & 6.8 & 2268 . & 8228 . & 302.9 + 144 & 219.66109 & -1.44992 & 34.7 & 0.9 & 2.9 & 4.6 & 805 . & 4151 . & 107.6 + 145 & 219.67558 & -2.17179 & 30.2 & 1.2 & 2.4 & 3.8 & 1610 . & 6050 . & 220.9 + 146 & 219.68265 & -1.18561 & 17.7 & 0.7 & 1.2 & 1.7 & 139 . & 958 . & 20.3 + 147 & 219.71339 & -2.37602 & 12.6 & 0.8 & 0.8 & 1.6 & 260 . & 1090 . & 38.5 + 148 & 219.73047 & -2.04188 & 26.9 & 1.5 & 2.1 & 4.7 & 3284 . & 11769 . & 458.6 + 149 & 219.73797 & -0.11036 & 36.7 & 1.0 & 3.1 & 4.6 & 1363 . & 5075 . & 180.0 + 150 & 219.77673 & -1.44166 & 32.2 & 1.4 & 2.6 & 4.0 & 1114 . & 7827 . & 151.2 + 151 & 219.78109 & -1.25631 & 36.5 & 2.1 & 3.1 & 3.5 & 984 . & 15995 . & 130.1 + 152 & 219.83611 & -0.92944 & 25.1 & 1.4 & 1.9 & 3.4 & 2307 . & 7222 . & 325.1 + 153 & 219.84160 & -2.01182 & 30.9 & 0.9 & 2.5 & 6.5 & 2812 . & 5238 . & 384.3 + 154 & 219.84398 & -2.19972 & 12.2 & 0.9 & 0.7 & 1.4 & 194 . & 1331 . & 28.8 + 155 & 219.84890 & -1.27179 & 18.5 & 1.1 & 1.3 & 3.4 & 304 . & 4271 . & 44.2 + 156 & 219.86069 & -1.00864 & 17.9 & 2.4 & 1.2 & 2.1 & 393 . & 12785 . & 57.2 + 157 & 219.86273 & -2.21884 & 29.1 & 0.9 & 2.3 & 4.0 & 973 . & 3543 . & 134.4 + 158 & 219.86409 & -1.37585 & 19.5 & 1.7 & 1.4 & 2.7 & 451 . & 8542 . & 65.2 + 159 & 219.96747 & -0.13501 & 33.7 & 1.6 & 2.8 & 6.8 & 1097 . & 17718 . & 147.6 + 160 & 219.97538 & -2.37066 & 12.6 & 0.8 & 0.8 & 1.3 & 176 . & 817 . & 26.1 + 161 & 220.02312 & -1.90182 & 31.0 & 1.2 & 2.5 & 6.4 & 3694 . & 9119 . & 505.1 + 162 & 220.04748 & -1.97550 & 34.2 & 1.2 & 2.8 & 8.1 & 4155 . & 12371 . & 557.6 + 163 & 220.04813 & -2.14261 & 13.2 & 0.9 & 0.8 & 2.1 & 525 . & 1737 . & 77.7 + 164 & 220.07720 & -0.57029 & 26.9 & 1.8 & 2.0 & 5.4 & 974 . & 17223 . & 136.1 + 165 & 220.12973 & -1.20465 & 33.7 & 1.3 & 2.7 & 7.6 & 4004 . & 12966 . & 539.3 + 166 & 220.13756 & -1.26730 & 18.5 & 0.8 & 1.3 & 3.6 & 650 . & 2180 . & 94.3 + 167 & 220.14007 & -0.28800 & 30.1 & 1.0 & 2.4 & 3.2 & 1805 . & 3631 . & 248.2 + 168 & 220.14528 & -0.44460 & 24.4 & 1.0 & 1.8 & 4.8 & 1426 . & 4544 . & 201.8 + 169 & 220.14630 & -0.12833 & 36.0 & 0.6 & 3.0 & 6.4 & 1865 . & 2241 . & 247.7 + 170 & 220.16245 & -1.50227 & 33.9 & 1.0 & 2.8 & 5.7 & 1621 . & 5583 . & 218.1 + 171 & 220.16500 & -1.44552 & 36.2 & 0.9 & 3.0 & 7.5 & 2678 . & 6900 . & 355.2 + 172 & 220.22442 & -1.95845 & 33.0 & 1.0 & 2.7 & 6.4 & 4533 . & 6825 . & 613.2 + 173 & 220.27914 & -0.68848 & 34.9 & 2.2 & 2.9 & 5.7 & 2129 . & 27799 . & 284.9 + 174 & 220.29300 & -0.85491 & 22.2 & 1.4 & 1.6 & 3.7 & 642 . & 7797 . & 91.8 + 175 & 220.29833 & -2.26271 & 12.3 & 1.0 & 0.7 & 1.4 & 288 . & 1330 . & 42.8 + 176 & 220.30473 & -2.08209 & 13.1 & 0.7 & 0.8 & 1.6 & 281 . & 721 . & 41.7 + 177 & 220.32080 & -1.79356 & 14.0 & 1.0 & 0.9 & 2.5 & 743 . & 2573 . & 109.7 + 178 & 220.34555 & -1.24850 & 18.2 & 1.1 & 1.2 & 1.7 & 332 . & 2297 . & 48.3 + 179 & 220.39158 & -1.99274 & 34.1 & 0.8 & 2.8 & 5.2 & 2416 . & 3575 . & 324.9 + 180 & 220.43577 & -1.37462 & 18.3 & 1.2 & 1.2 & 1.8 & 280 . & 2712 . & 40.8 + 181 & 220.44935 & -0.69346 & 22.2 & 1.1 & 1.6 & 2.8 & 328 . & 3782 . & 46.9 + 182 & 220.45987 & -0.25588 & 28.8 & 1.7 & 2.2 & 2.2 & 907 . & 7035 . & 125.7 + 183 & 220.46326 & -2.09991 & 11.3 & 0.9 & 0.6 & 1.6 & 377 . & 1263 . & 56.2 + 184 & 220.47238 & -0.62280 & 29.1 & 1.9 & 2.2 & 2.8 & 3729 . & 10703 . & 516.1 + 185 & 220.50764 & -2.23354 & 34.5 & 2.8 & 2.8 & 6.0 & 1354 . & 49782 . & 181.7 + 186 & 220.50895 & -1.64215 & 14.4 & 1.3 & 0.9 & 1.9 & 435 . & 3046 . & 64.2 + 187 & 220.56194 & -0.14878 & 36.8 & 1.5 & 3.1 & 6.7 & 2350 . & 15867 . & 311.0 + 188 & 220.62506 & -1.90139 & 12.1 & 1.1 & 0.7 & 1.8 & 1114 . & 2116 . & 165.7 + 189 & 220.70520 & -2.13950 & 9.1 & 1.2 & 0.5 & 0.8 & 146 . & 1183 . & 21.9 + 190 & 220.74079 & -2.36335 & 33.3 & 1.3 & 2.7 & 6.5 & 2391 . & 12118 . & 323.4 + 191 & 220.74275 & -2.04401 & 33.4 & 0.4 & 2.7 & 2.9 & 978 . & 580 . & 132.2 + 192 & 220.77745 & -1.68182 & 13.4 & 1.1 & 0.8 & 2.0 & 1884 . & 2433 . & 278.9 + 193 & 220.82142 & -2.36739 & 13.4 & 0.9 & 0.8 & 1.4 & 432 . & 1176 . & 64.0 + 194 & 220.84033 & -0.33599 & 24.5 & 0.8 & 1.8 & 3.1 & 873 . & 2079 . & 123.7 + 195 & 220.85550 & -2.01005 & 9.0 & 1.0 & 0.5 & 0.9 & 144 . & 917 . & 21.6 + 196 & 220.86121 & -2.23778 & 33.7 & 1.4 & 2.7 & 4.4 & 981 . & 9584 . & 132.4 + 197 & 221.00045 & -1.86797 & 8.4 & 1.2 & 0.4 & 0.8 & 98 . & 1145 . & 14.7 + 198 & 221.02278 & -2.15159 & 35.6 & 1.3 & 2.9 & 6.9 & 2433 . & 12875 . & 325.0 + 199 & 221.10790 & -2.06077 & 34.0 & 1.6 & 2.7 & 7.0 & 1417 . & 19366 . & 191.2 + 200 & 221.19192 & -1.82072 & 7.9 & 1.3 & 0.4 & 0.8 & 132 . & 1506 . & 19.9 + 201 & 221.21485 & -1.67292 & 8.3 & 1.4 & 0.4 & 0.7 & 145 . & 1425 . & 21.9 + 202 & 221.29949 & -1.51836 & 9.6 & 1.0 & 0.5 & 1.2 & 179 . & 1371 . & 26.9 + 203 & 221.32065 & -1.86544 & 36.5 & 0.5 & 3.0 & 6.4 & 914 . & 1910 . & 121.7 + 204 & 221.43579 & -2.24274 & 39.0 & 0.9 & 3.3 & 5.8 & 6760 . & 5322 . & 885.3 + 205 & 221.51510 & -2.01515 & 36.4 & 1.6 & 3.0 & 8.0 & 1626 . & 20661 . & 216.6 + 206 & 221.58965 & -1.71627 & 34.3 & 1.5 & 2.7 & 3.0 & 747 . & 7316 . & 100.8 + 207 & 221.64320 & -1.93632 & 40.2 & 1.3 & 3.4 & 8.0 & 9524 . & 13011 . & 1239.4 + 208 & 221.66099 & -0.65995 & 25.1 & 2.7 & 1.8 & 2.9 & 699 . & 21336 . & 99.0 + 209 & 221.66236 & -2.17210 & 38.1 & 1.4 & 3.2 & 7.3 & 6080 . & 14427 . & 802.1 + 210 & 221.85319 & -1.49250 & 17.2 & 1.0 & 1.1 & 1.9 & 408 . & 1775 . & 59.7 + 211 & 221.88716 & -2.03011 & 40.0 & 1.1 & 3.4 & 8.8 & 47071 . & 11043 . & 6138.4 + 212 & 221.94101 & -1.18234 & 21.4 & 1.1 & 1.5 & 2.9 & 706 . & 3945 . & 101.6 + 213 & 222.04347 & -0.46200 & 18.4 & 0.9 & 1.2 & 1.4 & 128 . & 1086 . & 18.7 + 214 & 222.13298 & -1.65963 & 17.6 & 0.9 & 1.2 & 2.6 & 743 . & 2100 . & 108.5 + 215 & 222.14348 & -1.25561 & 20.6 & 0.9 & 1.4 & 3.8 & 1415 . & 3079 . & 204.3 + 216 & 222.15225 & -2.23575 & 38.7 & 1.2 & 3.2 & 10.6 & 35038 . & 16903 . & 4613.9 + 217 & 222.17169 & -1.42468 & 19.4 & 1.0 & 1.3 & 4.0 & 717 . & 4469 . & 104.1 + 218 & 222.17464 & -0.50214 & 16.5 & 0.6 & 1.1 & 2.2 & 468 . & 721 . & 68.6 + 219 & 222.26914 & -2.39338 & 16.1 & 0.8 & 1.0 & 1.2 & 210 . & 823 . & 30.9 + 220 & 222.28893 & -1.74660 & 20.2 & 1.4 & 1.4 & 2.0 & 354 . & 4065 . & 51.1 + 221 & 222.38925 & -0.57429 & 17.8 & 1.3 & 1.2 & 3.0 & 313 . & 5431 . & 45.8 + 222 & 222.41979 & -1.50987 & 17.9 & 1.2 & 1.2 & 4.5 & 787 . & 6976 . & 114.9 + 223 & 222.45446 & -0.23248 & 51.1 & 0.9 & 4.7 & 5.7 & 3606 . & 4721 . & 434.4 + 224 & 222.47657 & -0.40751 & 16.2 & 1.0 & 1.0 & 2.7 & 639 . & 2742 . & 93.8 + 225 & 222.55848 & -2.00582 & 15.3 & 1.3 & 1.0 & 1.9 & 323 . & 3287 . & 47.5 + 226 & 222.59226 & -1.65167 & 17.8 & 0.9 & 1.2 & 2.8 & 1503 . & 2238 . & 219.6 + 227 & 222.62328 & -1.60328 & 42.4 & 1.3 & 3.6 & 3.9 & 6567 . & 6525 . & 845.8 + 228 & 222.71451 & -0.65952 & 16.6 & 1.0 & 1.1 & 3.4 & 770 . & 3844 . & 113.0 + 229 & 222.76605 & -0.37975 & 16.1 & 1.0 & 1.0 & 2.6 & 1378 . & 2691 . & 202.4 + 230 & 222.80675 & -1.60301 & 17.9 & 1.1 & 1.2 & 4.1 & 3314 . & 5337 . & 484.0 + 231 & 222.86894 & -1.86472 & 39.4 & 1.5 & 3.2 & 7.2 & 1854 . & 17842 . & 243.7 + 232 & 222.88445 & -2.01860 & 38.8 & 1.3 & 3.2 & 7.0 & 7979 . & 12375 . & 1053.3 + 233 & 222.89596 & -1.60599 & 41.9 & 1.2 & 3.5 & 4.1 & 4531 . & 6365 . & 586.4 + 234 & 222.91114 & -1.02965 & 15.9 & 1.1 & 1.0 & 1.5 & 80 . & 1778 . & 11.8 + 235 & 222.94071 & -0.49868 & 15.1 & 0.6 & 0.9 & 2.5 & 407 . & 872 . & 60.0 + 236 & 222.95052 & -1.51563 & 12.8 & 1.1 & 0.7 & 1.1 & 155 . & 1412 . & 23.0 + 237 & 222.95085 & -0.38800 & 12.7 & 1.0 & 0.7 & 1.4 & 275 . & 1394 . & 40.8 + 238 & 222.97828 & -0.36949 & 17.4 & 0.9 & 1.1 & 3.0 & 1167 . & 2454 . & 170.8 + 239 & 222.98334 & -0.61279 & 58.6 & 0.6 & 5.7 & 8.4 & 3511 . & 3351 . & 396.4 + 240 & 223.02400 & -1.33220 & 13.2 & 0.8 & 0.8 & 1.5 & 274 . & 1065 . & 40.6 + 241 & 223.02604 & -1.47150 & 18.5 & 0.9 & 1.2 & 3.3 & 1233 . & 2679 . & 179.7 + 242 & 223.04577 & -2.21752 & 15.7 & 1.2 & 1.0 & 2.9 & 438 . & 4517 . & 64.4 + 243 & 223.08384 & -1.81147 & 40.9 & 0.8 & 3.4 & 6.4 & 2759 . & 4361 . & 359.7 + 244 & 223.10717 & -0.84201 & 15.4 & 0.9 & 1.0 & 3.2 & 1845 . & 2620 . & 272.0 + 245 & 223.11403 & -1.87932 & 17.9 & 1.1 & 1.2 & 3.4 & 2839 . & 3989 . & 414.7 + 246 & 223.12868 & -0.39210 & 14.0 & 1.0 & 0.8 & 1.6 & 404 . & 1583 . & 59.8 + 247 & 223.15225 & -0.17586 & 13.2 & 1.2 & 0.8 & 1.4 & 189 . & 2283 . & 28.1 + 248 & 223.16135 & -0.57153 & 14.9 & 0.9 & 0.9 & 1.9 & 372 . & 1720 . & 54.9 + 249 & 223.25247 & -1.67840 & 39.9 & 1.1 & 3.3 & 5.4 & 2652 . & 7355 . & 348.2 + 250 & 223.26401 & -0.33130 & 15.4 & 1.3 & 1.0 & 2.3 & 256 . & 4055 . & 37.8 + 251 & 223.27004 & -1.78248 & 38.8 & 0.8 & 3.1 & 4.6 & 1825 . & 3325 . & 241.4 + 252 & 223.30090 & -2.18953 & 19.5 & 1.6 & 1.3 & 4.9 & 935 . & 12393 . & 135.8 + 253 & 223.37318 & -1.80454 & 18.1 & 1.1 & 1.2 & 2.9 & 4904 . & 3619 . & 715.9 + 254 & 223.43536 & -0.61934 & 13.1 & 0.9 & 0.8 & 2.1 & 421 . & 1911 . & 62.6 + 255 & 223.46962 & -0.92311 & 15.9 & 1.1 & 1.0 & 3.0 & 2036 . & 3562 . & 299.7 + 256 & 223.50709 & -2.04035 & 17.1 & 1.1 & 1.1 & 2.5 & 1860 . & 3317 . & 272.5 + 257 & 223.51112 & -2.06330 & 12.4 & 1.2 & 0.7 & 2.0 & 208 . & 3114 . & 31.0 + 258 & 223.57972 & -2.34839 & 18.6 & 1.0 & 1.2 & 1.9 & 172 . & 1882 . & 25.1 + 259 & 223.62754 & -1.83456 & 17.8 & 1.2 & 1.1 & 2.8 & 5876 . & 4133 . & 859.2 + 260 & 223.62901 & -2.19556 & 11.4 & 1.9 & 0.6 & 0.7 & 35 . & 2614 . & 5.2 + 261 & 223.68933 & -2.15481 & 16.8 & 1.4 & 1.1 & 1.2 & 228 . & 2596 . & 33.4 + 262 & 223.72345 & -0.52264 & 12.2 & 2.0 & 0.7 & 1.8 & 38 . & 7317 . & 5.7 + 263 & 223.73529 & -0.09304 & 15.4 & 0.5 & 1.0 & 1.3 & 113 . & 330 . & 16.6 + 264 & 223.74755 & -1.97876 & 16.9 & 1.1 & 1.1 & 2.4 & 2976 . & 2796 . & 436.5 + 265 & 223.74931 & -0.81232 & 13.5 & 1.5 & 0.8 & 2.5 & 1680 . & 5411 . & 249.3 + 266 & 223.76004 & -2.01628 & 63.3 & 1.0 & 6.3 & 10.7 & 14505 . & 12030 . & 1570.3 + 267 & 223.81334 & -0.61099 & 14.1 & 1.5 & 0.8 & 2.4 & 218 . & 5438 . & 32.2 + 268 & 223.84369 & -0.09735 & 51.8 & 1.1 & 4.6 & 6.4 & 2904 . & 7427 . & 351.4 + 269 & 223.86839 & -1.53879 & 19.2 & 1.1 & 1.3 & 2.2 & 1914 . & 2818 . & 278.4 + 270 & 223.88749 & -0.49225 & 14.4 & 1.7 & 0.9 & 1.4 & 109 . & 4250 . & 16.1 + 271 & 223.90961 & -1.86607 & 17.7 & 1.1 & 1.1 & 2.7 & 6299 . & 3240 . & 921.5 + 272 & 223.94819 & -2.31370 & 14.7 & 1.0 & 0.9 & 1.2 & 63 . & 1261 . & 9.2 + 273 & 223.94945 & -0.96932 & 42.9 & 0.9 & 3.6 & 4.7 & 1487 . & 3598 . & 192.2 + 274 & 223.96870 & -0.07250 & 11.3 & 0.7 & 0.6 & 0.5 & 41 . & 283 . & 6.2 + 275 & 224.01014 & -1.17245 & 15.1 & 1.0 & 0.9 & 2.8 & 3807 . & 3187 . & 562.0 + 276 & 224.02807 & -1.70343 & 17.4 & 1.3 & 1.1 & 3.0 & 2434 . & 5476 . & 356.6 + 277 & 224.09656 & -0.11342 & 15.3 & 0.8 & 0.9 & 1.7 & 283 . & 1112 . & 41.8 + 278 & 224.10830 & -0.87756 & 13.3 & 1.4 & 0.8 & 2.4 & 2738 . & 4954 . & 406.6 + 279 & 224.16174 & -1.93486 & 17.8 & 1.5 & 1.1 & 2.9 & 9700 . & 6698 . & 1418.7 + 280 & 224.18161 & -1.64760 & 21.7 & 1.2 & 1.5 & 3.4 & 1383 . & 4749 . & 199.4 + 281 & 224.18974 & -0.56697 & 12.6 & 1.2 & 0.7 & 2.6 & 542 . & 3862 . & 80.7 + 282 & 224.20744 & -0.30617 & 14.4 & 1.2 & 0.9 & 2.8 & 533 . & 3991 . & 78.9 + 283 & 224.27956 & -1.03729 & 14.3 & 1.6 & 0.9 & 1.9 & 2159 . & 4838 . & 319.7 + 284 & 224.29481 & -2.15138 & 16.7 & 1.3 & 1.1 & 2.3 & 2776 . & 4290 . & 407.6 + 285 & 224.38363 & -1.97467 & 17.1 & 1.3 & 1.1 & 2.5 & 5231 . & 4636 . & 767.1 + 286 & 224.42319 & -0.71043 & 15.1 & 1.5 & 0.9 & 3.5 & 6564 . & 8356 . & 969.5 + 287 & 224.44950 & -1.22910 & 15.3 & 0.9 & 0.9 & 1.9 & 649 . & 1607 . & 95.7 + 288 & 224.45196 & -2.37347 & 9.6 & 1.1 & 0.5 & 0.6 & 255 . & 643 . & 38.3 + 289 & 224.46809 & -1.68803 & 14.8 & 2.3 & 0.9 & 2.5 & 320 . & 14214 . & 47.3 + 290 & 224.48688 & -0.10934 & 15.5 & 1.8 & 0.9 & 2.8 & 459 . & 9218 . & 67.7 + 291 & 224.49204 & -2.14882 & 15.1 & 1.4 & 0.9 & 2.5 & 133 . & 4820 . & 19.7 + 292 & 224.54490 & -2.14914 & 18.2 & 0.8 & 1.2 & 2.1 & 361 . & 1417 . & 52.7 + 293 & 224.58083 & -0.96230 & 16.5 & 1.1 & 1.0 & 3.2 & 2881 . & 3944 . & 423.5 + 294 & 224.58199 & -1.52493 & 15.6 & 0.9 & 1.0 & 2.3 & 406 . & 1918 . & 59.8 + 295 & 224.59567 & -0.38858 & 16.0 & 1.6 & 1.0 & 3.7 & 4324 . & 10140 . & 636.8 + 296 & 224.61948 & -1.85638 & 18.8 & 1.2 & 1.2 & 4.1 & 1701 . & 6079 . & 248.0 + 297 & 224.65739 & -1.32624 & 16.8 & 0.9 & 1.1 & 2.4 & 610 . & 1798 . & 89.6 + 298 & 224.67817 & -2.01929 & 16.1 & 0.9 & 1.0 & 1.9 & 313 . & 1639 . & 46.1 + 299 & 224.72063 & -0.96201 & 17.7 & 1.1 & 1.1 & 3.0 & 946 . & 3740 . & 138.6 + 300 & 224.73413 & -1.71988 & 16.5 & 1.4 & 1.0 & 3.0 & 3720 . & 5979 . & 546.9 + 301 & 224.74864 & -0.15232 & 13.0 & 0.9 & 0.7 & 3.6 & 174 . & 2951 . & 25.9 + 302 & 224.81018 & -1.56172 & 16.9 & 0.8 & 1.1 & 4.2 & 118 . & 2522 . & 17.3 + 303 & 224.81874 & -0.84184 & 18.6 & 1.2 & 1.2 & 3.0 & 192 . & 4308 . & 28.0 + 304 & 224.83916 & -1.49400 & 18.7 & 1.3 & 1.2 & 3.4 & 736 . & 5640 . & 107.3 + 305 & 224.85275 & -0.41591 & 15.7 & 1.5 & 1.0 & 2.9 & 2919 . & 6942 . & 430.4 + 306 & 224.92372 & -1.15640 & 12.7 & 0.6 & 0.7 & 2.1 & 324 . & 689 . & 48.2 + 307 & 224.92913 & -1.17375 & 9.8 & 1.0 & 0.5 & 1.2 & 68 . & 1155 . & 10.2 + 308 & 225.05476 & -0.20742 & 17.0 & 1.0 & 1.1 & 3.1 & 1894 . & 3097 . & 278.1 + 309 & 225.06605 & -0.59853 & 15.5 & 1.5 & 0.9 & 2.5 & 1095 . & 6073 . & 161.6 + 310 & 225.15878 & -0.74363 & 19.7 & 0.9 & 1.3 & 2.5 & 437 . & 2220 . & 63.6 + 311 & 225.19247 & -0.33050 & 15.0 & 0.8 & 0.9 & 2.4 & 567 . & 1669 . & 83.8 + 312 & 225.24214 & -1.07539 & 13.9 & 0.8 & 0.8 & 1.9 & 649 . & 1310 . & 96.2 + 313 & 225.25760 & -0.33575 & 17.1 & 0.7 & 1.1 & 2.4 & 1076 . & 1323 . & 158.0 + 314 & 225.26637 & -0.91563 & 14.7 & 0.8 & 0.9 & 2.2 & 829 . & 1584 . & 122.6 + 315 & 225.27990 & -1.34988 & 13.4 & 1.1 & 0.8 & 1.5 & 142 . & 1794 . & 21.2 + 316 & 225.31169 & -0.55449 & 17.3 & 1.3 & 1.1 & 2.6 & 1790 . & 4904 . & 262.7 + 317 & 225.34649 & -0.11049 & 17.6 & 1.1 & 1.1 & 2.3 & 1116 . & 2962 . & 163.5 + 318 & 225.37235 & -0.51791 & 11.8 & 0.8 & 0.6 & 0.8 & 55 . & 602 . & 8.1 + 319 & 225.42389 & -0.23851 & 15.7 & 1.1 & 1.0 & 2.0 & 1025 . & 2284 . & 151.2 + 320 & 225.43577 & -0.97221 & 10.7 & 0.9 & 0.6 & 1.1 & 67 . & 903 . & 10.0 + 321 & 225.44266 & -0.38711 & 17.5 & 1.0 & 1.1 & 1.8 & 883 . & 1842 . & 129.5 + lcccccccccccccc 1 & 216.66360 & -1.66179 & 2.6 & & & 0.43@xmath199 0.04 & 0.28@xmath199 0.02 & 0.17@xmath199 0.03 & 21.0 & 6.4 & 11.9 & & 0.4 & 1 + 2 & 216.68802 & -1.42173 & 2.6 & & & 0.29@xmath199 0.07 & 0.29@xmath199 0.01 & 0.12@xmath199 0.02 & 13.6 & 5.9 & 11.5 & & 0.3 & 1 + 3 & 216.69028 & -1.37625 & 2.6 & & & 0.31@xmath199 0.09 & 0.27@xmath199 0.02 & 0.23@xmath199 0.01 & 21.9 & 17.5 & 9.3 & & 0.4 & 1 + 4 & 216.69757 & -1.41449 & 2.6 & & & 0.89@xmath199 0.07 & 0.79@xmath199 0.01 & 0.51@xmath199 0.05 & 31.0 & 33.4 & 10.1 & & 1.1 & 1 + 5 & 216.74443 & -0.39604 & 2.1 & & & 0.46@xmath199 0.11 & 0.52@xmath199 0.01 & 0.27@xmath199 0.04 & 25.6 & 13.4 & 9.9 & & 0.4 & 1 + 6 & 216.78496 & -1.06081 & 2.6 & 0.87@xmath199 0.03 & 2.45@xmath199 0.12 & 3.99@xmath199 0.15 & 2.61@xmath199 0.02 & 1.36@xmath199 0.01 & 16.2 & 53.1 & 12.1 & 0.01 & 17.2 & 2 + 7 & 216.78586 & -1.04352 & 2.6 & 24.01@xmath199 0.14 & 26.32@xmath199 0.24 & 20.11@xmath199 0.18 & 8.66@xmath199 0.02 & 3.61@xmath199 0.01 & 19.4 & 55.6 & 18.1 & 3.18 & 322.1 & 2 + 8 & 216.79010 & -1.33692 & 2.7 & & & 0.35@xmath199 0.06 & 0.30@xmath199 0.01 & 0.19@xmath199 0.01 & 16.7 & 12.2 & 10.4 & & 0.4 & 1 + 9 & 216.82965 & -1.30140 & 2.7 & & & 0.62@xmath199 0.05 & 0.39@xmath199 0.01 & 0.22@xmath199 0.02 & 17.7 & 8.0 & 12.6 & & 0.6 & 1 + 10 & 216.84781 & -1.32426 & 2.7 & & & 0.83@xmath199 0.10 & 0.66@xmath199 0.02 & 0.36@xmath199 0.03 & 23.0 & 18.8 & 11.3 & & 1.0 & 1 + 11 & 216.85027 & -1.54283 & 2.6 & & & 0.14@xmath199 0.03 & 0.20@xmath199 0.01 & 0.10@xmath199 0.02 & 15.1 & 8.8 & 9.3 & & 0.2 & 1 + 12 & 216.94028 & -1.58797 & 5.1 & & & 0.15@xmath199 0.05 & 0.14@xmath199 0.01 & 0.07@xmath199 0.01 & 10.5 & 14.3 & 11.0 & & 0.7 & 1 + 13 & 216.99194 & -1.31877 & 2.4 & & & 0.31@xmath199 0.07 & 0.27@xmath199 0.02 & 0.16@xmath199 0.01 & 20.9 & 7.5 & 10.7 & & 0.3 & 1 + 14 & 217.01163 & -0.11884 & 1.5 & 1.32@xmath199 0.04 & 2.90@xmath199 0.10 & 3.37@xmath199 0.14 & 1.87@xmath199 0.02 & 0.73@xmath199 0.02 & 13.2 & 6.4 & 14.6 & 0.02 & 6.6 & 2 + 15 & 217.02290 & -0.12783 & 1.5 & & & 0.63@xmath199 0.04 & 0.67@xmath199 0.02 & 0.48@xmath199 0.02 & 21.6 & 13.7 & 9.2 & & 0.3 & 1 + 16 & 217.03499 & -1.61789 & 5.1 & & & 0.53@xmath199 0.04 & 0.42@xmath199 0.01 & 0.24@xmath199 0.01 & 19.4 & 47.1 & 11.0 & & 2.2 & 1 + 17 & 217.05092 & -0.72714 & 2.0 & & & 0.84@xmath199 0.06 & 0.58@xmath199 0.01 & 0.40@xmath199 0.02 & 20.9 & 11.2 & 10.9 & & 0.5 & 1 + 18 & 217.06100 & -1.71599 & 1.9 & 0.38@xmath199 0.03 & 0.40@xmath199 0.06 & 0.28@xmath199 0.06 & 0.11@xmath199 0.01 & & 11.0 & 0.3 & 19.9 & 0.01 & 2.2 & 2 + 19 & 217.06459 & -1.61942 & 5.1 & & & 0.34@xmath199 0.06 & 0.26@xmath199 0.01 & 0.17@xmath199 0.02 & 13.9 & 35.2 & 10.7 & & 1.4 & 1 + 20 & 217.06668 & -0.71351 & 2.0 & & & 0.58@xmath199 0.05 & 0.31@xmath199 0.01 & 0.29@xmath199 0.02 & 15.3 & 6.0 & 11.2 & & 0.3 & 1 + 21 & 217.09753 & -1.62461 & 5.1 & & & 0.12@xmath199 0.11 & 0.15@xmath199 0.01 & 0.08@xmath199 0.02 & 14.8 & 27.0 & 9.5 & & 0.7 & 1 + 22 & 217.12683 & -1.57470 & 5.1 & & & 0.54@xmath199 0.08 & 0.48@xmath199 0.01 & 0.22@xmath199 0.04 & 19.5 & 40.1 & 11.6 & & 2.3 & 1 + 23 & 217.13768 & -1.53865 & 5.1 & 1.02@xmath199 0.05 & 1.18@xmath199 0.07 & 0.94@xmath199 0.10 & 0.42@xmath199 0.01 & 0.13@xmath199 0.01 & 9.5 & 8.6 & 18.6 & 0.03 & 43.3 & 2 + 24 & 217.14229 & -1.39884 & 5.3 & 0.48@xmath199 0.02 & 0.53@xmath199 0.06 & 0.65@xmath199 0.11 & 0.66@xmath199 0.03 & 0.29@xmath199 0.05 & 14.6 & 49.7 & 12.0 & 0.07 & 31.3 & 2 + 25 & 217.16275 & -1.23764 & 5.8 & 0.43@xmath199 0.02 & 0.34@xmath199 0.05 & 0.26@xmath199 0.09 & 0.14@xmath199 0.01 & & 13.6 & 5.0 & 16.8 & & 12.4 & 2 + 26 & 217.16357 & -1.83950 & 2.8 & & & 0.28@xmath199 0.04 & 0.27@xmath199 0.01 & 0.20@xmath199 0.01 & 19.9 & 17.9 & 9.4 & & 0.4 & 1 + 27 & 217.16840 & -0.04018 & 2.1 & & & 1.36@xmath199 0.09 & 0.77@xmath199 0.02 & 0.60@xmath199 0.01 & 20.7 & 14.5 & 11.6 & & 0.8 & 1 + 28 & 217.17271 & -1.35582 & 5.3 & & & 0.42@xmath199 0.24 & 0.53@xmath199 0.01 & 0.42@xmath199 0.02 & 27.3 & 216.3 & 8.4 & & 3.0 & 1 + 29 & 217.17929 & -0.05392 & 2.1 & & & 0.94@xmath199 0.12 & 1.53@xmath199 0.01 & 1.10@xmath199 0.02 & 28.6 & 112.6 & 8.0 & & 1.2 & 1 + 30 & 217.19145 & -1.49453 & 2.7 & 1.05@xmath199 0.03 & 3.79@xmath199 0.15 & 4.73@xmath199 0.18 & 2.53@xmath199 0.01 & 1.30@xmath199 0.03 & 25.3 & 40.8 & 13.6 & 0.07 & 26.6 & 2 + 31 & 217.19412 & -0.51546 & 2.1 & & & 0.20@xmath199 0.04 & 0.15@xmath199 0.01 & 0.10@xmath199 0.01 & 17.1 & 3.7 & 10.6 & & 0.1 & 1 + 32 & 217.20294 & -0.59034 & 2.1 & & & 1.09@xmath199 0.07 & 0.92@xmath199 0.01 & 0.56@xmath199 0.01 & 22.1 & 21.1 & 10.5 & & 0.8 & 1 + 33 & 217.21700 & -1.48283 & 4.7 & & & 1.70@xmath199 0.13 & 0.97@xmath199 0.02 & 0.40@xmath199 0.02 & 23.2 & 29.0 & 15.4 & & 4.4 & 1 + 34 & 217.22528 & -0.26994 & 1.1 & & & 1.24@xmath199 0.09 & 1.42@xmath199 0.01 & 1.07@xmath199 0.02 & 26.0 & 19.6 & 8.8 & & 0.3 & 1 + 35 & 217.23386 & -0.25371 & 1.1 & 0.35@xmath199 0.02 & 0.43@xmath199 0.06 & 1.25@xmath199 0.07 & 1.72@xmath199 0.01 & & 21.4 & 22.0 & 8.8 & 0.10 & 1.5 & 2 + 36 & 217.23518 & -0.24417 & 1.1 & & & 2.38@xmath199 0.29 & 2.12@xmath199 0.01 & 1.52@xmath199 0.02 & 25.7 & 19.1 & 9.8 & & 0.5 & 1 + 37 & 217.24686 & -0.23750 & 1.1 & & & 1.24@xmath199 0.18 & 1.59@xmath199 0.01 & 0.86@xmath199 0.02 & 22.8 & 13.9 & 9.4 & & 0.3 & 1 + 38 & 217.25539 & -0.21553 & 1.1 & 0.81@xmath199 0.03 & 0.74@xmath199 0.06 & 1.19@xmath199 0.12 & 0.93@xmath199 0.02 & 0.38@xmath199 0.01 & 14.8 & 3.0 & 12.1 & 0.20 & 2.4 & 2 + 39 & 217.26141 & -1.46862 & 4.7 & 1.76@xmath199 0.06 & 3.36@xmath199 0.10 & 3.80@xmath199 0.14 & 2.38@xmath199 0.01 & 1.04@xmath199 0.02 & 20.1 & 97.7 & 13.9 & 0.18 & 98.5 & 2 + 40 & 217.26904 & -1.36247 & 5.1 & 0.68@xmath199 0.04 & 2.50@xmath199 0.12 & 1.25@xmath199 0.12 & 0.75@xmath199 0.02 & 0.27@xmath199 0.01 & 16.2 & 14.5 & 18.9 & 0.02 & 48.0 & 2 + 41 & 217.27194 & -0.34158 & 1.1 & & & 0.50@xmath199 0.11 & 0.40@xmath199 0.01 & 0.38@xmath199 0.01 & 12.4 & 4.6 & 9.4 & & 0.1 & 1 + 42 & 217.27429 & -0.36459 & 1.1 & & & 0.78@xmath199 0.09 & 0.59@xmath199 0.01 & 0.57@xmath199 0.02 & 31.7 & 6.5 & 9.5 & & 0.2 & 1 + 43 & 217.29121 & -1.31510 & 5.1 & 0.23@xmath199 0.04 & 0.63@xmath199 0.09 & 0.86@xmath199 0.15 & 0.51@xmath199 0.01 & 0.47@xmath199 0.01 & 10.4 & 44.8 & 12.3 & 0.02 & 19.0 & 2 + 44 & 217.30083 & -1.12665 & 2.5 & & & 1.28@xmath199 0.11 & 0.75@xmath199 0.03 & 0.47@xmath199 0.05 & 33.3 & 14.7 & 12.3 & & 1.1 & 1 + 45 & 217.30141 & -0.05822 & 2.1 & 27.95@xmath199 0.20 & 36.45@xmath199 0.22 & 31.76@xmath199 0.25 & 15.53@xmath199 0.02 & 6.53@xmath199 0.01 & 21.6 & 84.3 & 16.3 & 9.90 & 344.0 & 2 + 46 & 217.30147 & -1.32625 & 5.1 & & 0.82@xmath199 0.08 & 1.20@xmath199 0.09 & 1.05@xmath199 0.01 & 0.84@xmath199 0.01 & 16.9 & 133.9 & 10.9 & 0.01 & 20.3 & 2 + 47 & 217.30151 & -1.78669 & 2.8 & & & 0.47@xmath199 0.05 & 0.52@xmath199 0.02 & 0.31@xmath199 0.02 & 19.8 & 28.4 & 9.6 & & 0.8 & 1 + 48 & 217.30252 & -0.03870 & 2.1 & & 4.39@xmath199 0.17 & 9.50@xmath199 0.28 & 7.82@xmath199 0.02 & 4.27@xmath199 0.01 & 19.9 & 163.9 & 10.7 & 0.32 & 28.7 & 2 + 49 & 217.30620 & -0.04772 & 2.1 & 15.24@xmath199 0.14 & 22.90@xmath199 0.21 & 21.77@xmath199 0.35 & 17.66@xmath199 0.02 & 6.64@xmath199 0.02 & 24.8 & 128.7 & 13.9 & 0.31 & 127.6 & 2 + 50 & 217.31033 & -0.21533 & 1.1 & & & 0.36@xmath199 0.07 & 0.37@xmath199 0.02 & 0.25@xmath199 0.03 & 25.5 & 3.6 & 9.5 & & 0.1 & 1 + 51 & 217.31052 & -0.28071 & 1.1 & & & 2.28@xmath199 0.10 & 1.58@xmath199 0.01 & 0.94@xmath199 0.03 & 26.6 & 7.2 & 11.6 & & 0.4 & 1 + 52 & 217.31207 & -1.38271 & 5.1 & 2.12@xmath199 0.11 & 5.52@xmath199 0.26 & 6.61@xmath199 0.17 & 3.78@xmath199 0.01 & & 20.8 & 187.4 & 14.0 & 0.28 & 169.7 & 2 + 53 & 217.31659 & -1.37286 & 5.1 & & & 10.10@xmath199 0.20 & 6.08@xmath199 0.01 & 2.74@xmath199 0.02 & 24.2 & 240.3 & 14.8 & & 33.0 & 1 + 54 & 217.31665 & -1.33835 & 5.1 & & 4.47@xmath199 0.22 & 5.46@xmath199 0.16 & 3.13@xmath199 0.02 & & 24.7 & 162.2 & 13.8 & 0.08 & 99.1 & 2 + 55 & 217.31862 & -0.30418 & 1.1 & & & 0.93@xmath199 0.07 & 0.73@xmath199 0.01 & 0.56@xmath199 0.01 & 21.6 & 5.4 & 10.1 & & 0.2 & 1 + 56 & 217.32750 & -1.36968 & 5.1 & 8.43@xmath199 0.14 & 18.17@xmath199 0.33 & 21.88@xmath199 0.20 & 13.75@xmath199 0.01 & 6.35@xmath199 0.01 & 16.9 & 728.5 & 13.6 & 0.42 & 553.9 & 2 + 57 & 217.33861 & -0.23874 & 1.1 & & & 0.71@xmath199 0.09 & 0.58@xmath199 0.02 & 0.32@xmath199 0.03 & 32.4 & 2.9 & 11.1 & & 0.1 & 1 + 58 & 217.33931 & -1.36862 & 5.1 & 1.10@xmath199 0.05 & & 10.16@xmath199 0.19 & 7.94@xmath199 0.01 & & 22.7 & 585.4 & 12.1 & 0.32 & 158.1 & 2 + 59 & 217.34460 & -1.37803 & 5.1 & 4.88@xmath199 0.13 & 5.39@xmath199 0.16 & 4.23@xmath199 0.09 & 1.93@xmath199 0.01 & & 13.3 & 49.9 & 17.7 & 0.23 & 217.6 & 2 + 60 & 217.35243 & -1.43530 & 4.7 & & & 3.10@xmath199 0.34 & 1.68@xmath199 33.58 & 0.71@xmath199 0.04 & 32.9 & 49.1 & 15.7 & & 7.9 & 1 + 61 & 217.35413 & -1.41616 & 5.1 & & & 1.82@xmath199 0.09 & 1.24@xmath199 0.01 & 0.64@xmath199 0.01 & 22.6 & 92.5 & 12.3 & & 6.7 & 1 + 62 & 217.35962 & -1.32541 & 5.1 & 0.43@xmath199 0.03 & 0.51@xmath199 0.07 & 0.64@xmath199 0.11 & 0.40@xmath199 0.01 & & 8.4 & 22.7 & 13.3 & 0.02 & 21.1 & 2 + 63 & 217.36540 & -1.33517 & 5.1 & 0.56@xmath199 0.03 & 0.84@xmath199 0.07 & 1.62@xmath199 0.12 & 1.09@xmath199 0.02 & & 20.4 & 103.2 & 11.6 & 0.02 & 32.7 & 2 + 64 & 217.37611 & -0.08454 & 2.1 & 827.64@xmath199 1.66 & 319.99@xmath199 0.64 & 222.51@xmath199 0.89 & 87.32@xmath199 0.00 & 33.92@xmath199 0.00 & 25.4 & 310.9 & 19.3 & 474.50 & 9679.4 & 2 + 65 & 217.47652 & -0.04724 & 1.8 & 2.61@xmath199 0.05 & 3.76@xmath199 0.13 & 3.39@xmath199 0.18 & 2.49@xmath199 0.03 & 0.75@xmath199 0.03 & 10.7 & 8.9 & 15.4 & 0.88 & 23.9 & 2 + 66 & 217.47820 & -0.18994 & 2.1 & & & 0.81@xmath199 0.07 & 0.68@xmath199 0.01 & 0.55@xmath199 0.02 & 21.1 & 25.0 & 9.6 & & 0.7 & 1 + 67 & 217.49167 & -0.08088 & 1.8 & 1.27@xmath199 0.05 & 2.48@xmath199 0.12 & 5.59@xmath199 0.20 & 3.20@xmath199 0.01 & & 17.1 & 38.8 & 11.6 & & 8.7 & 2 + 68 & 217.49518 & -0.07074 & 1.8 & 8.64@xmath199 0.11 & 17.84@xmath199 0.27 & 17.70@xmath199 0.18 & 10.09@xmath199 0.01 & 4.40@xmath199 0.03 & 21.4 & 51.2 & 14.8 & 1.07 & 73.8 & 2 + 69 & 217.50027 & -0.20677 & 2.1 & & & 0.68@xmath199 0.08 & 0.86@xmath199 0.02 & 0.44@xmath199 0.01 & 19.5 & 24.2 & 9.5 & & 0.6 & 1 + 70 & 217.52646 & -0.07732 & 1.8 & & & 1.94@xmath199 0.18 & 1.86@xmath199 0.02 & 1.11@xmath199 0.03 & 25.1 & 36.3 & 10.1 & & 1.2 & 1 + 71 & 217.57440 & -0.09783 & 1.8 & & & 1.64@xmath199 0.16 & 1.05@xmath199 0.01 & 0.56@xmath199 0.04 & 28.2 & 9.4 & 12.6 & & 0.7 & 1 + 72 & 217.60297 & -0.05246 & 1.8 & & & 0.98@xmath199 0.15 & 0.51@xmath199 0.01 & 0.28@xmath199 0.02 & 24.8 & 3.4 & 14.0 & & 0.4 & 1 + 73 & 217.62495 & -0.15439 & 2.1 & 2.46@xmath199 0.22 & 3.77@xmath199 0.12 & 7.61@xmath199 0.27 & 3.48@xmath199 0.03 & 1.36@xmath199 0.03 & 28.1 & 26.4 & 13.6 & 0.22 & 26.3 & 2 + 74 & 217.62689 & -0.18399 & 2.1 & 49.31@xmath199 0.30 & 40.47@xmath199 0.24 & 35.20@xmath199 0.28 & 14.31@xmath199 0.01 & 5.91@xmath199 0.02 & 24.8 & 64.3 & 17.5 & 11.56 & 461.0 & 2 + 75 & 217.63626 & -0.13185 & 2.1 & & & 0.55@xmath199 0.12 & 0.65@xmath199 0.01 & 0.34@xmath199 0.01 & 13.9 & 17.4 & 9.7 & & 0.5 & 1 + 76 & 217.65610 & -0.19365 & 2.1 & 9.23@xmath199 0.15 & 10.97@xmath199 0.19 & 4.72@xmath199 0.12 & 1.99@xmath199 0.01 & 0.87@xmath199 0.02 & 18.4 & 4.9 & 22.7 & 0.64 & 69.8 & 2 + 77 & 217.69716 & -0.23935 & 1.6 & & & 1.30@xmath199 0.11 & 0.97@xmath199 0.01 & 0.57@xmath199 0.02 & 30.9 & 10.4 & 11.3 & & 0.5 & 1 + 78 & 217.69861 & -0.20510 & 2.1 & & & 1.64@xmath199 0.19 & 1.09@xmath199 0.01 & 0.66@xmath199 0.03 & 28.5 & 16.8 & 11.8 & & 1.0 & 1 + 79 & 217.72137 & -0.16784 & 2.1 & & & 0.58@xmath199 0.09 & 0.32@xmath199 0.01 & 0.15@xmath199 0.01 & 16.5 & 2.3 & 14.6 & & 0.3 & 0 + 80 & 217.73511 & -0.69747 & 2.3 & & & 0.33@xmath199 0.08 & 0.31@xmath199 0.02 & 0.20@xmath199 0.01 & 19.5 & 11.1 & 9.9 & & 0.3 & 1 + 81 & 217.76306 & -1.90009 & 2.9 & & & 0.86@xmath199 0.05 & 0.65@xmath199 0.01 & 0.47@xmath199 0.01 & 21.1 & 31.0 & 10.4 & & 1.2 & 1 + 82 & 217.79570 & -1.88198 & 2.9 & 4.20@xmath199 0.06 & 3.75@xmath199 0.09 & 3.09@xmath199 0.19 & 2.52@xmath199 0.02 & 1.05@xmath199 0.02 & 16.9 & 33.6 & 14.0 & 0.66 & 69.7 & 2 + 83 & 217.79695 & -0.13617 & 2.1 & & & 0.73@xmath199 0.07 & 0.53@xmath199 0.02 & 0.34@xmath199 0.04 & 23.5 & 9.9 & 11.1 & & 0.5 & 1 + 84 & 217.80267 & -0.28371 & 1.6 & & & 0.39@xmath199 0.06 & 0.26@xmath199 0.01 & 0.13@xmath199 0.02 & 15.1 & 1.7 & 12.9 & & 0.1 & 1 + 85 & 217.83006 & -0.18935 & 2.1 & & & 0.37@xmath199 0.09 & 0.53@xmath199 0.01 & 0.34@xmath199 0.02 & 19.1 & 25.0 & 8.6 & & 0.4 & 1 + 86 & 217.84331 & -1.81442 & 2.1 & & & 0.28@xmath199 0.04 & 0.23@xmath199 0.02 & 0.12@xmath199 0.02 & 17.7 & 3.5 & 11.4 & & 0.2 & 1 + 87 & 217.84694 & -0.25950 & 2.1 & & & 1.55@xmath199 0.09 & 1.31@xmath199 0.02 & 0.74@xmath199 0.02 & 21.6 & 25.4 & 10.8 & & 1.1 & 1 + 88 & 217.86577 & -1.90944 & 2.1 & & 0.57@xmath199 0.03 & 0.10@xmath199 0.05 & 0.05@xmath199 0.01 & & 6.3 & 0.0 & 43.8 & 0.01 & 1.3 & 2 + 89 & 217.87485 & -1.83482 & 2.1 & & & 0.28@xmath199 0.04 & 0.25@xmath199 0.01 & 0.12@xmath199 0.03 & 21.0 & 4.3 & 11.1 & & 0.2 & 1 + 90 & 217.88361 & -0.27692 & 2.3 & & & 0.87@xmath199 0.13 & 0.63@xmath199 0.01 & 0.37@xmath199 0.02 & 27.6 & 12.1 & 11.6 & & 0.7 & 1 + 91 & 217.88567 & -0.18413 & 2.3 & & & 1.07@xmath199 0.07 & 0.98@xmath199 0.02 & 0.59@xmath199 0.04 & 21.6 & 28.7 & 10.2 & & 1.0 & 1 + 92 & 217.89963 & -0.29547 & 2.3 & 0.49@xmath199 0.02 & 0.61@xmath199 0.09 & 0.92@xmath199 0.13 & 0.63@xmath199 0.02 & 0.36@xmath199 0.03 & 16.3 & 10.5 & 12.0 & 0.02 & 5.0 & 2 + 93 & 217.91693 & -1.88418 & 2.7 & & & 0.42@xmath199 0.04 & 0.40@xmath199 0.01 & 0.26@xmath199 0.01 & 17.8 & 19.5 & 9.9 & & 0.6 & 1 + 94 & 217.91704 & -0.29507 & 2.3 & & & 0.89@xmath199 0.08 & 0.76@xmath199 0.01 & 0.46@xmath199 0.02 & 18.5 & 20.7 & 10.5 & & 0.8 & 1 + 95 & 217.91763 & -1.81642 & 2.1 & & & 0.27@xmath199 0.05 & 0.31@xmath199 0.02 & 0.24@xmath199 0.02 & 20.7 & 15.4 & 8.7 & & 0.3 & 1 + 96 & 217.92119 & -0.26952 & 2.3 & & & 1.27@xmath199 0.11 & 0.97@xmath199 0.02 & 0.65@xmath199 0.02 & 30.1 & 26.7 & 10.6 & & 1.1 & 1 + 97 & 217.92220 & -0.24550 & 2.3 & & & 2.26@xmath199 0.14 & 1.43@xmath199 0.04 & 0.81@xmath199 0.02 & 36.5 & 22.5 & 12.3 & & 1.6 & 1 + 98 & 217.94304 & -0.28449 & 2.3 & & & 2.19@xmath199 0.12 & 1.72@xmath199 0.02 & 0.95@xmath199 0.04 & 30.0 & 37.1 & 11.1 & & 1.8 & 1 + 99 & 217.95029 & -0.29615 & 2.3 & & & 1.02@xmath199 20.42 & 1.24@xmath199 0.01 & 0.96@xmath199 0.01 & 18.2 & 81.7 & 8.6 & & 1.3 & 1 + 100 & 217.95697 & -0.49270 & 2.1 & & & 1.11@xmath199 0.54 & 0.67@xmath199 0.01 & 0.39@xmath199 0.02 & 23.9 & 8.7 & 12.5 & & 0.6 & 1 + 101 & 217.95921 & -0.12667 & 2.1 & & & 0.52@xmath199 0.05 & 0.29@xmath199 0.02 & 0.21@xmath199 0.02 & 20.5 & 4.6 & 12.1 & & 0.3 & 1 + 102 & 217.96873 & -1.82791 & 2.7 & & & 0.21@xmath199 0.06 & 0.19@xmath199 0.02 & 0.10@xmath199 0.02 & 15.0 & 5.2 & 11.1 & & 0.3 & 1 + 103 & 217.99861 & -0.31560 & 2.3 & & & 2.82@xmath199 0.23 & 5.10@xmath199 0.01 & 3.36@xmath199 0.02 & 28.1 & 315.2 & 8.3 & & 4.5 & 1 + 104 & 218.01694 & -1.95616 & 2.0 & & & 0.58@xmath199 0.05 & 0.47@xmath199 0.01 & 0.24@xmath199 0.03 & 17.7 & 6.5 & 11.4 & & 0.3 & 1 + 105 & 218.01711 & -1.87673 & 2.7 & & & 0.46@xmath199 0.09 & 0.30@xmath199 0.03 & 0.15@xmath199 0.02 & 23.1 & 5.4 & 12.8 & & 0.5 & 1 + 106 & 218.01738 & -0.48248 & 1.5 & & & 0.94@xmath199 0.06 & 0.64@xmath199 0.02 & 0.32@xmath199 0.02 & 17.3 & 4.1 & 12.5 & & 0.3 & 1 + 107 & 218.02229 & -0.31524 & 2.3 & 29.95@xmath199 0.21 & 34.01@xmath199 0.27 & 28.04@xmath199 0.14 & 13.74@xmath199 0.01 & 5.63@xmath199 0.01 & 18.1 & 79.5 & 16.7 & 2.90 & 299.4 & 2 + 108 & 218.03098 & -0.26996 & 2.3 & & & 1.45@xmath199 0.11 & 1.03@xmath199 0.02 & 0.55@xmath199 0.03 & 28.5 & 16.5 & 12.0 & & 1.1 & 1 + 109 & 218.03860 & -1.81585 & 2.7 & & & 0.17@xmath199 0.07 & 0.25@xmath199 0.01 & 0.18@xmath199 0.01 & 15.1 & 27.9 & 8.1 & & 0.3 & 1 + 110 & 218.04095 & -0.32733 & 2.3 & 2.90@xmath199 0.08 & 6.37@xmath199 0.15 & 8.71@xmath199 0.15 & 6.04@xmath199 0.01 & 3.52@xmath199 0.01 & 19.4 & 91.1 & 12.3 & 0.12 & 39.0 & 2 + 111 & 218.04378 & -0.48542 & 2.0 & & & 0.69@xmath199 0.11 & 0.55@xmath199 0.01 & 0.28@xmath199 0.01 & 21.1 & 7.4 & 11.6 & & 0.4 & 1 + 112 & 218.05252 & -0.11740 & 2.1 & 28.07@xmath199 0.20 & 45.77@xmath199 0.27 & 39.40@xmath199 0.43 & 20.01@xmath199 0.02 & 7.60@xmath199 0.02 & 17.6 & 104.2 & 16.4 & 3.16 & 289.2 & 2 + 113 & 218.06357 & -0.32522 & 2.3 & & 1.64@xmath199 0.13 & 3.62@xmath199 0.13 & 2.28@xmath199 0.01 & 2.13@xmath199 0.02 & 23.3 & 55.6 & 11.1 & 0.10 & 11.7 & 2 + 114 & 218.07256 & -0.27425 & 2.3 & 0.79@xmath199 0.04 & 0.91@xmath199 0.06 & 1.26@xmath199 0.10 & 0.71@xmath199 0.01 & & 13.9 & 8.2 & 13.3 & 0.01 & 6.8 & 2 + 115 & 218.07767 & -0.28704 & 2.3 & & 0.68@xmath199 0.07 & 1.69@xmath199 0.14 & 1.54@xmath199 0.01 & 0.78@xmath199 0.02 & 18.1 & 40.4 & 10.3 & 0.02 & 4.5 & 2 + 116 & 218.08118 & -0.30117 & 2.3 & 1.17@xmath199 0.03 & 1.02@xmath199 0.17 & 3.43@xmath199 0.19 & 2.71@xmath199 0.02 & 1.63@xmath199 0.04 & 15.0 & 101.7 & 9.7 & 0.11 & 14.2 & 2 + 117 & 218.08861 & -0.05262 & 2.1 & & & 0.69@xmath199 0.05 & 0.42@xmath199 0.01 & 0.30@xmath199 0.02 & 20.2 & 7.7 & 11.6 & & 0.4 & 1 + 118 & 218.10060 & -1.87255 & 2.2 & & & 0.63@xmath199 0.21 & 0.45@xmath199 0.02 & 0.28@xmath199 0.01 & 23.2 & 9.8 & 11.1 & & 0.5 & 1 + 119 & 218.10332 & -0.36435 & 1.5 & 42.69@xmath199 0.26 & 59.18@xmath199 0.36 & 56.90@xmath199 0.40 & 28.07@xmath199 0.03 & 10.93@xmath199 0.02 & 20.6 & 81.8 & 16.0 & 5.45 & 225.0 & 2 + 120 & 218.10861 & -0.03431 & 2.1 & & & 0.38@xmath199 0.04 & 0.25@xmath199 0.01 & 0.18@xmath199 0.02 & 15.1 & 5.4 & 11.1 & & 0.3 & 1 + 121 & 218.11208 & -0.29306 & 2.3 & 0.89@xmath199 0.05 & 2.37@xmath199 0.11 & 4.02@xmath199 0.11 & 2.54@xmath199 0.03 & 1.65@xmath199 0.03 & 21.0 & 49.8 & 11.7 & 0.35 & 18.9 & 2 + 122 & 218.11674 & -0.12198 & 2.1 & & & 1.10@xmath199 0.06 & 0.61@xmath199 0.01 & 0.38@xmath199 0.02 & 21.9 & 7.8 & 12.7 & & 0.6 & 1 + 123 & 218.12117 & -0.31474 & 1.5 & 0.56@xmath199 0.03 & 1.16@xmath199 0.07 & 4.36@xmath199 0.15 & 4.43@xmath199 0.01 & 2.71@xmath199 0.02 & 25.2 & 93.5 & 9.1 & & 4.5 & 2 + 124 & 218.12663 & -0.27104 & 2.3 & & & 1.16@xmath199 0.14 & 0.84@xmath199 0.02 & 0.49@xmath199 0.03 & 21.7 & 16.8 & 11.4 & & 0.9 & 1 + 125 & 218.12708 & -1.74818 & 2.2 & & 0.40@xmath199 0.05 & 0.32@xmath199 0.09 & 0.17@xmath199 0.01 & & 12.9 & 0.9 & 16.5 & 0.01 & 1.5 & 2 + 126 & 218.13034 & -0.30450 & 1.5 & & & 3.32@xmath199 0.25 & 2.89@xmath199 0.01 & 1.53@xmath199 0.02 & 24.2 & 28.4 & 11.0 & & 1.3 & 1 + 127 & 218.13336 & -0.37784 & 2.3 & & & 5.26@xmath199 0.25 & 3.45@xmath199 0.02 & 1.75@xmath199 0.04 & 30.1 & 45.3 & 12.8 & & 3.7 & 1 + 128 & 218.13341 & -0.31053 & 1.5 & 0.40@xmath199 0.04 & 0.98@xmath199 0.13 & 2.36@xmath199 0.19 & 1.32@xmath199 0.01 & & 17.7 & 13.0 & 11.3 & 0.06 & 3.4 & 2 + 129 & 218.13341 & -0.51962 & 2.0 & 0.87@xmath199 0.04 & 2.39@xmath199 0.12 & 3.41@xmath199 0.15 & 2.19@xmath199 0.01 & 0.86@xmath199 0.01 & 13.9 & 17.0 & 13.3 & 0.18 & 11.9 & 2 + 130 & 218.13423 & -1.43585 & 7.4 & & & 0.54@xmath199 0.04 & 0.42@xmath199 0.02 & 0.17@xmath199 0.17 & 24.8 & 48.7 & 13.0 & & 4.3 & 1 + 131 & 218.13649 & -0.53308 & 2.0 & 1.57@xmath199 0.05 & 1.52@xmath199 0.10 & 0.95@xmath199 0.10 & 0.55@xmath199 0.01 & & 3.8 & 1.8 & 18.0 & 0.01 & 8.0 & 2 + 132 & 218.14186 & -0.52209 & 2.0 & 0.46@xmath199 0.03 & 0.45@xmath199 0.07 & 1.83@xmath199 0.16 & 1.97@xmath199 0.01 & 1.06@xmath199 0.01 & 17.4 & 66.3 & 9.0 & 0.06 & 5.2 & 2 + 133 & 218.14290 & -1.89307 & 2.2 & & & 0.67@xmath199 0.07 & 0.54@xmath199 0.03 & 0.34@xmath199 0.04 & 28.6 & 13.5 & 10.6 & & 0.5 & 1 + 134 & 218.14362 & -0.28641 & 1.5 & & & 1.91@xmath199 0.32 & 1.37@xmath199 0.01 & 1.35@xmath199 0.02 & 25.5 & 28.1 & 9.7 & & 0.8 & 1 + 135 & 218.14444 & -0.18696 & 2.1 & & & 1.49@xmath199 0.06 & 1.12@xmath199 0.01 & 0.73@xmath199 0.01 & 17.7 & 25.1 & 10.8 & & 1.1 & 1 + 136 & 218.14633 & -0.56713 & 2.0 & 7.83@xmath199 0.11 & 10.85@xmath199 0.17 & 11.46@xmath199 0.25 & 7.11@xmath199 0.02 & 3.54@xmath199 0.01 & 20.5 & 54.0 & 13.9 & 3.90 & 96.6 & 2 + 137 & 218.14888 & -0.33179 & 1.5 & & & 2.10@xmath199 0.21 & 1.56@xmath199 0.01 & 1.32@xmath199 0.21 & 19.5 & 25.1 & 10.1 & & 0.8 & 1 + 138 & 218.15030 & -0.64382 & 2.0 & 15.22@xmath199 0.17 & 16.92@xmath199 0.24 & 16.67@xmath199 0.17 & 8.73@xmath199 0.01 & 4.08@xmath199 0.03 & 21.0 & 52.7 & 15.1 & 1.31 & 113.8 & 2 + 139 & 218.15474 & -0.40847 & 2.0 & 1.06@xmath199 0.04 & 1.52@xmath199 0.10 & 2.43@xmath199 0.10 & 1.38@xmath199 0.02 & & 20.6 & 14.5 & 12.6 & 0.19 & 10.7 & 2 + 140 & 218.15652 & -0.58218 & 2.0 & 8.55@xmath199 0.09 & 11.95@xmath199 0.20 & 11.13@xmath199 0.65 & 6.66@xmath199 0.02 & 3.20@xmath199 0.03 & 30.5 & 43.4 & 14.6 & 0.61 & 67.5 & 2 + 141 & 218.15736 & -0.37530 & 2.3 & & 2.04@xmath199 0.09 & 2.50@xmath199 0.19 & 1.60@xmath199 0.01 & 1.44@xmath199 0.01 & 18.2 & 27.0 & 12.3 & 0.03 & 9.0 & 2 + 142 & 218.15930 & -0.18826 & 2.1 & & & 0.37@xmath199 0.08 & 0.40@xmath199 0.01 & 0.24@xmath199 0.01 & 15.6 & 12.4 & 9.7 & & 0.3 & 1 + 143 & 218.16008 & -0.34305 & 1.5 & 0.59@xmath199 0.04 & 1.53@xmath199 0.10 & 4.13@xmath199 0.17 & 2.92@xmath199 0.01 & & 19.4 & 39.7 & 10.4 & 0.02 & 5.0 & 2 + 144 & 218.16364 & -0.35937 & 1.5 & 0.55@xmath199 0.04 & 1.83@xmath199 0.09 & 6.03@xmath199 0.20 & 5.87@xmath199 0.01 & 3.29@xmath199 0.02 & 22.8 & 99.0 & 9.5 & 0.24 & 8.2 & 2 + 145 & 218.16907 & -1.84870 & 2.2 & & & 0.45@xmath199 0.08 & 0.41@xmath199 0.02 & 0.27@xmath199 0.03 & 19.3 & 12.3 & 10.1 & & 0.4 & 1 + 146 & 218.17120 & -0.67295 & 2.0 & & & 8.16@xmath199 0.27 & 8.68@xmath199 0.02 & 4.78@xmath199 0.03 & 21.1 & 194.5 & 10.0 & & 6.2 & 1 + 147 & 218.17444 & -0.63119 & 2.0 & & & 2.76@xmath199 0.12 & 3.03@xmath199 0.02 & 1.69@xmath199 0.01 & 19.2 & 74.2 & 9.8 & & 2.1 & 1 + 148 & 218.17519 & -0.37541 & 2.3 & 2.01@xmath199 0.05 & 7.74@xmath199 0.23 & 9.24@xmath199 0.21 & 4.66@xmath199 0.02 & 4.09@xmath199 0.02 & 28.5 & 63.2 & 13.3 & 0.10 & 36.2 & 2 + 149 & 218.17607 & -0.50686 & 2.0 & 0.79@xmath199 0.04 & 1.59@xmath199 0.10 & 3.28@xmath199 0.15 & 2.60@xmath199 0.02 & 1.29@xmath199 0.01 & 19.1 & 39.1 & 11.1 & 0.10 & 9.7 & 2 + 150 & 218.18240 & -0.58864 & 2.0 & 0.91@xmath199 0.05 & 1.32@xmath199 0.09 & 1.57@xmath199 0.15 & 0.92@xmath199 0.01 & & 16.6 & 6.8 & 13.9 & 0.09 & 8.0 & 2 + 151 & 218.18529 & -0.43450 & 2.0 & & & 0.58@xmath199 0.07 & 0.70@xmath199 0.01 & 0.39@xmath199 0.03 & 19.5 & 19.3 & 9.4 & & 0.5 & 1 + 152 & 218.19066 & -0.38409 & 2.3 & 1.10@xmath199 0.09 & 6.89@xmath199 0.13 & 13.48@xmath199 0.43 & 8.96@xmath199 0.03 & 3.99@xmath199 0.02 & 16.9 & 139.4 & 11.9 & 0.42 & 38.8 & 2 + 153 & 218.19389 & -0.60201 & 2.0 & 2.36@xmath199 0.06 & 3.64@xmath199 0.14 & 3.92@xmath199 0.22 & 2.14@xmath199 0.01 & 0.92@xmath199 0.02 & 14.2 & 12.7 & 14.8 & 0.12 & 18.8 & 2 + 154 & 218.19931 & -0.39083 & 2.0 & 17.48@xmath199 0.16 & 20.38@xmath199 0.14 & 13.10@xmath199 0.42 & 5.41@xmath199 0.02 & & 25.5 & 13.1 & 20.5 & 1.92 & 131.9 & 2 + 155 & 218.20279 & -0.31413 & 1.5 & & 0.49@xmath199 0.07 & 1.18@xmath199 0.07 & 1.26@xmath199 0.02 & 0.84@xmath199 0.02 & 20.9 & 22.2 & 9.6 & & 0.9 & 1 + 156 & 218.22362 & -0.39818 & 2.0 & & 2.10@xmath199 0.08 & 2.87@xmath199 0.16 & 1.66@xmath199 0.02 & & 23.8 & 14.5 & 13.3 & 0.11 & 9.0 & 2 + 157 & 218.22694 & -2.00283 & 2.0 & & & 0.27@xmath199 0.06 & 0.42@xmath199 0.01 & 0.24@xmath199 0.02 & 19.2 & 17.1 & 8.6 & & 0.3 & 1 + 158 & 218.23599 & -2.02763 & 2.8 & & & 1.15@xmath199 0.06 & 1.01@xmath199 0.02 & 0.62@xmath199 0.02 & 24.5 & 45.0 & 10.4 & & 1.6 & 1 + 159 & 218.23843 & -0.30095 & 2.1 & & 1.41@xmath199 0.07 & 1.71@xmath199 0.09 & 1.18@xmath199 0.01 & 0.73@xmath199 0.02 & 23.0 & 15.2 & 12.6 & 0.05 & 6.4 & 2 + 160 & 218.24062 & -0.39514 & 2.1 & & 0.87@xmath199 0.09 & 1.54@xmath199 0.09 & 0.87@xmath199 0.01 & & 22.8 & 11.7 & 12.3 & 0.02 & 4.0 & 2 + 161 & 218.24518 & -0.60239 & 2.0 & & & 3.19@xmath199 0.18 & 2.32@xmath199 0.03 & 1.38@xmath199 0.03 & 26.2 & 27.0 & 12.1 & & 1.9 & 1 + 162 & 218.25404 & -2.04922 & 2.8 & & 0.60@xmath199 0.04 & 1.05@xmath199 0.09 & 1.01@xmath199 0.02 & 0.60@xmath199 0.01 & 18.7 & 38.8 & 10.6 & & 2.8 & 1 + 163 & 218.25458 & -0.35242 & 2.1 & & 1.58@xmath199 0.06 & 2.08@xmath199 0.23 & 1.36@xmath199 0.02 & 0.70@xmath199 0.03 & 24.2 & 15.7 & 12.8 & 0.04 & 6.6 & 2 + 164 & 218.26382 & -0.39015 & 2.1 & & & 2.13@xmath199 0.16 & 1.21@xmath199 0.02 & 0.58@xmath199 0.03 & 27.7 & 9.9 & 14.3 & & 1.2 & 1 + 165 & 218.28824 & -2.01099 & 2.0 & & & 0.93@xmath199 0.09 & 1.01@xmath199 0.01 & 0.70@xmath199 0.03 & 22.5 & 39.0 & 9.1 & & 0.8 & 1 + 166 & 218.29642 & -0.80482 & 2.3 & & & 0.31@xmath199 0.05 & 0.27@xmath199 0.01 & 0.13@xmath199 0.03 & 19.4 & 5.0 & 11.4 & & 0.3 & 1 + 167 & 218.29813 & -1.44999 & 7.4 & & & 0.25@xmath199 0.05 & 0.28@xmath199 0.02 & 0.17@xmath199 0.01 & 14.8 & 113.6 & 9.5 & & 2.9 & 1 + 168 & 218.30522 & -0.11718 & 2.2 & & & 1.04@xmath199 0.08 & 0.88@xmath199 0.02 & 0.70@xmath199 0.02 & 28.4 & 36.6 & 9.5 & & 0.9 & 1 + 169 & 218.30528 & -0.37284 & 2.1 & & & 0.56@xmath199 0.07 & 0.32@xmath199 0.02 & 0.18@xmath199 0.02 & 22.0 & 3.6 & 13.1 & & 0.3 & 1 + 170 & 218.31230 & -0.44349 & 2.1 & & & 1.25@xmath199 0.11 & 0.75@xmath199 0.02 & 0.39@xmath199 0.02 & 21.5 & 7.9 & 13.3 & & 0.7 & 1 + 171 & 218.31528 & -0.08900 & 2.2 & & & 1.07@xmath199 0.11 & 0.74@xmath199 0.02 & 0.56@xmath199 0.04 & 25.3 & 20.1 & 10.6 & & 0.8 & 1 + 172 & 218.33670 & -0.07875 & 2.2 & & & 1.77@xmath199 0.09 & 1.42@xmath199 0.03 & 0.91@xmath199 0.04 & 27.9 & 36.1 & 10.6 & & 1.4 & 1 + 173 & 218.33705 & -2.02509 & 2.0 & & & 0.65@xmath199 0.04 & 0.58@xmath199 0.01 & 0.42@xmath199 0.05 & 20.6 & 17.6 & 9.7 & & 0.5 & 1 + 174 & 218.33766 & -2.04468 & 2.1 & & & 0.41@xmath199 0.04 & 0.37@xmath199 0.01 & 0.25@xmath199 0.05 & 24.1 & 10.6 & 9.9 & & 0.3 & 1 + 175 & 218.35033 & -2.03794 & 2.1 & 0.43@xmath199 0.02 & & 0.74@xmath199 0.09 & 0.49@xmath199 0.01 & 0.24@xmath199 0.04 & 9.7 & 5.2 & 12.8 & 0.04 & 4.2 & 2 + 176 & 218.36565 & -1.11563 & 2.7 & & & 0.76@xmath199 1.12 & 0.71@xmath199 0.02 & 0.43@xmath199 0.03 & 25.1 & 31.1 & 10.1 & & 1.0 & 1 + 177 & 218.39359 & -1.44219 & 7.5 & & & 0.24@xmath199 0.06 & 0.32@xmath199 0.01 & 0.25@xmath199 0.02 & 21.0 & 265.2 & 8.3 & & 3.5 & 1 + 178 & 218.39362 & -0.18950 & 2.2 & & & 0.38@xmath199 0.10 & 0.30@xmath199 0.02 & 0.21@xmath199 0.02 & 28.0 & 8.4 & 10.4 & & 0.3 & 1 + 179 & 218.39891 & -2.06174 & 2.1 & & & 0.57@xmath199 0.05 & 0.44@xmath199 0.01 & 0.20@xmath199 0.02 & 15.8 & 5.0 & 12.3 & & 0.4 & 1 + 180 & 218.40973 & -1.12670 & 2.7 & & & 0.43@xmath199 0.04 & 0.34@xmath199 0.01 & 0.20@xmath199 0.02 & 16.8 & 11.6 & 10.9 & & 0.5 & 1 + 181 & 218.46768 & -1.97681 & 2.1 & & & 0.59@xmath199 0.07 & 0.60@xmath199 0.03 & 0.38@xmath199 0.02 & 25.1 & 18.3 & 9.7 & & 0.5 & 1 + 182 & 218.55058 & -0.93413 & 1.6 & & & 0.20@xmath199 0.04 & 0.16@xmath199 0.01 & 0.11@xmath199 0.01 & 21.9 & 1.8 & 11.1 & & 0.1 & 0 + 183 & 218.56168 & -1.90193 & 2.7 & & 0.60@xmath199 0.04 & 0.32@xmath199 0.05 & 0.16@xmath199 0.02 & & 15.0 & 0.7 & 20.5 & 0.01 & 3.0 & 2 + 184 & 218.58128 & -1.44798 & 2.4 & & & 0.29@xmath199 0.04 & 0.31@xmath199 0.02 & 0.24@xmath199 0.03 & 21.0 & 18.9 & 8.9 & & 0.4 & 1 + 185 & 218.68935 & -0.64691 & 2.5 & & & 0.42@xmath199 0.04 & 0.31@xmath199 0.03 & 0.10@xmath199 0.03 & 24.9 & 2.5 & 14.7 & & 0.3 & 0 + 186 & 218.69077 & -0.58594 & 2.1 & & & 0.19@xmath199 0.03 & 0.17@xmath199 0.01 & 0.11@xmath199 0.01 & 17.8 & 4.9 & 10.0 & & 0.2 & 1 + 187 & 218.70361 & -0.81409 & 2.5 & & & 0.54@xmath199 0.07 & 0.45@xmath199 0.02 & 0.28@xmath199 0.02 & 18.6 & 14.2 & 10.6 & & 0.6 & 1 + 188 & 218.70598 & -0.49142 & 5.8 & & & 0.55@xmath199 0.06 & 0.60@xmath199 0.01 & 0.38@xmath199 0.03 & 26.5 & 152.0 & 9.5 & & 3.8 & 1 + 189 & 218.73402 & -0.45868 & 5.8 & & 0.37@xmath199 0.04 & 0.88@xmath199 0.06 & 0.81@xmath199 0.02 & 0.61@xmath199 0.02 & 19.7 & 203.7 & 9.8 & 0.01 & 17.4 & 2 + 190 & 218.74429 & -0.44515 & 5.8 & & & 0.47@xmath199 0.06 & 0.43@xmath199 0.01 & 0.23@xmath199 0.01 & 11.2 & 55.6 & 11.1 & & 2.7 & 1 + 191 & 218.76701 & -1.78701 & 2.9 & & & 1.11@xmath199 0.06 & 1.28@xmath199 0.01 & 0.58@xmath199 0.05 & 22.6 & 48.5 & 10.4 & & 1.8 & 1 + 192 & 218.76704 & -0.42141 & 5.8 & & & 0.34@xmath199 0.05 & 0.40@xmath199 0.02 & 0.30@xmath199 0.05 & 24.1 & 152.1 & 8.8 & & 2.7 & 1 + 193 & 218.78169 & -0.44537 & 5.8 & & & 0.19@xmath199 0.09 & 0.18@xmath199 0.01 & 0.16@xmath199 0.02 & 16.1 & 71.7 & 8.9 & & 1.3 & 1 + 194 & 218.80585 & -0.38295 & 1.6 & & & 0.50@xmath199 0.10 & 0.41@xmath199 0.01 & 0.24@xmath199 0.01 & 15.8 & 5.1 & 10.8 & & 0.2 & 1 + 195 & 218.84952 & -1.55865 & 4.3 & & & 0.17@xmath199 0.02 & 0.16@xmath199 0.01 & 0.11@xmath199 0.01 & 15.1 & 19.7 & 9.9 & & 0.6 & 1 + 196 & 218.91626 & -1.72617 & 2.6 & & & 0.54@xmath199 0.03 & 0.34@xmath199 0.02 & 0.28@xmath199 0.01 & 21.6 & 12.4 & 10.8 & & 0.5 & 1 + 197 & 218.98193 & -1.28618 & 2.9 & & & 0.11@xmath199 0.04 & 0.07@xmath199 0.01 & 0.06@xmath199 0.02 & 15.4 & 3.6 & 10.5 & & 0.1 & 1 + 198 & 219.22278 & -1.08200 & 1.9 & & & 0.44@xmath199 0.05 & 0.43@xmath199 0.01 & 0.25@xmath199 0.02 & 25.4 & 8.9 & 10.2 & & 0.3 & 1 + 199 & 219.23773 & -1.11610 & 1.9 & & & 0.69@xmath199 0.05 & 0.38@xmath199 0.07 & 0.37@xmath199 0.02 & 26.0 & 7.2 & 11.1 & & 0.4 & 1 + 200 & 219.31848 & -1.08639 & 1.2 & & & 0.72@xmath199 0.06 & 0.37@xmath199 0.10 & 0.37@xmath199 0.02 & 20.2 & 2.3 & 11.5 & & 0.1 & 1 + 201 & 219.33624 & -1.12155 & 1.2 & & & 0.96@xmath199 0.11 & 0.88@xmath199 0.01 & 0.78@xmath199 0.02 & 23.7 & 13.0 & 9.1 & & 0.3 & 1 + 202 & 219.33771 & -1.10507 & 1.2 & & & 0.92@xmath199 0.05 & 0.94@xmath199 0.02 & 0.63@xmath199 0.01 & 17.9 & 10.1 & 9.5 & & 0.3 & 1 + 203 & 219.35236 & -1.13380 & 1.2 & & & 1.31@xmath199 0.16 & 1.35@xmath199 0.01 & 0.84@xmath199 0.02 & 30.1 & 13.0 & 9.7 & & 0.4 & 1 + 204 & 219.40442 & -1.26784 & 2.6 & & & 0.30@xmath199 0.02 & 0.25@xmath199 0.02 & 0.13@xmath199 0.02 & 16.9 & 6.9 & 11.1 & & 0.3 & 1 + 205 & 219.54990 & -0.80885 & 2.6 & & & 0.39@xmath199 0.04 & 0.25@xmath199 0.01 & 0.17@xmath199 0.01 & 17.9 & 7.1 & 11.5 & & 0.4 & 1 + 206 & 219.57233 & -1.57568 & 2.5 & & & 0.37@xmath199 0.05 & 0.29@xmath199 0.01 & 0.25@xmath199 0.02 & 20.3 & 14.8 & 9.7 & & 0.4 & 1 + 207 & 219.66856 & -1.34391 & 2.6 & & & 0.43@xmath199 0.04 & 0.41@xmath199 0.01 & 0.21@xmath199 0.04 & 24.5 & 11.8 & 10.7 & & 0.5 & 1 + 208 & 219.67012 & -1.90204 & 2.1 & & & 0.41@xmath199 0.02 & 0.27@xmath199 0.01 & 0.19@xmath199 0.03 & 21.8 & 5.4 & 11.1 & & 0.3 & 1 + 209 & 219.72528 & -1.30127 & 2.6 & & & 0.31@xmath199 0.06 & 0.28@xmath199 0.02 & 0.23@xmath199 0.03 & 19.9 & 17.2 & 9.3 & & 0.4 & 1 + 210 & 219.81528 & -0.83951 & 1.9 & & & 0.33@xmath199 0.03 & 0.23@xmath199 0.01 & 0.17@xmath199 0.01 & 19.8 & 4.2 & 10.8 & & 0.2 & 1 + 211 & 219.85861 & -0.95960 & 1.9 & & & 0.33@xmath199 0.06 & 0.33@xmath199 0.01 & 0.17@xmath199 0.01 & 20.0 & 6.7 & 10.1 & & 0.2 & 1 + 212 & 219.91673 & -1.95510 & 2.5 & & & 0.44@xmath199 0.11 & 0.33@xmath199 0.04 & 0.26@xmath199 0.02 & 22.0 & 13.1 & 10.2 & & 0.4 & 1 + 213 & 220.08527 & -2.02981 & 2.8 & & & 0.51@xmath199 0.05 & 0.37@xmath199 0.02 & 0.23@xmath199 0.02 & 21.0 & 12.3 & 11.2 & & 0.6 & 1 + 214 & 220.10861 & -0.61273 & 2.0 & & 0.63@xmath199 0.07 & 1.36@xmath199 0.06 & 1.16@xmath199 0.01 & 0.74@xmath199 0.02 & 19.0 & 27.3 & 10.4 & 0.05 & 4.1 & 2 + 215 & 220.15361 & -1.20618 & 2.7 & & & 0.25@xmath199 0.14 & 0.26@xmath199 0.01 & 0.15@xmath199 0.03 & 24.3 & 11.8 & 9.9 & & 0.4 & 1 + 216 & 220.15628 & -0.29484 & 2.4 & & & 0.46@xmath199 0.07 & 0.38@xmath199 0.01 & 0.25@xmath199 0.01 & 11.1 & 11.9 & 10.4 & & 0.4 & 1 + 217 & 220.17960 & -0.36337 & 2.4 & & & 0.71@xmath199 0.04 & 0.46@xmath199 0.01 & 0.32@xmath199 0.01 & 19.2 & 11.0 & 11.3 & & 0.6 & 1 + 218 & 220.18361 & -0.33092 & 2.4 & & & 1.41@xmath199 0.10 & 0.88@xmath199 0.02 & 0.52@xmath199 0.02 & 23.4 & 15.4 & 12.3 & & 1.1 & 1 + 219 & 220.18413 & -1.87562 & 2.7 & & & 0.50@xmath199 0.05 & 0.41@xmath199 0.01 & 0.25@xmath199 0.02 & 19.5 & 14.0 & 10.7 & & 0.6 & 1 + 220 & 220.18965 & -1.20219 & 2.7 & & & 0.40@xmath199 0.04 & 0.28@xmath199 0.02 & 0.14@xmath199 0.08 & 22.3 & 5.7 & 12.5 & & 0.4 & 1 + 221 & 220.20215 & -0.34194 & 2.4 & 1.08@xmath199 0.04 & 0.98@xmath199 0.06 & 1.79@xmath199 0.04 & 1.21@xmath199 0.01 & 0.61@xmath199 0.03 & 19.6 & 22.0 & 11.8 & 0.49 & 17.8 & 2 + 222 & 220.20824 & -0.16897 & 3.0 & & & 0.51@xmath199 0.03 & 0.32@xmath199 0.01 & 0.17@xmath199 0.02 & 25.4 & 7.0 & 13.0 & & 0.6 & 1 + 223 & 220.25204 & -1.91249 & 2.7 & & & 0.58@xmath199 0.07 & 0.43@xmath199 0.02 & 0.21@xmath199 0.01 & 16.0 & 9.8 & 11.8 & & 0.6 & 1 + 224 & 220.27026 & -1.76359 & 0.9 & & & 0.50@xmath199 0.09 & 0.41@xmath199 0.02 & 0.18@xmath199 0.02 & 20.9 & 0.9 & 12.1 & & 0.1 & 0 + 225 & 220.29817 & -1.71690 & 0.9 & & & 1.45@xmath199 0.08 & 1.24@xmath199 0.01 & 0.74@xmath199 0.03 & 26.7 & 5.4 & 10.1 & & 0.2 & 1 + 226 & 220.30891 & -1.78707 & 0.9 & & & 0.50@xmath199 0.23 & 0.84@xmath199 0.03 & 0.49@xmath199 0.02 & 26.4 & 7.2 & 8.4 & & 0.1 & 1 + 227 & 220.31667 & -1.74675 & 0.9 & & & 1.42@xmath199 0.10 & 1.19@xmath199 0.02 & 0.54@xmath199 0.03 & 34.2 & 2.7 & 11.9 & & 0.2 & 1 + 228 & 220.32027 & -1.89012 & 0.9 & & & 0.74@xmath199 0.18 & 0.52@xmath199 0.01 & 0.37@xmath199 0.02 & 25.3 & 1.8 & 11.1 & & 0.1 & 1 + 229 & 220.32063 & -1.95297 & 0.8 & & & 0.40@xmath199 0.07 & 0.30@xmath199 0.02 & 0.16@xmath199 0.03 & 20.9 & 0.6 & 11.9 & & 0.0 & 0 + 230 & 220.32780 & -1.76004 & 0.9 & & & 1.27@xmath199 0.09 & 1.16@xmath199 0.02 & 0.87@xmath199 0.03 & 30.3 & 7.6 & 9.5 & & 0.2 & 1 + 231 & 220.33427 & -1.89749 & 0.9 & & & 1.97@xmath199 0.19 & 1.61@xmath199 0.03 & 0.95@xmath199 0.05 & 33.7 & 5.7 & 10.8 & & 0.2 & 1 + 232 & 220.36496 & -1.96519 & 0.8 & & & 0.71@xmath199 0.09 & 0.75@xmath199 0.03 & 0.34@xmath199 0.03 & 23.1 & 1.9 & 10.7 & & 0.1 & 1 + 233 & 220.37860 & -1.78618 & 0.9 & & & 1.36@xmath199 0.09 & 1.02@xmath199 0.01 & 0.61@xmath199 0.03 & 30.7 & 3.5 & 11.1 & & 0.2 & 1 + 234 & 220.38187 & -1.97019 & 2.8 & & & 0.52@xmath199 0.07 & 0.30@xmath199 0.01 & 0.26@xmath199 0.05 & 18.9 & 11.4 & 11.1 & & 0.6 & 1 + 235 & 220.39813 & -1.90991 & 2.8 & & & 0.34@xmath199 0.04 & 0.33@xmath199 0.01 & 0.18@xmath199 0.01 & 16.8 & 12.7 & 10.4 & & 0.5 & 1 + 236 & 220.42284 & -0.58607 & 2.2 & 0.54@xmath199 0.05 & & 0.79@xmath199 0.08 & 0.43@xmath199 0.01 & & 25.2 & 2.5 & 16.2 & 0.04 & 5.6 & 2 + 237 & 220.45834 & -0.60897 & 2.2 & 22.97@xmath199 0.18 & 26.19@xmath199 0.24 & 27.72@xmath199 0.22 & 12.02@xmath199 0.02 & 5.74@xmath199 0.02 & 20.2 & 91.3 & 15.5 & 6.60 & 293.1 & 2 + 238 & 220.46141 & -0.60237 & 2.2 & 19.45@xmath199 0.18 & 21.09@xmath199 0.23 & 11.30@xmath199 0.15 & 5.29@xmath199 0.02 & & 14.3 & 16.4 & 20.8 & 3.15 & 200.7 & 2 + 239 & 220.47417 & -1.59828 & 0.9 & & & 0.36@xmath199 0.06 & 0.39@xmath199 0.02 & 0.24@xmath199 0.08 & 24.1 & 2.4 & 9.5 & & 0.1 & 1 + 240 & 220.48645 & -1.93346 & 2.8 & & & 0.61@xmath199 0.10 & 0.62@xmath199 0.02 & 0.28@xmath199 0.02 & 24.5 & 18.4 & 10.9 & & 0.8 & 1 + 241 & 220.50360 & -1.93499 & 0.7 & & & 0.52@xmath199 0.05 & 0.47@xmath199 0.02 & 0.23@xmath199 0.01 & 20.3 & 0.9 & 11.1 & & 0.0 & 1 + 242 & 220.57458 & -1.94131 & 0.7 & & & 0.75@xmath199 0.11 & 0.54@xmath199 0.01 & 0.40@xmath199 0.05 & 24.8 & 1.6 & 10.5 & & 0.1 & 1 + 243 & 220.57959 & -2.01522 & 0.7 & & & 0.47@xmath199 0.12 & 0.48@xmath199 0.02 & 0.25@xmath199 0.02 & 24.2 & 1.2 & 10.2 & & 0.0 & 1 + 244 & 220.58614 & -1.99846 & 0.7 & & & 0.59@xmath199 0.07 & 0.38@xmath199 0.03 & 0.24@xmath199 0.02 & 31.6 & 0.7 & 11.7 & & 0.0 & 0 + 245 & 220.61360 & -1.93542 & 0.7 & & & 0.75@xmath199 0.10 & 0.88@xmath199 0.03 & 0.53@xmath199 0.01 & 15.4 & 3.5 & 9.3 & & 0.1 & 1 + 246 & 220.62610 & -1.86985 & 0.7 & & & 0.56@xmath199 0.06 & 0.38@xmath199 0.01 & 0.21@xmath199 0.02 & 20.5 & 0.6 & 12.1 & & 0.0 & 0 + 247 & 220.62987 & -1.98768 & 0.7 & & & 0.22@xmath199 0.05 & 0.32@xmath199 0.00 & 0.28@xmath199 0.01 & 8.9 & 3.2 & 7.9 & & 0.0 & 1 + 248 & 220.63036 & -2.02230 & 0.7 & & & 0.75@xmath199 0.06 & 0.73@xmath199 0.02 & 0.40@xmath199 0.03 & 24.5 & 1.9 & 10.3 & & 0.1 & 1 + 249 & 220.64902 & -1.68893 & 0.8 & & & 0.54@xmath199 0.05 & 0.41@xmath199 0.02 & 0.16@xmath199 0.08 & 19.2 & 0.5 & 13.6 & & 0.1 & 0 + 250 & 220.65720 & -1.96683 & 0.7 & 0.70@xmath199 0.03 & 0.72@xmath199 0.07 & 1.99@xmath199 0.07 & 1.69@xmath199 0.02 & 0.86@xmath199 0.06 & 18.5 & 4.5 & 10.2 & 0.23 & 1.1 & 2 + 251 & 220.66861 & -1.96927 & 0.7 & & & 1.23@xmath199 0.12 & 1.39@xmath199 0.02 & 0.39@xmath199 0.04 & 22.3 & 1.1 & 12.8 & & 0.1 & 1 + 252 & 220.67680 & -1.97724 & 0.7 & & & 1.30@xmath199 0.08 & 1.35@xmath199 0.02 & 0.95@xmath199 0.05 & 25.3 & 5.9 & 9.3 & & 0.1 & 1 + 253 & 220.68027 & -1.61378 & 0.8 & & & 0.48@xmath199 0.05 & 0.32@xmath199 0.02 & 0.28@xmath199 0.02 & 18.9 & 1.4 & 10.4 & & 0.1 & 1 + 254 & 220.68097 & -1.86537 & 0.7 & 1.13@xmath199 0.05 & 2.93@xmath199 0.10 & 7.24@xmath199 0.20 & 4.90@xmath199 0.01 & 4.20@xmath199 0.01 & 15.5 & 17.6 & 10.1 & 0.12 & 2.2 & 2 + 255 & 220.68314 & -1.79279 & 0.7 & & & 1.45@xmath199 0.10 & 1.93@xmath199 0.01 & 1.21@xmath199 0.03 & 22.1 & 9.7 & 8.8 & & 0.2 & 1 + 256 & 220.68343 & -1.82118 & 0.7 & & & 2.45@xmath199 0.12 & 2.43@xmath199 0.02 & 2.11@xmath199 0.03 & 25.6 & 14.2 & 8.9 & & 0.3 & 1 + 257 & 220.68738 & -1.64756 & 0.8 & & 0.83@xmath199 0.07 & 2.70@xmath199 0.09 & 2.03@xmath199 0.02 & 1.10@xmath199 0.01 & 20.3 & 8.1 & 10.0 & & 0.5 & 1 + 258 & 220.68959 & -1.84020 & 0.7 & & & 1.40@xmath199 0.10 & 3.24@xmath199 0.01 & 2.47@xmath199 0.01 & 21.0 & 45.9 & 7.1 & & 0.3 & 1 + 259 & 220.69145 & -1.76785 & 0.7 & & & 0.66@xmath199 0.06 & 0.62@xmath199 0.01 & 0.41@xmath199 0.01 & 18.9 & 2.2 & 9.8 & & 0.1 & 1 + 260 & 220.71892 & -2.09471 & 0.5 & & & 0.48@xmath199 0.13 & 0.54@xmath199 0.03 & 0.24@xmath199 0.03 & 26.7 & 0.5 & 10.5 & & 0.0 & 0 + 261 & 220.71959 & -1.74963 & 0.8 & & 0.60@xmath199 0.06 & 2.69@xmath199 0.13 & 2.15@xmath199 0.02 & 1.22@xmath199 0.03 & 18.3 & 11.7 & 9.3 & & 0.5 & 1 + 262 & 220.72655 & -2.02186 & 2.7 & & & 0.20@xmath199 0.06 & 0.16@xmath199 0.01 & 0.09@xmath199 0.01 & 12.7 & 4.3 & 11.4 & & 0.2 & 1 + 263 & 220.74297 & -1.71511 & 0.8 & & 0.61@xmath199 0.08 & 1.86@xmath199 0.13 & 1.53@xmath199 0.01 & 0.83@xmath199 0.02 & 12.7 & 6.1 & 10.0 & 0.03 & 0.6 & 2 + 264 & 220.77162 & -1.72315 & 0.8 & 3.18@xmath199 0.09 & 22.07@xmath199 0.26 & 29.11@xmath199 0.38 & 17.91@xmath199 0.02 & 8.56@xmath199 0.03 & 18.8 & 27.3 & 13.2 & 0.49 & 12.2 & 2 + 265 & 220.77525 & -1.66612 & 0.8 & & & 1.58@xmath199 0.14 & 1.18@xmath199 0.02 & 0.51@xmath199 0.03 & 23.6 & 1.7 & 13.0 & & 0.2 & 1 + 266 & 220.77763 & -2.07407 & 2.7 & & & 0.58@xmath199 0.05 & 0.45@xmath199 0.01 & 0.32@xmath199 0.01 & 16.3 & 21.9 & 10.1 & & 0.7 & 1 + 267 & 220.78081 & -2.06005 & 2.7 & & & 1.17@xmath199 0.20 & 0.96@xmath199 0.02 & 0.76@xmath199 0.02 & 35.2 & 55.9 & 9.7 & & 1.5 & 1 + 268 & 220.78998 & -1.71551 & 0.8 & 350.99@xmath199 0.70 & 116.76@xmath199 0.58 & 76.06@xmath199 0.61 & 26.79@xmath199 0.03 & 9.98@xmath199 0.03 & 30.9 & 11.7 & 21.2 & 123.21 & 493.6 & 2 + 269 & 220.80859 & -1.74176 & 0.8 & & 3.84@xmath199 0.12 & 4.16@xmath199 0.12 & 2.10@xmath199 0.01 & 1.10@xmath199 0.03 & 20.5 & 2.6 & 14.5 & & 0.9 & 1 + 270 & 220.81714 & -1.72115 & 0.8 & 2.10@xmath199 0.08 & 5.23@xmath199 0.11 & 5.19@xmath199 0.20 & 2.48@xmath199 0.02 & 1.20@xmath199 0.03 & 20.4 & 2.5 & 15.4 & 0.23 & 3.9 & 2 + 271 & 220.84007 & -2.00722 & 0.5 & & & 1.18@xmath199 0.11 & 0.81@xmath199 0.01 & 0.54@xmath199 0.03 & 28.7 & 0.8 & 11.2 & & 0.0 & 1 + 272 & 220.85739 & -2.12180 & 0.5 & & & 0.63@xmath199 0.15 & 0.48@xmath199 0.01 & 0.29@xmath199 0.02 & 30.0 & 0.4 & 11.1 & & 0.0 & 0 + 273 & 220.86861 & -1.98951 & 0.5 & & & 2.03@xmath199 0.14 & 1.91@xmath199 0.01 & 1.09@xmath199 0.03 & 25.6 & 2.2 & 10.3 & & 0.1 & 1 + 274 & 220.90346 & -1.97350 & 0.5 & & & 0.81@xmath199 0.07 & 0.70@xmath199 0.01 & 0.58@xmath199 0.02 & 19.5 & 1.3 & 9.4 & & 0.0 & 1 + 275 & 220.91592 & -1.84368 & 0.4 & & & 0.27@xmath199 0.04 & 0.21@xmath199 0.03 & 0.07@xmath199 0.03 & 24.5 & 0.1 & 13.8 & & 0.0 & 0 + 276 & 220.95256 & -1.94321 & 0.4 & & & 1.49@xmath199 0.15 & 0.91@xmath199 0.02 & 0.57@xmath199 0.02 & 28.1 & 0.5 & 12.1 & & 0.0 & 0 + 277 & 221.01361 & -0.30341 & 1.8 & & & 0.29@xmath199 0.02 & 0.13@xmath199 0.02 & 0.10@xmath199 0.03 & 16.1 & 1.2 & 13.1 & & 0.1 & 0 + 278 & 221.01779 & -1.88110 & 0.4 & & 0.34@xmath199 0.04 & 1.13@xmath199 0.07 & 0.78@xmath199 0.02 & 0.37@xmath199 0.03 & 23.9 & 0.6 & 10.4 & & 0.0 & 1 + 279 & 221.02028 & -1.85618 & 0.4 & & & 0.37@xmath199 0.06 & 0.28@xmath199 0.02 & 0.18@xmath199 0.01 & 20.2 & 0.2 & 10.7 & & 0.0 & 0 + 280 & 221.04073 & -2.01474 & 2.7 & & & 0.34@xmath199 0.07 & 0.27@xmath199 0.01 & 0.17@xmath199 0.02 & 23.6 & 10.2 & 10.7 & & 0.4 & 1 + 281 & 221.04996 & -1.84921 & 0.4 & & & 0.33@xmath199 0.11 & 0.28@xmath199 0.02 & 0.21@xmath199 0.04 & 21.5 & 0.3 & 9.8 & & 0.0 & 0 + 282 & 221.06265 & -1.82691 & 0.4 & & & 0.46@xmath199 0.08 & 0.31@xmath199 0.01 & 0.16@xmath199 0.03 & 17.4 & 0.1 & 12.4 & & 0.0 & 0 + 283 & 221.06631 & -1.89868 & 0.4 & & & 0.73@xmath199 0.07 & 0.43@xmath199 0.02 & 0.23@xmath199 0.04 & 27.6 & 0.2 & 13.1 & & 0.0 & 0 + 284 & 221.07498 & -2.09293 & 2.7 & & & 0.34@xmath199 0.04 & 0.26@xmath199 0.01 & 0.16@xmath199 0.01 & 17.0 & 8.6 & 11.1 & & 0.4 & 1 + 285 & 221.10945 & -1.80266 & 0.4 & & & 0.61@xmath199 0.08 & 0.53@xmath199 0.02 & 0.41@xmath199 0.02 & 26.2 & 0.6 & 9.6 & & 0.0 & 1 + 286 & 221.12860 & -1.78952 & 0.4 & & & 0.69@xmath199 0.11 & 0.69@xmath199 0.01 & 0.41@xmath199 0.02 & 25.5 & 0.6 & 9.9 & & 0.0 & 1 + 287 & 221.13835 & -1.77993 & 0.4 & & & 0.73@xmath199 0.44 & 0.78@xmath199 0.01 & 0.46@xmath199 0.03 & 30.3 & 0.8 & 9.5 & & 0.0 & 1 + 288 & 221.14677 & -1.70049 & 0.4 & & & 0.94@xmath199 0.21 & 0.80@xmath199 0.03 & 0.48@xmath199 0.04 & 25.4 & 0.6 & 10.6 & & 0.0 & 1 + 289 & 221.15402 & -1.76118 & 0.4 & & & 0.62@xmath199 0.05 & 0.47@xmath199 0.02 & 0.28@xmath199 0.02 & 23.9 & 0.3 & 10.7 & & 0.0 & 0 + 290 & 221.18008 & -1.60550 & 0.4 & & & 0.38@xmath199 0.05 & 0.33@xmath199 0.01 & 0.20@xmath199 0.04 & 25.9 & 0.3 & 10.4 & & 0.0 & 0 + 291 & 221.20360 & -1.70795 & 0.4 & & & 1.05@xmath199 0.12 & 1.29@xmath199 0.02 & 0.82@xmath199 0.03 & 22.2 & 1.8 & 9.1 & & 0.0 & 1 + 292 & 221.21077 & -1.69173 & 0.4 & & & 1.91@xmath199 0.13 & 1.90@xmath199 0.03 & 1.45@xmath199 0.04 & 28.6 & 2.9 & 9.2 & & 0.1 & 1 + 293 & 221.21512 & -1.74699 & 0.4 & & & 0.48@xmath199 0.07 & 0.31@xmath199 0.02 & 0.20@xmath199 0.06 & 23.6 & 0.2 & 11.7 & & 0.0 & 0 + 294 & 221.22061 & -1.70672 & 0.4 & & & 1.61@xmath199 0.08 & 1.38@xmath199 0.03 & 1.00@xmath199 0.05 & 30.5 & 1.6 & 9.9 & & 0.0 & 1 + 295 & 221.23824 & -1.68452 & 0.4 & & & 0.67@xmath199 0.05 & 0.65@xmath199 0.01 & 0.31@xmath199 0.02 & 18.7 & 0.4 & 11.1 & & 0.0 & 0 + 296 & 221.24693 & -1.66071 & 0.4 & & & 0.83@xmath199 0.12 & 0.50@xmath199 0.02 & 0.31@xmath199 0.02 & 30.5 & 0.4 & 11.1 & & 0.0 & 0 + 297 & 221.25288 & -1.88401 & 3.0 & & & 0.56@xmath199 0.05 & 0.41@xmath199 0.01 & 0.28@xmath199 0.01 & 18.9 & 19.4 & 10.7 & & 0.8 & 1 + 298 & 221.27051 & -1.86055 & 3.0 & & & 0.64@xmath199 0.08 & 0.40@xmath199 0.02 & 0.22@xmath199 0.02 & 22.3 & 9.8 & 12.7 & & 0.8 & 1 + 299 & 221.28821 & -1.54326 & 0.5 & & & 0.91@xmath199 0.05 & 0.60@xmath199 0.01 & 0.19@xmath199 0.03 & 26.4 & 0.2 & 16.3 & & 0.0 & 0 + 300 & 221.31360 & -1.49345 & 0.5 & & & 0.64@xmath199 0.04 & 0.64@xmath199 0.01 & 0.38@xmath199 0.02 & 15.8 & 1.5 & 9.1 & & 0.0 & 1 + 301 & 221.40959 & -1.45116 & 0.5 & & & 0.39@xmath199 0.04 & 0.53@xmath199 0.01 & 0.25@xmath199 0.01 & 16.7 & 0.8 & 9.6 & & 0.0 & 1 + 302 & 221.41096 & -1.46285 & 0.5 & & & 0.51@xmath199 0.05 & 0.70@xmath199 0.01 & 0.64@xmath199 0.02 & 25.8 & 3.6 & 7.9 & & 0.0 & 1 + 303 & 221.41943 & -1.88986 & 0.4 & & & 0.44@xmath199 0.07 & 0.37@xmath199 0.01 & 0.24@xmath199 0.01 & 16.1 & 0.3 & 10.3 & & 0.0 & 0 + 304 & 221.44730 & -1.45209 & 0.5 & & & 0.83@xmath199 0.04 & 0.62@xmath199 0.02 & 0.39@xmath199 0.09 & 19.2 & 0.6 & 11.1 & & 0.0 & 1 + 305 & 221.45702 & -1.43383 & 0.5 & & & 0.67@xmath199 0.06 & 0.46@xmath199 0.02 & 0.42@xmath199 0.02 & 19.7 & 0.9 & 10.1 & & 0.0 & 1 + 306 & 221.47028 & -1.44233 & 0.5 & & & 0.96@xmath199 0.12 & 0.72@xmath199 0.02 & 0.44@xmath199 0.03 & 28.3 & 0.9 & 10.9 & & 0.0 & 1 + 307 & 221.61063 & -1.96285 & 3.4 & & & 0.43@xmath199 0.07 & 0.42@xmath199 0.01 & 0.26@xmath199 0.01 & 13.1 & 30.8 & 9.9 & & 0.9 & 1 + 308 & 221.62140 & -1.97294 & 3.4 & & & 0.67@xmath199 0.21 & 0.80@xmath199 0.01 & 0.73@xmath199 0.01 & 23.9 & 146.9 & 8.3 & & 1.9 & 1 + 309 & 221.64833 & -0.64641 & 1.8 & & & 0.89@xmath199 0.26 & 0.69@xmath199 0.01 & 0.44@xmath199 0.04 & 32.8 & 11.5 & 10.7 & & 0.5 & 1 + 310 & 221.65941 & -1.97432 & 3.4 & & & 1.34@xmath199 0.06 & 1.04@xmath199 0.02 & 0.59@xmath199 0.03 & 24.3 & 47.7 & 11.2 & & 2.4 & 1 + 311 & 221.67384 & -0.65854 & 1.8 & & & 0.43@xmath199 0.04 & 0.29@xmath199 0.01 & 0.17@xmath199 0.01 & 18.1 & 3.3 & 11.8 & & 0.2 & 1 + 312 & 221.67769 & -1.89724 & 3.4 & 0.68@xmath199 0.04 & 0.83@xmath199 0.04 & 1.26@xmath199 0.11 & 0.89@xmath199 0.02 & 0.70@xmath199 0.01 & 20.3 & 44.0 & 11.4 & 1.06 & 39.2 & 2 + 313 & 221.68977 & -1.91388 & 3.4 & & & 0.91@xmath199 0.07 & 0.68@xmath199 0.01 & 0.50@xmath199 0.01 & 21.0 & 46.6 & 10.4 & & 1.7 & 1 + 314 & 221.76068 & -1.87185 & 3.4 & & & 0.93@xmath199 0.06 & 0.66@xmath199 0.02 & 0.24@xmath199 0.02 & 24.1 & 11.0 & 14.5 & & 1.4 & 1 + 315 & 221.77640 & -1.89046 & 3.4 & & & 1.28@xmath199 0.08 & 0.91@xmath199 0.01 & 0.61@xmath199 0.03 & 26.4 & 49.1 & 10.9 & & 2.2 & 1 + 316 & 221.78694 & -1.46119 & 1.1 & & & 0.73@xmath199 0.07 & 0.52@xmath199 0.01 & 0.27@xmath199 0.02 & 24.1 & 2.0 & 12.1 & & 0.1 & 1 + 317 & 221.79367 & -1.93285 & 3.4 & & & 4.06@xmath199 0.08 & 3.67@xmath199 0.05 & 1.79@xmath199 0.01 & 21.8 & 154.7 & 11.1 & & 7.6 & 1 + 318 & 221.79500 & -1.85506 & 3.4 & & & 0.68@xmath199 0.05 & 0.45@xmath199 0.02 & 0.27@xmath199 0.03 & 19.4 & 18.3 & 11.7 & & 1.1 & 1 + 319 & 221.84097 & -1.99511 & 3.4 & 2.81@xmath199 0.13 & 6.09@xmath199 0.16 & 3.05@xmath199 0.19 & 1.30@xmath199 0.01 & & 17.3 & 7.9 & 22.4 & 0.08 & 63.5 & 2 + 320 & 221.85423 & -1.97388 & 3.4 & 2.35@xmath199 0.10 & & 3.08@xmath199 0.83 & 1.41@xmath199 0.02 & & 24.6 & 11.2 & 19.9 & 0.07 & 45.5 & 2 + 321 & 221.85603 & -2.00270 & 3.4 & 1.45@xmath199 0.10 & 7.16@xmath199 0.24 & 6.35@xmath199 0.27 & 3.16@xmath199 0.03 & & 26.4 & 43.3 & 16.2 & 0.28 & 68.0 & 2 + 322 & 221.87639 & -1.50934 & 1.1 & & & 0.61@xmath199 0.08 & 0.40@xmath199 0.02 & 0.17@xmath199 0.04 & 21.8 & 1.0 & 13.7 & & 0.1 & 0 + 323 & 221.87692 & -1.54629 & 1.1 & & & 0.83@xmath199 0.08 & 0.55@xmath199 0.02 & 0.40@xmath199 0.02 & 25.8 & 3.3 & 11.1 & & 0.2 & 1 + 324 & 221.88266 & -2.00579 & 3.4 & 7.42@xmath199 0.24 & 5.69@xmath199 0.15 & 4.13@xmath199 0.16 & 1.74@xmath199 0.02 & & 25.7 & 16.7 & 18.7 & 0.30 & 121.2 & 2 + 325 & 221.88997 & -1.94734 & 3.4 & 0.23@xmath199 0.03 & 2.20@xmath199 0.13 & 6.09@xmath199 0.13 & 4.80@xmath199 0.03 & 2.86@xmath199 0.04 & 24.8 & 327.3 & 10.1 & 0.06 & 28.9 & 2 + 326 & 221.89005 & -1.97390 & 3.4 & & 2.90@xmath199 0.10 & 4.01@xmath199 0.14 & 1.77@xmath199 0.02 & 0.62@xmath199 0.02 & 25.2 & 26.2 & 15.3 & 0.05 & 27.4 & 2 + 327 & 221.90764 & -1.94720 & 3.4 & 0.56@xmath199 0.03 & 1.53@xmath199 0.13 & 4.87@xmath199 0.24 & 4.04@xmath199 0.03 & & 30.0 & 328.4 & 9.7 & 0.19 & 31.8 & 2 + 328 & 221.92062 & -1.94709 & 3.4 & & & 3.91@xmath199 0.23 & 3.37@xmath199 0.03 & 2.24@xmath199 0.04 & 30.7 & 237.7 & 10.1 & & 7.7 & 1 + 329 & 221.94850 & -1.98371 & 3.4 & 14.56@xmath199 0.38 & 9.23@xmath199 0.23 & 12.94@xmath199 0.25 & 5.66@xmath199 0.01 & 3.14@xmath199 0.02 & 25.7 & 135.1 & 14.0 & & 137.3 & 2 + 330 & 221.96013 & -1.99328 & 3.4 & 51.84@xmath199 0.26 & 49.44@xmath199 0.40 & 30.59@xmath199 0.43 & 16.56@xmath199 0.02 & 8.54@xmath199 0.02 & 22.4 & 211.4 & 17.0 & 15.03 & 1367.9 & 2 + 331 & 221.98361 & -1.66118 & 1.2 & & & 0.47@xmath199 0.04 & 0.40@xmath199 0.02 & 0.23@xmath199 0.02 & 23.4 & 2.4 & 10.7 & & 0.1 & 1 + 332 & 222.02690 & -2.03539 & 3.4 & 0.92@xmath199 0.03 & 1.61@xmath199 0.08 & 3.22@xmath199 0.08 & 2.68@xmath199 0.01 & 1.57@xmath199 0.01 & 20.8 & 116.2 & 11.1 & 0.08 & 29.1 & 2 + 333 & 222.03059 & -2.01666 & 3.4 & & & 1.45@xmath199 0.07 & 1.24@xmath199 0.01 & 0.77@xmath199 0.01 & 20.8 & 90.0 & 10.1 & & 2.8 & 1 + 334 & 222.05750 & -2.03285 & 3.4 & & & 1.47@xmath199 0.07 & 1.02@xmath199 0.01 & 0.98@xmath199 0.02 & 21.6 & 89.0 & 9.9 & & 2.8 & 1 + 335 & 222.06302 & -1.35355 & 1.4 & & & 0.51@xmath199 0.05 & 0.38@xmath199 0.02 & 0.20@xmath199 0.01 & 19.4 & 2.5 & 11.8 & & 0.2 & 1 + 336 & 222.06860 & -1.72828 & 1.2 & & & 0.43@xmath199 0.03 & 0.28@xmath199 0.01 & 0.15@xmath199 0.01 & 19.1 & 1.0 & 12.7 & & 0.1 & 0 + 337 & 222.07347 & -1.61539 & 1.2 & & & 0.32@xmath199 0.04 & 0.27@xmath199 0.01 & 0.10@xmath199 0.07 & 16.9 & 0.7 & 13.0 & & 0.1 & 0 + 338 & 222.10800 & -1.66619 & 1.2 & & & 0.34@xmath199 0.05 & 0.23@xmath199 0.01 & 0.07@xmath199 0.01 & 11.0 & 0.3 & 16.3 & & 0.1 & 0 + 339 & 222.11191 & -1.98560 & 3.2 & & 1.39@xmath199 0.07 & 1.42@xmath199 0.12 & 0.99@xmath199 0.01 & 0.50@xmath199 0.02 & 21.9 & 21.1 & 13.5 & & 4.9 & 1 + 340 & 222.20792 & -1.72119 & 1.4 & & & 0.43@xmath199 0.09 & 0.30@xmath199 0.01 & 0.17@xmath199 0.02 & 20.3 & 2.2 & 11.6 & & 0.1 & 1 + 341 & 222.21552 & -1.74081 & 1.4 & & & 1.00@xmath199 0.10 & 0.84@xmath199 0.02 & 0.47@xmath199 0.06 & 32.3 & 6.9 & 10.9 & & 0.3 & 1 + 342 & 222.22256 & -1.70764 & 1.4 & & & 0.40@xmath199 0.04 & 0.25@xmath199 0.01 & 0.14@xmath199 0.01 & 10.4 & 1.3 & 12.6 & & 0.1 & 1 + 343 & 222.22420 & -0.52362 & 1.1 & & & 0.44@xmath199 0.09 & 0.47@xmath199 0.01 & 0.31@xmath199 0.01 & 22.1 & 4.3 & 9.3 & & 0.1 & 1 + 344 & 222.22839 & -0.48183 & 1.1 & & & 0.75@xmath199 0.50 & 0.87@xmath199 0.01 & 0.54@xmath199 0.02 & 27.5 & 8.0 & 9.3 & & 0.2 & 1 + 345 & 222.22861 & -1.97452 & 3.2 & & & 0.36@xmath199 0.04 & 0.22@xmath199 0.03 & 0.14@xmath199 0.01 & 17.8 & 7.8 & 12.0 & & 0.5 & 1 + 346 & 222.23027 & -1.71933 & 1.4 & & & 0.82@xmath199 0.09 & 0.84@xmath199 0.02 & 0.50@xmath199 0.03 & 20.9 & 10.3 & 9.8 & & 0.3 & 1 + 347 & 222.25806 & -1.65205 & 1.4 & 0.21@xmath199 0.01 & 0.41@xmath199 0.06 & 0.47@xmath199 0.07 & 0.24@xmath199 0.02 & 0.09@xmath199 0.03 & 17.3 & 0.6 & 15.1 & 0.01 & 0.9 & 2 + 348 & 222.28526 & -1.97143 & 3.2 & & & 0.76@xmath199 0.05 & 0.51@xmath199 0.01 & 0.23@xmath199 0.03 & 16.1 & 11.5 & 13.1 & & 1.0 & 1 + 349 & 222.29474 & -1.94483 & 3.2 & & & 0.92@xmath199 0.11 & 0.54@xmath199 0.02 & 0.51@xmath199 0.06 & 25.5 & 31.0 & 10.8 & & 1.4 & 1 + 350 & 222.30548 & -1.96571 & 3.2 & & & 1.27@xmath199 0.13 & 0.94@xmath199 0.01 & 0.63@xmath199 0.07 & 26.6 & 47.5 & 10.8 & & 2.0 & 1 + 351 & 222.33113 & -1.26244 & 1.4 & & & 0.39@xmath199 0.04 & 0.28@xmath199 0.01 & 0.17@xmath199 0.01 & 18.7 & 2.1 & 11.6 & & 0.1 & 1 + 352 & 222.40762 & -0.43437 & 1.0 & & & 0.80@xmath199 0.05 & 0.73@xmath199 0.01 & 0.45@xmath199 0.02 & 17.6 & 4.5 & 10.2 & & 0.2 & 1 + 353 & 222.43056 & -0.37747 & 1.0 & & & 0.28@xmath199 0.04 & 0.21@xmath199 0.03 & 0.16@xmath199 0.01 & 19.0 & 1.3 & 10.4 & & 0.0 & 1 + 354 & 222.44499 & -0.19538 & 4.7 & 0.95@xmath199 0.04 & 1.21@xmath199 0.06 & 1.83@xmath199 0.04 & 1.31@xmath199 0.01 & 0.71@xmath199 0.01 & 17.1 & 91.5 & 12.0 & 0.08 & 45.7 & 2 + 355 & 222.46452 & -0.21392 & 4.7 & & & 0.33@xmath199 0.05 & 0.25@xmath199 0.02 & 0.18@xmath199 0.03 & 17.0 & 31.3 & 10.4 & & 1.2 & 1 + 356 & 222.48000 & -0.36760 & 1.0 & & & 0.51@xmath199 0.09 & 0.36@xmath199 0.01 & 0.24@xmath199 0.02 & 24.1 & 1.8 & 11.0 & & 0.1 & 1 + 357 & 222.50493 & -1.67362 & 1.2 & & & 1.98@xmath199 0.37 & 0.92@xmath199 0.01 & 0.54@xmath199 0.02 & 31.9 & 2.3 & 14.5 & & 0.3 & 0 + 358 & 222.53471 & -1.59515 & 1.2 & & & 1.20@xmath199 0.10 & 0.71@xmath199 0.02 & 0.71@xmath199 0.01 & 24.5 & 5.9 & 10.5 & & 0.2 & 1 + 359 & 222.56842 & -1.58987 & 3.6 & & 0.68@xmath199 0.07 & 2.85@xmath199 0.47 & 2.17@xmath199 0.01 & 1.26@xmath199 0.03 & 28.5 & 205.3 & 9.5 & 0.01 & 13.5 & 2 + 360 & 222.58058 & -1.63926 & 1.2 & & & 0.33@xmath199 0.12 & 0.43@xmath199 0.01 & 0.41@xmath199 0.01 & 11.8 & 11.5 & 7.9 & & 0.1 & 1 + 361 & 222.58540 & -1.62588 & 1.2 & 0.17@xmath199 0.02 & & 2.97@xmath199 0.12 & 2.30@xmath199 0.01 & 1.52@xmath199 0.02 & 22.1 & 16.3 & 10.6 & 0.03 & 1.8 & 2 + 362 & 222.58748 & -1.66868 & 1.2 & & & 2.91@xmath199 0.96 & 1.44@xmath199 0.01 & 0.78@xmath199 0.02 & 30.5 & 3.7 & 14.3 & & 0.4 & 1 + 363 & 222.59723 & -1.54188 & 3.6 & & & 5.02@xmath199 0.85 & 1.51@xmath199 0.05 & 0.83@xmath199 0.07 & 60.0 & 13.8 & 21.5 & & 5.5 & 0 + 364 & 222.59981 & -1.60397 & 3.6 & 8.69@xmath199 0.21 & 16.68@xmath199 0.13 & 18.08@xmath199 0.18 & 8.82@xmath199 0.02 & 3.63@xmath199 0.03 & 21.2 & 161.2 & 15.3 & 0.44 & 253.5 & 2 + 365 & 222.62677 & -1.61108 & 3.6 & 1.25@xmath199 0.04 & 8.10@xmath199 0.15 & 9.58@xmath199 0.18 & 4.60@xmath199 0.01 & & 27.4 & 99.2 & 14.7 & 0.13 & 79.8 & 2 + 366 & 222.63492 & -1.59053 & 3.6 & 10.82@xmath199 0.17 & 12.99@xmath199 0.31 & 5.05@xmath199 0.13 & 2.83@xmath199 0.01 & 1.50@xmath199 0.01 & 20.2 & 30.6 & 19.3 & 1.00 & 256.8 & 2 + 367 & 222.67039 & -0.36185 & 1.0 & & & 0.75@xmath199 0.05 & 0.69@xmath199 0.01 & 0.53@xmath199 0.01 & 15.1 & 6.2 & 9.4 & & 0.1 & 1 + 368 & 222.68245 & -1.57599 & 3.6 & 0.95@xmath199 0.08 & 3.19@xmath199 0.09 & 5.73@xmath199 0.15 & 3.23@xmath199 0.02 & 1.72@xmath199 0.03 & 33.8 & 122.0 & 12.3 & 0.03 & 43.0 & 2 + 369 & 222.69707 & -1.57972 & 3.6 & 0.94@xmath199 0.07 & 2.87@xmath199 0.13 & 3.84@xmath199 0.10 & 2.26@xmath199 0.01 & 0.92@xmath199 0.02 & 20.9 & 54.1 & 13.7 & 0.07 & 38.7 & 2 + 370 & 222.69827 & -1.38536 & 1.2 & & & 0.79@xmath199 0.06 & 0.48@xmath199 0.01 & 0.25@xmath199 0.18 & 22.1 & 1.5 & 13.2 & & 0.1 & 0 + 371 & 222.70326 & -1.39869 & 1.2 & & & 0.62@xmath199 0.04 & 0.52@xmath199 0.01 & 0.17@xmath199 0.13 & 15.1 & 1.1 & 13.8 & & 0.1 & 0 + 372 & 222.70799 & -1.69619 & 1.2 & & & 0.55@xmath199 0.18 & 0.45@xmath199 0.02 & 0.24@xmath199 0.04 & 30.1 & 2.5 & 11.1 & & 0.1 & 1 + 373 & 222.71811 & -1.58156 & 1.2 & & 3.05@xmath199 0.15 & 4.23@xmath199 0.17 & 2.09@xmath199 0.02 & 1.05@xmath199 0.02 & 23.4 & 6.1 & 13.7 & 0.08 & 3.9 & 2 + 374 & 222.73155 & -1.69959 & 1.2 & & & 0.87@xmath199 17.38 & 0.62@xmath199 0.02 & 0.30@xmath199 0.04 & 31.6 & 2.2 & 12.5 & & 0.2 & 0 + 375 & 222.74771 & -1.63462 & 1.2 & & & 1.26@xmath199 0.10 & 0.76@xmath199 0.01 & 0.42@xmath199 0.02 & 28.8 & 2.7 & 12.9 & & 0.2 & 1 + 376 & 222.75381 & -1.58264 & 1.2 & & & 0.83@xmath199 0.10 & 1.77@xmath199 0.01 & 0.89@xmath199 0.01 & 12.7 & 27.5 & 8.1 & & 0.4 & 1 + 377 & 222.76486 & -0.33217 & 1.0 & & 0.54@xmath199 0.07 & 2.11@xmath199 0.07 & 1.87@xmath199 0.01 & 1.26@xmath199 0.01 & 19.2 & 18.8 & 9.1 & 0.07 & 1.2 & 2 + 378 & 222.77548 & -1.57931 & 1.2 & & & 4.44@xmath199 0.29 & 2.23@xmath199 0.01 & 1.57@xmath199 31.30 & 23.9 & 8.8 & 12.8 & & 0.7 & 1 + 379 & 222.78471 & -1.57365 & 1.2 & & 0.85@xmath199 0.06 & 1.31@xmath199 0.09 & 1.18@xmath199 0.01 & & 23.3 & 5.6 & 11.4 & 0.08 & 1.6 & 2 + 380 & 222.78636 & -1.64347 & 1.2 & & & 1.26@xmath199 0.09 & 0.93@xmath199 0.04 & 0.66@xmath199 0.03 & 22.9 & 7.0 & 10.5 & & 0.3 & 1 + 381 & 222.78694 & -0.32338 & 1.0 & & & 1.12@xmath199 0.06 & 0.96@xmath199 0.01 & 0.78@xmath199 0.02 & 21.6 & 8.5 & 9.5 & & 0.2 & 1 + 382 & 222.78970 & -1.65013 & 1.2 & & & 0.75@xmath199 0.07 & 0.67@xmath199 0.03 & 0.61@xmath199 0.02 & 18.6 & 10.0 & 9.1 & & 0.2 & 1 + 383 & 222.81094 & -1.65026 & 1.2 & & & 1.97@xmath199 0.10 & 1.45@xmath199 0.01 & 0.92@xmath199 0.02 & 26.9 & 10.1 & 10.7 & & 0.4 & 1 + 384 & 222.81493 & -2.06935 & 3.2 & & & 0.43@xmath199 0.06 & 0.29@xmath199 0.01 & 0.17@xmath199 0.02 & 20.1 & 9.9 & 11.9 & & 0.6 & 1 + 385 & 222.81668 & -1.60961 & 1.2 & 0.68@xmath199 0.03 & 4.65@xmath199 0.13 & 5.41@xmath199 0.12 & 2.81@xmath199 0.02 & 1.20@xmath199 0.04 & 26.9 & 6.2 & 14.6 & & 4.0 & 2 + 386 & 222.81987 & -1.58147 & 1.2 & & 1.68@xmath199 0.07 & 1.18@xmath199 0.10 & 0.63@xmath199 0.01 & & 16.0 & 0.8 & 17.5 & 0.06 & 2.0 & 2 + 387 & 222.82413 & -1.55637 & 1.2 & & 0.22@xmath199 0.06 & 0.14@xmath199 0.04 & 0.07@xmath199 0.01 & & 6.0 & 0.1 & 19.3 & & 0.1 & 0 + 388 & 222.82747 & -0.40408 & 1.0 & & & 0.32@xmath199 0.08 & 0.48@xmath199 0.01 & 0.41@xmath199 0.01 & 17.6 & 12.7 & 7.4 & & 0.1 & 1 + 389 & 222.83113 & -0.38453 & 1.0 & & & 1.14@xmath199 0.11 & 0.89@xmath199 0.03 & 0.51@xmath199 0.07 & 23.7 & 5.4 & 10.1 & & 0.2 & 1 + 390 & 222.84862 & -0.34452 & 1.0 & & & 0.42@xmath199 0.06 & 0.30@xmath199 0.01 & 0.18@xmath199 0.03 & 18.5 & 1.2 & 11.5 & & 0.1 & 1 + 391 & 222.85194 & -1.40858 & 1.2 & & & 0.32@xmath199 0.04 & 0.27@xmath199 0.01 & 0.16@xmath199 0.01 & 16.7 & 2.0 & 10.6 & & 0.1 & 1 + 392 & 222.85612 & -1.59384 & 3.5 & & 0.89@xmath199 0.09 & 2.86@xmath199 0.07 & 1.73@xmath199 0.02 & 1.10@xmath199 0.01 & 18.2 & 113.5 & 10.5 & 0.03 & 15.4 & 2 + 393 & 222.86479 & -1.98605 & 3.2 & & & 0.17@xmath199 0.07 & 0.15@xmath199 0.01 & 0.08@xmath199 0.01 & 14.6 & 6.8 & 10.7 & & 0.3 & 1 + 394 & 222.87193 & -1.51286 & 0.7 & & & 0.49@xmath199 0.06 & 0.37@xmath199 0.02 & 0.17@xmath199 0.02 & 19.3 & 0.5 & 12.4 & & 0.0 & 0 + 395 & 222.87569 & -1.59096 & 3.5 & 0.24@xmath199 0.04 & 1.46@xmath199 0.12 & 1.27@xmath199 0.11 & 0.64@xmath199 0.01 & 0.27@xmath199 0.01 & 8.9 & 8.6 & 16.4 & 0.01 & 12.3 & 2 + 396 & 222.88792 & -1.58922 & 3.5 & & & 1.82@xmath199 0.11 & 1.66@xmath199 0.03 & 0.97@xmath199 0.02 & 25.1 & 126.8 & 10.1 & & 4.0 & 1 + 397 & 222.89861 & -2.09140 & 3.2 & & & 0.41@xmath199 0.09 & 0.31@xmath199 0.02 & 0.19@xmath199 0.01 & 20.9 & 13.5 & 11.1 & & 0.6 & 1 + 398 & 222.90027 & -1.37600 & 0.8 & & & 0.98@xmath199 19.52 & 0.58@xmath199 0.02 & 0.37@xmath199 0.02 & 32.8 & 1.2 & 12.1 & & 0.1 & 0 + 399 & 222.91457 & -1.63621 & 3.5 & & & 0.43@xmath199 0.12 & 1.00@xmath199 0.06 & 0.10@xmath199 0.01 & 17.7 & 5.1 & 14.8 & & 1.3 & 1 + 400 & 222.91504 & -0.85355 & 1.0 & & & 0.35@xmath199 0.03 & 0.24@xmath199 0.02 & 0.16@xmath199 0.02 & 19.9 & 1.0 & 11.1 & & 0.0 & 0 + 401 & 222.91597 & -1.56927 & 3.5 & & & 2.20@xmath199 0.19 & 1.53@xmath199 0.03 & 0.95@xmath199 0.03 & 27.1 & 76.2 & 11.4 & & 4.0 & 1 + 402 & 222.91953 & -0.41049 & 0.7 & & & 0.65@xmath199 0.07 & 0.41@xmath199 0.01 & 0.30@xmath199 0.01 & 23.6 & 1.0 & 11.2 & & 0.1 & 1 + 403 & 222.92062 & -1.51726 & 0.7 & & & 0.95@xmath199 0.08 & 0.77@xmath199 0.02 & 0.34@xmath199 0.02 & 24.4 & 1.2 & 12.2 & & 0.1 & 1 + 404 & 222.92625 & -1.59276 & 3.5 & & & 2.44@xmath199 0.09 & 1.50@xmath199 0.02 & 1.02@xmath199 0.03 & 19.4 & 70.4 & 11.7 & & 4.2 & 1 + 405 & 222.93190 & -1.66507 & 3.5 & & & 1.12@xmath199 0.06 & 0.80@xmath199 0.02 & 0.40@xmath199 0.03 & 19.5 & 27.1 & 12.4 & & 2.0 & 1 + 406 & 222.93993 & -1.62058 & 3.5 & & & 1.77@xmath199 0.09 & 1.14@xmath199 0.01 & 0.53@xmath199 0.03 & 24.6 & 26.4 & 13.6 & & 2.9 & 1 + 407 & 222.94356 & -1.05401 & 1.0 & & & 0.63@xmath199 0.04 & 0.44@xmath199 0.02 & 0.24@xmath199 0.03 & 23.0 & 1.5 & 11.7 & & 0.1 & 1 + 408 & 222.94872 & -1.03252 & 1.0 & & & 0.27@xmath199 0.04 & 0.14@xmath199 0.01 & 0.06@xmath199 0.02 & 11.6 & 0.2 & 15.2 & & 0.0 & 0 + 409 & 222.95105 & -1.60145 & 3.5 & 0.27@xmath199 0.03 & 1.47@xmath199 0.11 & 2.47@xmath199 0.16 & 1.65@xmath199 0.02 & 0.69@xmath199 0.03 & 15.1 & 52.2 & 12.3 & & 14.3 & 2 + 410 & 222.95894 & -1.81255 & 3.2 & & & 0.37@xmath199 0.06 & 0.23@xmath199 0.01 & 0.10@xmath199 0.03 & 22.9 & 6.7 & 12.3 & & 0.5 & 1 + 411 & 222.96379 & -1.62430 & 3.5 & & & 1.23@xmath199 0.11 & 0.67@xmath199 0.01 & 0.30@xmath199 0.03 & 18.3 & 12.7 & 15.0 & & 1.8 & 1 + 412 & 222.96652 & -1.58795 & 3.5 & & & 1.18@xmath199 0.08 & 0.58@xmath199 0.01 & 0.26@xmath199 0.03 & 25.8 & 9.5 & 15.9 & & 1.6 & 1 + 413 & 222.97255 & -1.55855 & 0.7 & 0.59@xmath199 0.04 & 0.94@xmath199 0.09 & 0.80@xmath199 0.13 & 0.71@xmath199 0.04 & & 14.5 & 0.4 & 14.8 & 0.05 & 0.7 & 2 + 414 & 222.97340 & -0.61340 & 5.7 & 1.30@xmath199 0.05 & 1.20@xmath199 0.06 & 1.52@xmath199 0.03 & 0.89@xmath199 0.02 & 0.45@xmath199 0.01 & 18.7 & 64.6 & 13.3 & 0.05 & 69.0 & 2 + 415 & 222.97495 & -0.44952 & 0.9 & & & 0.37@xmath199 0.06 & 0.31@xmath199 0.01 & 0.22@xmath199 0.02 & 23.0 & 1.7 & 10.2 & & 0.1 & 1 + 416 & 222.97861 & -1.66099 & 1.2 & & & 0.51@xmath199 0.11 & 0.43@xmath199 0.02 & 0.22@xmath199 0.05 & 17.5 & 2.2 & 11.2 & & 0.1 & 1 + 417 & 222.98337 & -1.45358 & 1.2 & & & 1.06@xmath199 0.17 & 1.30@xmath199 0.01 & 0.73@xmath199 0.02 & 25.7 & 13.6 & 9.4 & & 0.3 & 1 + 418 & 222.99214 & -1.43047 & 1.2 & 0.89@xmath199 0.04 & 1.16@xmath199 0.09 & 2.16@xmath199 0.11 & 1.48@xmath199 0.02 & 0.79@xmath199 0.03 & 16.8 & 7.9 & 11.6 & 0.07 & 3.0 & 2 + 419 & 222.99365 & -1.51953 & 0.7 & & & 1.95@xmath199 0.22 & 1.73@xmath199 0.02 & 0.45@xmath199 0.04 & 34.3 & 0.6 & 17.7 & & 0.2 & 0 + 420 & 222.99698 & -1.51120 & 0.7 & & & 1.26@xmath199 0.16 & 1.21@xmath199 0.01 & 0.79@xmath199 0.03 & 26.7 & 5.3 & 9.5 & & 0.1 & 1 + 421 & 223.00484 & -1.94195 & 1.2 & & & 0.42@xmath199 0.10 & 0.35@xmath199 0.01 & 0.14@xmath199 0.03 & 25.1 & 1.1 & 12.4 & & 0.1 & 0 + 422 & 223.00908 & -1.36770 & 0.8 & & & 0.47@xmath199 0.07 & 0.25@xmath199 0.02 & 0.07@xmath199 0.02 & 19.2 & 0.0 & 33.6 & & 0.0 & 0 + 423 & 223.01208 & -1.50200 & 1.2 & & 0.45@xmath199 0.05 & 1.01@xmath199 0.14 & 0.58@xmath199 0.02 & 0.30@xmath199 0.02 & 22.4 & 3.1 & 11.6 & & 0.4 & 1 + 424 & 223.01291 & -1.47151 & 1.2 & & & 0.71@xmath199 0.10 & 0.44@xmath199 0.01 & 0.28@xmath199 0.01 & 15.5 & 2.2 & 12.1 & & 0.1 & 1 + 425 & 223.01425 & -1.57421 & 1.2 & & & 2.14@xmath199 0.14 & 1.19@xmath199 0.01 & 0.65@xmath199 0.03 & 25.3 & 3.9 & 13.5 & & 0.4 & 1 + 426 & 223.02278 & -1.77786 & 3.4 & & & 0.52@xmath199 0.08 & 0.50@xmath199 0.01 & 0.27@xmath199 0.02 & 14.2 & 28.5 & 10.4 & & 1.1 & 1 + 427 & 223.02487 & -1.76720 & 3.4 & & 0.19@xmath199 0.04 & 0.41@xmath199 0.08 & 0.37@xmath199 0.01 & 0.13@xmath199 0.01 & 12.1 & 12.5 & 11.3 & 0.03 & 4.0 & 2 + 428 & 223.03162 & -1.30387 & 0.8 & & & 0.48@xmath199 0.04 & 0.49@xmath199 0.04 & 0.27@xmath199 0.01 & 19.7 & 1.6 & 10.1 & & 0.1 & 1 + 429 & 223.03275 & -1.40421 & 1.2 & & & 1.00@xmath199 0.07 & 0.61@xmath199 0.03 & 0.28@xmath199 0.03 & 24.8 & 1.9 & 13.6 & & 0.2 & 0 + 430 & 223.03484 & -1.56766 & 1.2 & & & 0.93@xmath199 0.14 & 0.68@xmath199 0.02 & 0.45@xmath199 0.02 & 22.4 & 4.7 & 10.9 & & 0.2 & 1 + 431 & 223.03651 & -1.77189 & 3.4 & 0.26@xmath199 0.03 & 0.32@xmath199 0.07 & 0.48@xmath199 0.07 & 0.36@xmath199 0.01 & 0.18@xmath199 0.01 & 10.6 & 12.5 & 12.0 & & 4.1 & 2 + 432 & 223.03957 & -0.44354 & 1.1 & & & 0.39@xmath199 0.04 & 0.38@xmath199 0.01 & 0.15@xmath199 0.01 & 15.1 & 1.3 & 11.7 & & 0.1 & 1 + 433 & 223.04083 & -1.64808 & 1.2 & & & 1.06@xmath199 0.11 & 0.62@xmath199 0.01 & 0.39@xmath199 0.02 & 32.9 & 2.6 & 12.3 & & 0.2 & 1 + 434 & 223.04205 & -1.48186 & 1.2 & 0.15@xmath199 0.01 & & 1.20@xmath199 0.08 & 0.90@xmath199 0.02 & 0.50@xmath199 0.02 & 23.3 & 4.9 & 11.5 & 0.07 & 1.2 & 2 + 435 & 223.05009 & -0.38869 & 1.1 & & & 0.35@xmath199 0.04 & 0.24@xmath199 0.01 & 0.14@xmath199 0.03 & 19.2 & 1.0 & 12.0 & & 0.1 & 0 + 436 & 223.05023 & -1.91044 & 1.2 & & & 0.30@xmath199 0.06 & 0.20@xmath199 0.01 & 0.10@xmath199 0.01 & 16.9 & 0.6 & 13.0 & & 0.1 & 0 + 437 & 223.05197 & -1.86582 & 1.2 & & & 1.02@xmath199 0.09 & 0.64@xmath199 0.02 & 0.44@xmath199 0.03 & 29.4 & 3.5 & 11.6 & & 0.2 & 1 + 438 & 223.05559 & -1.82892 & 3.4 & & & 0.28@xmath199 0.07 & 0.19@xmath199 0.01 & 0.10@xmath199 0.02 & 10.3 & 6.9 & 12.0 & & 0.5 & 1 + 439 & 223.05699 & -1.27096 & 0.8 & & & 0.47@xmath199 0.04 & 0.39@xmath199 0.01 & 0.20@xmath199 0.01 & 21.6 & 0.8 & 11.4 & & 0.0 & 0 + 440 & 223.05861 & -1.94953 & 1.2 & & & 0.46@xmath199 0.06 & 0.42@xmath199 0.01 & 0.21@xmath199 0.01 & 17.1 & 2.2 & 11.0 & & 0.1 & 1 + 441 & 223.06108 & -1.65532 & 1.2 & & & 0.86@xmath199 0.40 & 0.56@xmath199 0.02 & 0.29@xmath199 0.04 & 34.2 & 2.0 & 12.8 & & 0.2 & 0 + 442 & 223.06293 & -1.39705 & 0.8 & & & 0.60@xmath199 0.13 & 0.35@xmath199 0.02 & 0.17@xmath199 0.05 & 26.0 & 0.4 & 13.9 & & 0.0 & 0 + 443 & 223.06514 & -1.29952 & 0.8 & & & 0.37@xmath199 0.05 & 0.33@xmath199 0.01 & 0.24@xmath199 0.01 & 20.2 & 1.5 & 9.7 & & 0.0 & 1 + 444 & 223.06929 & -1.94058 & 1.2 & & & 0.37@xmath199 0.10 & 0.32@xmath199 0.01 & 0.27@xmath199 0.02 & 15.4 & 4.1 & 9.3 & & 0.1 & 1 + 445 & 223.07265 & -1.55845 & 1.2 & & & 1.65@xmath199 0.16 & 0.85@xmath199 0.01 & 0.31@xmath199 0.05 & 31.6 & 1.1 & 18.0 & & 0.3 & 0 + 446 & 223.07832 & -1.87203 & 1.2 & & & 0.57@xmath199 0.12 & 0.54@xmath199 0.02 & 0.35@xmath199 0.02 & 23.3 & 4.9 & 9.9 & & 0.1 & 1 + 447 & 223.08694 & -1.55981 & 1.2 & & & 1.52@xmath199 0.07 & 0.84@xmath199 0.02 & 0.35@xmath199 0.04 & 23.5 & 1.6 & 15.7 & & 0.3 & 0 + 448 & 223.08853 & -1.94225 & 1.2 & & & 0.82@xmath199 0.08 & 0.90@xmath199 0.01 & 0.44@xmath199 0.03 & 23.5 & 6.7 & 10.1 & & 0.2 & 1 + 449 & 223.09561 & -1.48521 & 1.2 & & & 1.61@xmath199 0.15 & 1.92@xmath199 0.01 & 1.24@xmath199 0.02 & 15.5 & 25.6 & 9.1 & & 0.5 & 1 + 450 & 223.09947 & -0.69093 & 0.9 & & & 0.74@xmath199 0.04 & 0.49@xmath199 0.01 & 0.30@xmath199 0.02 & 27.6 & 1.5 & 11.7 & & 0.1 & 1 + 451 & 223.10693 & -1.57471 & 1.2 & & & 0.56@xmath199 0.08 & 0.31@xmath199 0.01 & 0.20@xmath199 0.02 & 24.3 & 1.3 & 12.7 & & 0.1 & 0 + 452 & 223.12990 & -2.05228 & 1.2 & & & 0.43@xmath199 0.19 & 0.53@xmath199 0.01 & 0.25@xmath199 0.02 & 26.0 & 3.7 & 10.0 & & 0.1 & 1 + 453 & 223.13829 & -1.28272 & 0.8 & & & 0.44@xmath199 0.04 & 0.34@xmath199 0.02 & 0.25@xmath199 0.02 & 25.4 & 1.3 & 10.2 & & 0.0 & 1 + 454 & 223.13840 & -1.43799 & 1.2 & & 0.55@xmath199 0.04 & 0.92@xmath199 0.11 & 0.52@xmath199 0.02 & 0.29@xmath199 0.05 & 24.9 & 2.4 & 12.1 & & 0.4 & 1 + 455 & 223.14328 & -0.19754 & 0.8 & & & 0.38@xmath199 0.04 & 0.29@xmath199 0.01 & 0.19@xmath199 0.01 & 20.0 & 0.8 & 10.9 & & 0.0 & 1 + 456 & 223.14444 & -1.78344 & 3.4 & & 0.46@xmath199 0.05 & 1.27@xmath199 0.07 & 0.96@xmath199 0.01 & 0.50@xmath199 0.02 & 19.2 & 55.6 & 10.4 & & 4.0 & 1 + 457 & 223.16089 & -1.29860 & 0.8 & & & 0.94@xmath199 0.15 & 0.65@xmath199 0.02 & 0.59@xmath199 0.03 & 35.3 & 1.8 & 11.1 & & 0.1 & 1 + 458 & 223.16231 & -1.71619 & 1.2 & & & 0.40@xmath199 0.06 & 0.33@xmath199 0.01 & 0.20@xmath199 0.01 & 16.1 & 1.8 & 11.1 & & 0.1 & 1 + 459 & 223.17906 & -0.66603 & 0.9 & & & 0.67@xmath199 0.04 & 0.42@xmath199 0.02 & 0.20@xmath199 0.03 & 31.3 & 0.4 & 16.0 & & 0.1 & 0 + 460 & 223.17934 & -1.67436 & 3.3 & & & 0.38@xmath199 0.06 & 0.26@xmath199 0.01 & 0.11@xmath199 0.02 & 19.8 & 5.2 & 13.6 & & 0.5 & 1 + 461 & 223.19714 & -0.65160 & 0.9 & & & 0.29@xmath199 0.03 & 0.22@xmath199 0.01 & 0.15@xmath199 0.01 & 16.7 & 1.0 & 10.7 & & 0.0 & 1 + 462 & 223.20377 & -0.70562 & 1.0 & & & 0.68@xmath199 0.03 & 0.26@xmath199 0.01 & 0.25@xmath199 0.02 & 20.2 & 0.7 & 13.0 & & 0.1 & 0 + 463 & 223.24860 & -1.88349 & 3.1 & & & 1.50@xmath199 0.16 & 0.91@xmath199 0.02 & 0.43@xmath199 0.02 & 30.6 & 16.3 & 14.1 & & 1.9 & 1 + 464 & 223.24872 & -1.65500 & 3.3 & & & 0.97@xmath199 0.08 & 0.70@xmath199 0.01 & 0.35@xmath199 0.02 & 20.2 & 21.6 & 12.2 & & 1.5 & 1 + 465 & 223.26091 & -1.84888 & 3.1 & & & 0.57@xmath199 0.07 & 0.64@xmath199 0.01 & 0.41@xmath199 0.04 & 21.2 & 51.1 & 9.3 & & 1.2 & 1 + 466 & 223.27808 & -1.65952 & 3.3 & & & 1.34@xmath199 0.10 & 0.94@xmath199 0.01 & 0.40@xmath199 0.01 & 21.8 & 20.3 & 13.3 & & 2.0 & 1 + 467 & 223.28690 & -1.00619 & 1.0 & & & 0.32@xmath199 0.04 & 0.25@xmath199 0.01 & 0.20@xmath199 0.01 & 19.5 & 1.8 & 9.9 & & 0.1 & 1 + 468 & 223.28694 & -1.80786 & 3.1 & & & 0.75@xmath199 0.08 & 0.54@xmath199 0.02 & 0.28@xmath199 0.03 & 27.1 & 16.5 & 12.0 & & 1.1 & 1 + 469 & 223.29581 & -1.87834 & 3.1 & & & 0.75@xmath199 0.11 & 0.69@xmath199 0.01 & 0.37@xmath199 0.02 & 14.2 & 30.6 & 10.7 & & 1.3 & 1 + 470 & 223.29851 & -1.08897 & 1.0 & & & 0.43@xmath199 0.03 & 0.28@xmath199 0.01 & 0.19@xmath199 0.03 & 17.8 & 1.2 & 11.2 & & 0.1 & 1 + 471 & 223.31477 & -0.88728 & 1.0 & 0.48@xmath199 0.02 & 0.24@xmath199 0.03 & 0.17@xmath199 0.06 & 0.05@xmath199 0.01 & & 8.6 & 0.0 & 21.2 & 0.01 & 0.6 & 2 + 472 & 223.33730 & -1.87369 & 1.2 & & & 1.68@xmath199 0.33 & 1.50@xmath199 0.01 & 0.84@xmath199 0.04 & 31.3 & 8.6 & 11.1 & & 0.4 & 1 + 473 & 223.34309 & -1.72302 & 1.2 & & & 0.77@xmath199 0.14 & 0.49@xmath199 0.01 & 0.26@xmath199 0.01 & 19.4 & 1.8 & 12.7 & & 0.1 & 1 + 474 & 223.34862 & -0.74233 & 0.8 & & & 0.19@xmath199 0.02 & 0.16@xmath199 0.02 & 0.07@xmath199 0.03 & 23.2 & 0.3 & 11.9 & & 0.0 & 0 + 475 & 223.34940 & -1.86552 & 1.2 & & & 0.75@xmath199 15.06 & 1.26@xmath199 0.02 & 0.56@xmath199 0.03 & 27.7 & 11.8 & 9.1 & & 0.3 & 1 + 476 & 223.37361 & -1.59898 & 1.2 & & & 0.72@xmath199 0.06 & 0.65@xmath199 0.01 & 0.31@xmath199 0.02 & 23.9 & 2.9 & 11.6 & & 0.2 & 1 + 477 & 223.37543 & -1.86953 & 1.2 & & & 1.24@xmath199 0.07 & 0.74@xmath199 0.03 & 0.61@xmath199 0.03 & 19.9 & 5.0 & 11.1 & & 0.2 & 1 + 478 & 223.40527 & -1.80164 & 1.2 & & & 1.41@xmath199 0.13 & 1.04@xmath199 0.03 & 0.64@xmath199 0.03 & 28.9 & 6.3 & 11.1 & & 0.3 & 1 + 479 & 223.40964 & -0.94786 & 1.0 & & & 0.69@xmath199 0.04 & 0.60@xmath199 0.02 & 0.56@xmath199 0.01 & 18.7 & 6.5 & 9.1 & & 0.1 & 1 + 480 & 223.41028 & -0.85286 & 1.0 & & & 1.35@xmath199 0.06 & 1.24@xmath199 0.02 & 0.51@xmath199 0.01 & 22.1 & 3.3 & 11.9 & & 0.2 & 1 + 481 & 223.41595 & -1.84052 & 1.2 & & & 2.01@xmath199 0.13 & 2.00@xmath199 0.01 & 1.27@xmath199 0.02 & 27.0 & 19.2 & 9.7 & & 0.5 & 1 + 482 & 223.41833 & -1.83520 & 1.2 & 0.35@xmath199 0.03 & 0.81@xmath199 0.11 & 1.81@xmath199 0.10 & 2.14@xmath199 0.01 & 1.01@xmath199 0.01 & 19.9 & 15.0 & 10.1 & 0.16 & 2.2 & 2 + 483 & 223.42528 & -1.59261 & 1.2 & & & 0.38@xmath199 0.08 & 0.25@xmath199 0.02 & 0.09@xmath199 0.03 & 19.6 & 0.5 & 14.6 & & 0.1 & 0 + 484 & 223.42528 & -0.92787 & 1.0 & & & 2.44@xmath199 0.16 & 2.69@xmath199 0.02 & 2.13@xmath199 0.03 & 23.4 & 31.7 & 8.7 & & 0.5 & 1 + 485 & 223.43175 & -1.82718 & 1.2 & & & 1.99@xmath199 0.15 & 2.34@xmath199 0.01 & 0.80@xmath199 0.02 & 28.0 & 8.3 & 11.5 & & 0.5 & 1 + 486 & 223.43361 & -1.73632 & 1.2 & & & 10.85@xmath199 0.18 & 4.38@xmath199 0.03 & 3.85@xmath199 0.02 & 31.6 & 18.0 & 13.0 & & 1.7 & 1 + 487 & 223.43385 & -1.87511 & 1.2 & 0.29@xmath199 0.04 & 0.79@xmath199 0.08 & 0.48@xmath199 0.12 & 0.17@xmath199 0.01 & & 13.6 & 0.1 & 22.2 & 0.02 & 1.0 & 2 + 488 & 223.43474 & -1.81328 & 1.2 & 3.42@xmath199 0.08 & 4.60@xmath199 0.16 & 5.01@xmath199 0.24 & 2.69@xmath199 0.01 & 1.20@xmath199 0.01 & 13.2 & 6.1 & 14.7 & 0.84 & 12.3 & 2 + 489 & 223.43942 & -0.93338 & 1.0 & 0.86@xmath199 0.04 & 1.39@xmath199 0.09 & 2.36@xmath199 0.11 & 1.69@xmath199 0.01 & & 13.6 & 5.6 & 11.7 & 0.02 & 1.9 & 2 + 490 & 223.44583 & -1.74260 & 1.2 & & & 4.86@xmath199 0.12 & 2.84@xmath199 0.01 & 1.39@xmath199 0.02 & 19.1 & 7.9 & 13.8 & & 0.8 & 1 + 491 & 223.45030 & -0.86363 & 1.0 & & & 0.31@xmath199 0.06 & 0.44@xmath199 0.01 & 0.31@xmath199 0.03 & 25.9 & 5.5 & 8.4 & & 0.1 & 1 + 492 & 223.45251 & -0.93630 & 1.0 & 12.37@xmath199 0.14 & 11.01@xmath199 0.18 & 8.85@xmath199 0.22 & 4.81@xmath199 0.02 & 1.65@xmath199 0.03 & 14.7 & 4.6 & 17.0 & 3.13 & 27.5 & 2 + 493 & 223.45625 & -1.74560 & 1.2 & & & 2.67@xmath199 0.13 & 2.83@xmath199 0.01 & 1.80@xmath199 0.03 & 19.4 & 30.3 & 9.5 & & 0.8 & 1 + 494 & 223.46800 & -0.88045 & 1.0 & & & 0.34@xmath199 0.06 & 0.26@xmath199 0.02 & 0.10@xmath199 0.02 & 19.0 & 0.5 & 13.1 & & 0.0 & 0 + 495 & 223.46904 & -0.40646 & 0.8 & & 0.37@xmath199 0.02 & 0.30@xmath199 0.03 & 0.16@xmath199 0.02 & & 19.3 & 0.1 & 16.4 & 0.01 & 0.2 & 2 + 496 & 223.47890 & -0.91579 & 1.0 & & & 0.76@xmath199 0.06 & 0.81@xmath199 0.01 & 0.53@xmath199 0.02 & 18.5 & 6.4 & 9.4 & & 0.2 & 1 + 497 & 223.48235 & -0.99317 & 1.0 & & 1.24@xmath199 0.06 & 1.30@xmath199 0.15 & 0.85@xmath199 0.01 & 0.46@xmath199 0.02 & 26.0 & 1.7 & 13.6 & 0.01 & 0.9 & 2 + 498 & 223.48395 & -1.75978 & 1.2 & & & 6.58@xmath199 0.28 & 6.59@xmath199 0.01 & 3.55@xmath199 0.03 & 24.7 & 47.2 & 10.3 & & 1.7 & 1 + 499 & 223.49496 & -1.63762 & 1.2 & & & 0.68@xmath199 0.13 & 0.49@xmath199 0.02 & 0.32@xmath199 0.03 & 26.4 & 3.1 & 11.0 & & 0.1 & 1 + 500 & 223.49736 & -0.59913 & 0.8 & & & 0.40@xmath199 0.03 & 0.30@xmath199 0.02 & 0.14@xmath199 0.02 & 16.8 & 0.4 & 12.6 & & 0.0 & 0 + 501 & 223.50287 & -0.92286 & 1.0 & & & 0.64@xmath199 0.06 & 0.59@xmath199 0.01 & 0.46@xmath199 0.02 & 22.8 & 5.1 & 9.4 & & 0.1 & 1 + 502 & 223.51814 & -1.83536 & 1.1 & & & 2.77@xmath199 0.16 & 1.32@xmath199 0.02 & 1.29@xmath199 0.02 & 25.1 & 6.8 & 12.0 & & 0.5 & 1 + 503 & 223.51830 & -1.77699 & 1.1 & 1.90@xmath199 0.08 & 4.10@xmath199 0.20 & 9.36@xmath199 0.19 & 6.72@xmath199 0.01 & 4.88@xmath199 0.02 & 24.1 & 61.8 & 10.1 & 0.09 & 7.8 & 2 + 504 & 223.52119 & -0.56286 & 0.8 & & & 0.28@xmath199 0.04 & 0.25@xmath199 0.01 & 0.16@xmath199 0.02 & 14.6 & 0.9 & 10.2 & & 0.0 & 1 + 505 & 223.52808 & -1.79086 & 1.1 & 0.23@xmath199 0.04 & 1.87@xmath199 0.18 & 11.25@xmath199 0.21 & 10.77@xmath199 0.02 & 7.87@xmath199 0.02 & 28.6 & 208.2 & 8.3 & 0.05 & 5.2 & 2 + 506 & 223.53317 & -1.85961 & 1.1 & 0.26@xmath199 0.04 & 1.31@xmath199 0.16 & 7.91@xmath199 0.38 & 5.51@xmath199 0.02 & 3.60@xmath199 0.01 & 24.9 & 75.1 & 8.9 & 0.07 & 3.5 & 2 + 507 & 223.53740 & -0.93668 & 1.0 & & & 0.58@xmath199 0.06 & 0.54@xmath199 0.01 & 0.42@xmath199 0.01 & 15.1 & 4.8 & 9.4 & & 0.1 & 1 + 508 & 223.54027 & -1.80403 & 1.1 & & & 1.82@xmath199 0.19 & 1.83@xmath199 0.01 & 1.24@xmath199 0.01 & 17.2 & 19.2 & 9.5 & & 0.5 & 1 + 509 & 223.54720 & -0.95313 & 1.0 & & & 2.57@xmath199 0.09 & 2.25@xmath199 0.01 & 1.57@xmath199 0.03 & 24.4 & 15.3 & 9.9 & & 0.5 & 1 + 510 & 223.56526 & -1.87287 & 1.1 & & & 1.59@xmath199 0.14 & 1.35@xmath199 0.01 & 1.35@xmath199 0.02 & 24.5 & 19.9 & 9.1 & & 0.4 & 1 + 511 & 223.57730 & -1.81759 & 1.1 & & & 1.35@xmath199 0.18 & 1.17@xmath199 0.02 & 0.86@xmath199 0.01 & 16.7 & 11.2 & 9.8 & & 0.3 & 1 + 512 & 223.58498 & -1.76281 & 1.1 & 0.66@xmath199 0.06 & 1.36@xmath199 0.25 & 1.26@xmath199 0.22 & 0.98@xmath199 0.01 & 0.44@xmath199 0.01 & 21.2 & 2.5 & 13.6 & 0.08 & 2.3 & 2 + 513 & 223.58656 & -1.85283 & 1.1 & & & 0.98@xmath199 0.35 & 0.69@xmath199 0.01 & 0.28@xmath199 0.02 & 12.0 & 1.7 & 13.7 & & 0.2 & 1 + 514 & 223.59427 & -1.86906 & 1.1 & & & 1.83@xmath199 0.25 & 1.86@xmath199 0.02 & 0.86@xmath199 0.04 & 22.1 & 9.7 & 10.8 & & 0.4 & 1 + 515 & 223.59512 & -1.82017 & 1.1 & & & 4.36@xmath199 0.18 & 2.32@xmath199 0.04 & 1.69@xmath199 0.02 & 31.5 & 10.5 & 12.1 & & 0.7 & 1 + 516 & 223.59544 & -0.92339 & 1.0 & & & 1.81@xmath199 0.08 & 1.89@xmath199 0.02 & 1.11@xmath199 0.03 & 22.9 & 11.9 & 9.8 & & 0.3 & 1 + 517 & 223.60013 & -1.86221 & 1.1 & & 1.95@xmath199 0.12 & 2.96@xmath199 0.30 & 2.08@xmath199 0.02 & & 23.2 & 8.0 & 12.2 & 0.05 & 2.6 & 2 + 518 & 223.60364 & -1.77837 & 1.1 & 1.23@xmath199 0.18 & 2.23@xmath199 0.12 & 1.78@xmath199 0.12 & 0.72@xmath199 0.02 & & 22.1 & 0.8 & 18.1 & 0.05 & 3.3 & 2 + 519 & 223.60861 & -0.88459 & 1.0 & & & 0.67@xmath199 0.10 & 0.68@xmath199 0.03 & 0.57@xmath199 0.02 & 20.0 & 7.7 & 8.9 & & 0.1 & 1 + 520 & 223.61470 & -1.89882 & 1.1 & 4.93@xmath199 0.19 & 4.55@xmath199 0.20 & 1.95@xmath199 0.26 & 0.99@xmath199 0.03 & & 18.1 & 0.6 & 22.8 & 0.34 & 10.6 & 2 + 521 & 223.62488 & -1.72885 & 1.1 & & 1.01@xmath199 0.13 & 0.98@xmath199 0.12 & 0.55@xmath199 0.02 & & 25.7 & 1.1 & 15.0 & 0.01 & 0.9 & 2 + 522 & 223.63028 & -0.70208 & 0.8 & & & 0.29@xmath199 0.07 & 0.35@xmath199 0.01 & 0.16@xmath199 0.02 & 21.1 & 1.0 & 10.2 & & 0.0 & 1 + 523 & 223.63716 & -1.89807 & 1.1 & & & 1.59@xmath199 0.11 & 1.23@xmath199 0.01 & 0.76@xmath199 0.02 & 17.3 & 6.8 & 11.1 & & 0.3 & 1 + 524 & 223.64027 & -1.85328 & 1.1 & & & 1.67@xmath199 0.19 & 1.43@xmath199 0.01 & 1.01@xmath199 0.04 & 19.0 & 12.8 & 10.0 & & 0.4 & 1 + 525 & 223.64360 & -1.11120 & 1.0 & & & 0.72@xmath199 0.12 & 0.36@xmath199 0.02 & 0.23@xmath199 0.02 & 24.6 & 1.0 & 13.0 & & 0.1 & 0 + 526 & 223.64798 & -1.79046 & 1.1 & & 1.06@xmath199 0.11 & 1.73@xmath199 0.15 & 1.27@xmath199 0.01 & & 23.5 & 5.5 & 11.8 & 0.10 & 2.0 & 2 + 527 & 223.65361 & -0.86809 & 0.8 & & & 1.00@xmath199 0.15 & 0.88@xmath199 0.02 & 0.46@xmath199 0.05 & 26.2 & 2.3 & 10.9 & & 0.1 & 1 + 528 & 223.65729 & -1.96564 & 1.1 & 6.79@xmath199 0.14 & 6.48@xmath199 0.16 & 7.83@xmath199 0.34 & 3.76@xmath199 0.02 & 1.41@xmath199 0.03 & 25.9 & 6.1 & 15.2 & 0.41 & 13.7 & 2 + 529 & 223.66194 & -0.84791 & 0.8 & & & 0.43@xmath199 0.10 & 0.51@xmath199 0.02 & 0.38@xmath199 0.02 & 25.9 & 3.7 & 8.7 & & 0.1 & 1 + 530 & 223.66194 & -0.81619 & 0.8 & & & 0.36@xmath199 0.09 & 0.36@xmath199 0.01 & 0.23@xmath199 0.02 & 19.9 & 1.6 & 9.7 & & 0.0 & 1 + 531 & 223.66371 & -1.80012 & 1.1 & 5.87@xmath199 0.29 & 5.40@xmath199 0.59 & 5.70@xmath199 0.47 & 2.60@xmath199 0.03 & 1.22@xmath199 0.02 & 32.5 & 4.8 & 15.7 & 0.20 & 12.2 & 2 + 532 & 223.67334 & -0.74704 & 0.8 & & & 0.78@xmath199 0.06 & 0.69@xmath199 0.02 & 0.23@xmath199 0.02 & 24.6 & 0.6 & 13.6 & & 0.1 & 0 + 533 & 223.67926 & -1.10755 & 1.0 & & & 0.56@xmath199 0.05 & 0.31@xmath199 0.02 & 0.15@xmath199 0.02 & 17.6 & 0.5 & 14.6 & & 0.1 & 0 + 534 & 223.68062 & -1.88802 & 1.1 & & & 3.72@xmath199 0.16 & 3.10@xmath199 0.02 & 2.02@xmath199 0.03 & 23.4 & 23.1 & 10.4 & & 0.8 & 1 + 535 & 223.68073 & -1.75561 & 1.1 & 9.90@xmath199 0.80 & 4.25@xmath199 0.12 & 2.12@xmath199 0.44 & 0.91@xmath199 0.02 & & 23.9 & 0.6 & 22.4 & 0.10 & 13.2 & 2 + 536 & 223.68161 & -0.74865 & 0.8 & & & 0.79@xmath199 0.07 & 0.71@xmath199 0.02 & 0.49@xmath199 0.02 & 26.4 & 4.1 & 9.1 & & 0.1 & 1 + 537 & 223.69362 & -0.83443 & 0.8 & & & 0.40@xmath199 0.05 & 0.42@xmath199 0.01 & 0.29@xmath199 0.02 & 17.3 & 2.2 & 9.4 & & 0.1 & 1 + 538 & 223.70442 & -1.77953 & 1.1 & & & 2.00@xmath199 0.22 & 1.57@xmath199 0.02 & 0.78@xmath199 0.04 & 24.6 & 6.6 & 11.8 & & 0.4 & 1 + 539 & 223.72227 & -1.98775 & 1.1 & & 2.73@xmath199 0.25 & 3.25@xmath199 0.15 & 4.63@xmath199 0.02 & & 16.8 & 12.8 & 11.6 & 0.07 & 3.2 & 2 + 540 & 223.72473 & -0.84729 & 0.8 & 0.46@xmath199 0.02 & 0.67@xmath199 0.04 & 2.78@xmath199 0.09 & 2.36@xmath199 0.02 & 1.66@xmath199 0.02 & 25.7 & 14.6 & 9.1 & 0.13 & 1.2 & 2 + 541 & 223.72644 & -0.12151 & 1.0 & & & 0.96@xmath199 0.05 & 0.69@xmath199 0.01 & 0.44@xmath199 0.02 & 18.7 & 2.8 & 11.1 & & 0.1 & 1 + 542 & 223.73253 & -1.90505 & 1.1 & 2.51@xmath199 0.10 & 5.96@xmath199 0.21 & 4.92@xmath199 0.15 & 3.20@xmath199 0.00 & 1.53@xmath199 0.03 & 8.6 & 5.9 & 14.7 & 0.21 & 7.5 & 2 + 543 & 223.73430 & -0.82835 & 0.8 & & 0.76@xmath199 0.08 & 1.20@xmath199 0.11 & 1.25@xmath199 0.01 & 0.81@xmath199 0.01 & 14.0 & 4.1 & 10.5 & & 0.3 & 1 + 544 & 223.73595 & -1.78453 & 1.1 & & & 2.77@xmath199 0.27 & 2.12@xmath199 0.02 & 1.16@xmath199 0.04 & 30.5 & 10.2 & 11.5 & & 0.6 & 1 + 545 & 223.74008 & -0.78903 & 0.8 & & & 0.74@xmath199 0.07 & 0.74@xmath199 0.02 & 0.43@xmath199 0.02 & 19.9 & 2.4 & 10.1 & & 0.1 & 1 + 546 & 223.74031 & -1.76021 & 1.1 & & & 2.69@xmath199 0.35 & 1.85@xmath199 0.02 & 1.24@xmath199 0.02 & 28.7 & 11.0 & 11.1 & & 0.5 & 1 + 547 & 223.74309 & -1.67066 & 1.3 & & & 0.71@xmath199 0.13 & 0.55@xmath199 0.01 & 0.29@xmath199 0.01 & 15.2 & 3.1 & 11.5 & & 0.2 & 1 + 548 & 223.74458 & -1.90972 & 1.1 & 4.45@xmath199 0.09 & 4.57@xmath199 0.21 & 5.02@xmath199 0.23 & 3.89@xmath199 0.01 & 1.62@xmath199 0.04 & 16.0 & 8.9 & 13.3 & 1.43 & 13.3 & 2 + 549 & 223.74734 & -0.87097 & 0.8 & 1.29@xmath199 0.05 & 3.16@xmath199 0.11 & 3.52@xmath199 0.16 & 2.18@xmath199 0.02 & 1.09@xmath199 0.02 & 14.5 & 2.8 & 13.7 & 0.01 & 1.9 & 2 + 550 & 223.74777 & -1.11150 & 0.9 & & & 1.30@xmath199 0.07 & 1.00@xmath199 0.02 & 0.93@xmath199 0.03 & 23.4 & 7.7 & 9.5 & & 0.2 & 1 + 551 & 223.74796 & -1.14737 & 0.9 & & & 0.27@xmath199 0.07 & 0.19@xmath199 0.01 & 0.12@xmath199 0.01 & 14.3 & 1.1 & 10.1 & & 0.0 & 1 + 552 & 223.75101 & -1.08787 & 0.8 & & & 0.48@xmath199 0.08 & 0.35@xmath199 0.01 & 0.22@xmath199 0.02 & 16.4 & 1.0 & 11.1 & & 0.0 & 1 + 553 & 223.75519 & -1.81576 & 1.1 & & & 1.41@xmath199 0.20 & 1.27@xmath199 0.02 & 0.79@xmath199 0.02 & 18.1 & 9.8 & 10.2 & & 0.3 & 1 + 554 & 223.76877 & -0.06406 & 1.0 & & & 0.89@xmath199 0.03 & 0.43@xmath199 0.01 & 0.25@xmath199 0.02 & 22.4 & 0.8 & 14.1 & & 0.1 & 0 + 555 & 223.77863 & -1.82442 & 1.1 & & 0.90@xmath199 0.09 & 1.87@xmath199 0.23 & 1.39@xmath199 0.02 & & 23.7 & 7.9 & 11.0 & 0.06 & 1.6 & 2 + 556 & 223.77907 & -1.96897 & 1.1 & 0.84@xmath199 0.09 & 1.23@xmath199 0.13 & 1.01@xmath199 0.15 & 0.61@xmath199 0.01 & & 10.1 & 1.1 & 15.2 & 0.05 & 1.9 & 2 + 557 & 223.77994 & -0.05393 & 1.0 & & & 0.76@xmath199 0.04 & 0.52@xmath199 0.01 & 0.30@xmath199 0.02 & 22.0 & 1.6 & 11.8 & & 0.1 & 1 + 558 & 223.77998 & -1.09112 & 0.9 & & & 0.58@xmath199 0.11 & 0.38@xmath199 0.03 & 0.25@xmath199 0.02 & 23.4 & 1.2 & 11.6 & & 0.1 & 0 + 559 & 223.78171 & -0.80509 & 0.8 & & & 0.73@xmath199 0.10 & 0.64@xmath199 0.02 & 0.53@xmath199 0.04 & 20.2 & 3.8 & 9.3 & & 0.1 & 1 + 560 & 223.78377 & -0.81954 & 0.8 & & & 0.56@xmath199 0.10 & 0.55@xmath199 0.02 & 0.35@xmath199 0.02 & 20.5 & 2.3 & 9.8 & & 0.1 & 1 + 561 & 223.78526 & -1.81287 & 1.1 & & & 5.67@xmath199 2.09 & 3.71@xmath199 0.02 & 2.52@xmath199 0.03 & 39.1 & 20.7 & 11.3 & & 1.1 & 1 + 562 & 223.78952 & -1.96029 & 1.1 & 1.79@xmath199 0.06 & 5.31@xmath199 0.22 & 4.66@xmath199 0.15 & 2.28@xmath199 0.02 & 1.25@xmath199 0.02 & 24.1 & 4.0 & 15.5 & 0.12 & 5.8 & 2 + 563 & 223.79861 & -1.64567 & 1.3 & & 0.85@xmath199 0.08 & 0.94@xmath199 0.10 & 0.52@xmath199 0.01 & & 19.9 & 1.4 & 14.5 & 0.02 & 1.2 & 2 + 564 & 223.80055 & -1.82971 & 1.1 & & & 3.87@xmath199 0.17 & 3.37@xmath199 0.01 & 2.20@xmath199 0.02 & 26.0 & 26.2 & 10.2 & & 0.9 & 1 + 565 & 223.80098 & -0.84007 & 0.8 & & & 0.81@xmath199 0.13 & 0.70@xmath199 0.03 & 0.41@xmath199 0.05 & 24.1 & 1.8 & 11.1 & & 0.1 & 1 + 566 & 223.80861 & -1.48786 & 1.3 & & & 0.64@xmath199 0.05 & 0.38@xmath199 0.01 & 0.13@xmath199 0.02 & 15.1 & 0.7 & 16.2 & & 0.1 & 0 + 567 & 223.81619 & -1.64809 & 1.3 & & & 2.25@xmath199 0.27 & 1.11@xmath199 0.03 & 0.57@xmath199 0.02 & 33.6 & 1.8 & 17.7 & & 0.4 & 0 + 568 & 223.83205 & -1.83230 & 1.1 & & 1.33@xmath199 0.09 & 3.52@xmath199 0.25 & 2.28@xmath199 0.02 & 1.99@xmath199 0.01 & 23.3 & 19.5 & 10.3 & & 1.3 & 1 + 569 & 223.83426 & -1.52120 & 1.3 & & & 2.13@xmath199 0.12 & 1.59@xmath199 0.03 & 0.91@xmath199 0.03 & 24.6 & 9.8 & 11.4 & & 0.5 & 1 + 570 & 223.83655 & -1.53128 & 1.3 & & & 1.69@xmath199 0.17 & 1.67@xmath199 0.03 & 0.54@xmath199 0.03 & 21.1 & 5.0 & 12.7 & & 0.4 & 1 + 571 & 223.83768 & -1.86863 & 1.1 & & & 1.43@xmath199 0.13 & 1.37@xmath199 0.01 & 0.85@xmath199 0.03 & 13.9 & 10.8 & 10.1 & & 0.3 & 1 + 572 & 223.84377 & -1.48866 & 1.3 & & 1.06@xmath199 0.08 & 1.18@xmath199 0.13 & 0.60@xmath199 0.01 & 0.29@xmath199 0.01 & 20.0 & 2.4 & 13.6 & 0.02 & 1.4 & 2 + 573 & 223.84630 & -1.87010 & 1.1 & 0.68@xmath199 0.05 & 1.39@xmath199 0.17 & 2.66@xmath199 0.21 & 2.46@xmath199 0.04 & & 17.6 & 14.6 & 10.7 & 0.14 & 3.0 & 2 + 574 & 223.84694 & -1.93481 & 1.1 & & & 2.35@xmath199 0.17 & 1.51@xmath199 0.02 & 0.78@xmath199 0.03 & 24.2 & 5.0 & 12.8 & & 0.4 & 1 + 575 & 223.85141 & -0.89910 & 0.8 & & & 0.49@xmath199 0.05 & 0.34@xmath199 0.01 & 0.30@xmath199 0.01 & 17.7 & 1.5 & 10.1 & & 0.1 & 1 + 576 & 223.85164 & -1.18779 & 0.9 & & & 0.64@xmath199 0.17 & 0.52@xmath199 0.05 & 0.35@xmath199 0.02 & 16.4 & 1.8 & 11.1 & & 0.1 & 1 + 577 & 223.85423 & -1.78700 & 1.1 & 0.85@xmath199 0.05 & 1.63@xmath199 0.09 & 3.01@xmath199 0.28 & 1.86@xmath199 0.02 & 0.74@xmath199 0.03 & 31.7 & 5.7 & 12.6 & 0.03 & 2.8 & 2 + 578 & 223.85878 & -1.12276 & 0.9 & & & 0.84@xmath199 0.18 & 1.02@xmath199 0.01 & 0.63@xmath199 0.01 & 19.6 & 7.3 & 9.2 & & 0.2 & 1 + 579 & 223.86028 & -0.87139 & 0.8 & & & 0.47@xmath199 0.09 & 0.37@xmath199 0.01 & 0.33@xmath199 0.02 & 17.2 & 2.0 & 9.6 & & 0.1 & 1 + 580 & 223.86278 & -0.90451 & 0.8 & & & 0.42@xmath199 0.04 & 0.36@xmath199 0.01 & 0.23@xmath199 0.01 & 14.4 & 1.3 & 10.3 & & 0.0 & 1 + 581 & 223.87694 & -0.87620 & 0.8 & & & 0.48@xmath199 0.07 & 0.49@xmath199 0.01 & 0.32@xmath199 0.02 & 24.5 & 2.3 & 9.6 & & 0.1 & 1 + 582 & 223.87755 & -1.58461 & 1.3 & & & 0.74@xmath199 0.12 & 0.48@xmath199 0.01 & 0.31@xmath199 0.01 & 18.2 & 3.1 & 11.5 & & 0.2 & 1 + 583 & 223.87941 & -1.43669 & 1.3 & & 1.83@xmath199 0.07 & 2.20@xmath199 0.15 & 1.45@xmath199 0.01 & 0.52@xmath199 0.02 & 22.5 & 3.8 & 14.0 & 0.03 & 2.5 & 2 + 584 & 223.88095 & -1.12962 & 0.9 & & & 0.60@xmath199 0.09 & 0.61@xmath199 0.01 & 0.51@xmath199 0.01 & 12.9 & 5.9 & 8.9 & & 0.1 & 1 + 585 & 223.88525 & -2.00384 & 1.1 & & 1.44@xmath199 0.09 & 0.98@xmath199 0.18 & 0.46@xmath199 0.01 & & 19.7 & 0.5 & 18.5 & 0.02 & 1.2 & 2 + 586 & 223.88632 & -0.89515 & 0.8 & & & 0.39@xmath199 0.09 & 0.25@xmath199 0.02 & 0.17@xmath199 0.02 & 20.8 & 1.1 & 10.1 & & 0.0 & 1 + 587 & 223.88988 & -1.87494 & 1.1 & & 0.47@xmath199 0.18 & 1.48@xmath199 0.15 & 2.36@xmath199 0.01 & 1.99@xmath199 0.03 & 23.5 & 49.0 & 8.1 & 0.01 & 1.1 & 2 + 588 & 223.89125 & -1.71452 & 1.1 & & & 0.95@xmath199 0.04 & 0.45@xmath199 0.01 & 0.27@xmath199 0.01 & 8.9 & 1.2 & 14.1 & & 0.1 & 1 + 589 & 223.89288 & -1.84989 & 1.1 & 0.50@xmath199 0.04 & 0.92@xmath199 0.14 & 2.18@xmath199 0.20 & 1.77@xmath199 0.01 & 1.34@xmath199 0.01 & 17.6 & 14.8 & 10.1 & 0.03 & 1.9 & 2 + 590 & 223.89372 & -1.21371 & 0.9 & & & 0.53@xmath199 0.12 & 0.41@xmath199 0.01 & 0.35@xmath199 0.02 & 27.1 & 3.0 & 9.7 & & 0.1 & 1 + 591 & 223.89456 & -1.90002 & 1.1 & & & 0.57@xmath199 0.15 & 1.50@xmath199 0.04 & 0.21@xmath199 0.01 & 7.7 & 1.7 & 12.1 & & 0.2 & 1 + 592 & 223.89989 & -1.85853 & 1.1 & & 0.72@xmath199 0.13 & 1.13@xmath199 0.17 & 0.98@xmath199 0.01 & & 12.8 & 4.4 & 11.4 & & 0.5 & 1 + 593 & 223.90250 & -1.87620 & 1.1 & & & 2.39@xmath199 0.25 & 3.45@xmath199 0.01 & 2.12@xmath199 0.03 & 24.1 & 46.4 & 8.7 & & 0.8 & 1 + 594 & 223.90266 & -1.44495 & 1.3 & & 0.97@xmath199 0.06 & 1.00@xmath199 0.12 & 0.44@xmath199 0.01 & & 21.3 & 0.9 & 16.0 & 0.07 & 1.8 & 2 + 595 & 223.90669 & -1.80657 & 1.1 & & & 9.86@xmath199 0.79 & 5.66@xmath199 0.15 & 3.78@xmath199 0.06 & 52.4 & 24.3 & 12.2 & & 1.7 & 1 + 596 & 223.91028 & -1.46620 & 1.3 & & & 2.09@xmath199 0.09 & 1.23@xmath199 0.01 & 0.71@xmath199 0.02 & 23.6 & 5.4 & 12.7 & & 0.4 & 1 + 597 & 223.91878 & -1.49035 & 1.3 & & & 2.22@xmath199 0.80 & 2.06@xmath199 0.02 & 1.44@xmath199 0.02 & 22.2 & 25.0 & 9.7 & & 0.7 & 1 + 598 & 223.92104 & -1.60540 & 1.3 & & & 0.96@xmath199 0.21 & 0.80@xmath199 0.03 & 0.35@xmath199 0.03 & 21.9 & 3.3 & 12.3 & & 0.2 & 1 + 599 & 223.92487 & -1.15411 & 0.9 & & 0.41@xmath199 0.06 & 1.65@xmath199 0.44 & 2.69@xmath199 0.02 & 2.02@xmath199 0.19 & 21.9 & 31.4 & 8.3 & & 0.6 & 1 + 600 & 223.92694 & -0.78995 & 0.8 & & & 0.88@xmath199 0.11 & 0.61@xmath199 0.02 & 0.35@xmath199 0.03 & 23.9 & 1.3 & 11.8 & & 0.1 & 1 + 601 & 223.93051 & -1.88770 & 1.1 & 1.18@xmath199 0.07 & 2.16@xmath199 0.17 & 2.80@xmath199 0.25 & 1.87@xmath199 0.01 & 0.79@xmath199 0.02 & 14.1 & 4.3 & 13.6 & & 2.6 & 2 + 602 & 223.93158 & -1.65494 & 1.1 & & & 0.75@xmath199 0.09 & 0.46@xmath199 0.01 & 0.23@xmath199 0.04 & 25.0 & 1.3 & 13.3 & & 0.1 & 0 + 603 & 223.93387 & -1.49817 & 1.3 & & 2.38@xmath199 0.09 & 3.99@xmath199 0.12 & 2.59@xmath199 0.01 & 1.30@xmath199 0.02 & 25.3 & 12.5 & 12.1 & & 1.9 & 1 + 604 & 223.93506 & -1.50724 & 1.3 & & & 3.90@xmath199 0.12 & 2.16@xmath199 0.02 & 0.66@xmath199 0.02 & 25.5 & 2.4 & 19.1 & & 0.7 & 0 + 605 & 223.93608 & -1.89748 & 1.1 & 0.35@xmath199 0.07 & 1.28@xmath199 0.12 & 2.37@xmath199 0.17 & 1.82@xmath199 0.01 & & 27.4 & 9.3 & 11.2 & & 1.5 & 2 + 606 & 223.94341 & -1.97331 & 1.1 & & & 2.33@xmath199 0.13 & 1.62@xmath199 0.01 & 0.64@xmath199 0.01 & 22.8 & 3.5 & 14.0 & & 0.4 & 1 + 607 & 223.94472 & -0.96151 & 3.6 & & & 0.67@xmath199 0.08 & 0.56@xmath199 0.02 & 0.37@xmath199 0.03 & 24.7 & 40.4 & 10.4 & & 1.5 & 1 + 608 & 223.94691 & -1.06714 & 0.9 & & & 0.27@xmath199 0.05 & 0.22@xmath199 0.02 & 0.10@xmath199 0.03 & 21.5 & 0.5 & 12.1 & & 0.0 & 0 + 609 & 223.94861 & -1.16364 & 0.9 & & & 0.94@xmath199 0.06 & 0.97@xmath199 0.01 & 0.70@xmath199 0.01 & 21.9 & 7.5 & 9.2 & & 0.2 & 1 + 610 & 223.94992 & -0.90822 & 3.6 & & & 0.33@xmath199 0.06 & 0.25@xmath199 0.01 & 0.13@xmath199 0.05 & 18.3 & 10.3 & 11.8 & & 0.6 & 1 + 611 & 223.95480 & -0.58193 & 0.8 & 0.56@xmath199 0.04 & 0.68@xmath199 0.04 & 0.18@xmath199 0.05 & 0.01@xmath199 0.00 & & 1.2 & 0.0 & 14.8 & 0.04 & 0.7 & 2 + 612 & 223.95763 & -1.11681 & 0.9 & & & 1.30@xmath199 25.94 & 1.00@xmath199 0.02 & 0.63@xmath199 0.02 & 32.2 & 4.1 & 10.8 & & 0.2 & 1 + 613 & 223.96527 & -1.17333 & 0.9 & & & 0.41@xmath199 0.08 & 0.36@xmath199 0.01 & 0.29@xmath199 0.01 & 13.8 & 2.6 & 9.5 & & 0.1 & 1 + 614 & 223.96693 & -1.88535 & 1.1 & & & 2.22@xmath199 0.14 & 1.51@xmath199 0.01 & 0.71@xmath199 0.02 & 24.1 & 4.6 & 12.9 & & 0.4 & 1 + 615 & 223.97079 & -1.83067 & 1.1 & & 1.22@xmath199 0.16 & 7.63@xmath199 0.22 & 7.95@xmath199 0.02 & 6.10@xmath199 0.02 & 30.2 & 193.5 & 7.9 & & 3.0 & 1 + 616 & 223.98596 & -1.13722 & 0.9 & & & 0.94@xmath199 0.10 & 0.85@xmath199 0.02 & 0.39@xmath199 0.02 & 18.4 & 2.4 & 11.4 & & 0.1 & 1 + 617 & 223.98892 & -1.81974 & 1.1 & 3.95@xmath199 0.09 & 9.43@xmath199 0.18 & 12.85@xmath199 0.31 & 10.37@xmath199 0.01 & 6.91@xmath199 0.01 & 19.0 & 50.2 & 11.6 & 0.07 & 13.6 & 2 + 618 & 223.99164 & -1.88733 & 1.1 & & & 2.48@xmath199 0.23 & 1.33@xmath199 0.01 & 0.59@xmath199 0.03 & 25.9 & 2.4 & 15.4 & & 0.4 & 1 + 619 & 224.00452 & -1.73126 & 1.1 & 1.53@xmath199 0.07 & 3.67@xmath199 0.17 & 5.26@xmath199 0.21 & 3.20@xmath199 0.01 & 1.91@xmath199 0.02 & 20.5 & 11.7 & 12.5 & 0.12 & 5.5 & 2 + 620 & 224.01366 & -1.68320 & 1.1 & 2.45@xmath199 0.09 & 6.35@xmath199 0.19 & 7.87@xmath199 0.21 & 5.00@xmath199 0.06 & 2.85@xmath199 0.01 & 15.0 & 15.8 & 12.9 & 0.06 & 8.1 & 2 + 621 & 224.01607 & -1.71020 & 1.1 & 1.77@xmath199 0.07 & 3.00@xmath199 0.17 & 4.40@xmath199 0.29 & 3.31@xmath199 0.01 & 2.51@xmath199 0.03 & 19.9 & 17.1 & 11.4 & 0.72 & 7.5 & 2 + 622 & 224.01776 & -1.64680 & 1.1 & & & 1.12@xmath199 0.07 & 1.18@xmath199 0.00 & 0.84@xmath199 0.02 & 11.8 & 13.4 & 9.2 & & 0.3 & 1 + 623 & 224.01990 & -1.66953 & 1.1 & & & 1.20@xmath199 0.13 & 1.21@xmath199 0.02 & 0.81@xmath199 0.01 & 10.7 & 11.3 & 9.5 & & 0.3 & 1 + 624 & 224.02362 & -0.11928 & 0.9 & & & 0.34@xmath199 0.05 & 0.26@xmath199 0.01 & 0.21@xmath199 0.01 & 16.7 & 1.6 & 10.0 & & 0.1 & 1 + 625 & 224.02505 & -1.16522 & 0.9 & & & 1.25@xmath199 0.13 & 0.91@xmath199 0.02 & 0.44@xmath199 0.03 & 28.0 & 2.1 & 12.4 & & 0.2 & 1 + 626 & 224.02553 & -1.19359 & 0.9 & & & 1.53@xmath199 0.16 & 1.42@xmath199 0.02 & 0.79@xmath199 0.02 & 30.4 & 6.0 & 10.5 & & 0.2 & 1 + 627 & 224.02635 & -1.62812 & 1.1 & & & 1.47@xmath199 0.09 & 1.07@xmath199 0.01 & 0.50@xmath199 0.02 & 22.3 & 3.4 & 12.5 & & 0.3 & 1 + 628 & 224.05360 & -1.19887 & 0.9 & & & 0.66@xmath199 0.12 & 0.72@xmath199 0.02 & 0.28@xmath199 0.03 & 23.0 & 1.9 & 11.2 & & 0.1 & 1 + 629 & 224.05435 & -1.94756 & 1.1 & & 2.44@xmath199 0.20 & 7.47@xmath199 0.37 & 4.20@xmath199 0.03 & 2.31@xmath199 0.03 & 36.8 & 22.5 & 11.1 & & 2.4 & 1 + 630 & 224.05940 & -1.91428 & 1.1 & & & 1.90@xmath199 0.15 & 1.27@xmath199 0.01 & 0.62@xmath199 0.03 & 26.2 & 4.1 & 12.8 & & 0.3 & 1 + 631 & 224.06139 & -1.76787 & 1.1 & & & 2.33@xmath199 0.12 & 1.03@xmath199 0.02 & 0.38@xmath199 0.02 & 33.2 & 0.9 & 20.2 & & 0.3 & 0 + 632 & 224.07361 & -0.17304 & 0.9 & & & 0.35@xmath199 0.05 & 0.30@xmath199 0.01 & 0.17@xmath199 0.01 & 20.6 & 1.2 & 10.8 & & 0.1 & 1 + 633 & 224.07870 & -1.20919 & 0.9 & & & 0.67@xmath199 0.26 & 1.93@xmath199 0.04 & 0.42@xmath199 0.03 & 24.5 & 4.5 & 9.7 & & 0.2 & 1 + 634 & 224.07895 & -1.96155 & 1.1 & & & 2.82@xmath199 0.17 & 2.93@xmath199 0.01 & 2.81@xmath199 0.02 & 23.4 & 57.6 & 8.5 & & 0.8 & 1 + 635 & 224.08058 & -1.98522 & 1.1 & & & 2.22@xmath199 0.28 & 1.82@xmath199 0.01 & 1.56@xmath199 0.02 & 27.3 & 21.1 & 9.5 & & 0.5 & 1 + 636 & 224.08095 & -1.12886 & 0.9 & & & 1.84@xmath199 0.19 & 2.30@xmath199 0.03 & 1.82@xmath199 0.02 & 30.1 & 27.1 & 8.4 & & 0.4 & 1 + 637 & 224.08249 & -1.95217 & 1.1 & & 0.84@xmath199 0.22 & 3.37@xmath199 0.14 & 3.10@xmath199 0.01 & 1.49@xmath199 0.01 & 19.2 & 26.4 & 9.5 & 0.04 & 1.9 & 2 + 638 & 224.08719 & -1.96632 & 1.1 & 0.24@xmath199 0.05 & 0.45@xmath199 0.10 & 1.26@xmath199 0.12 & 1.27@xmath199 0.01 & & 14.6 & 13.3 & 9.4 & & 0.8 & 2 + 639 & 224.09500 & -1.87034 & 1.1 & & & 2.13@xmath199 0.19 & 1.17@xmath199 0.02 & 0.69@xmath199 0.06 & 29.4 & 3.9 & 13.1 & & 0.3 & 1 + 640 & 224.09650 & -1.84404 & 1.1 & & & 1.53@xmath199 0.28 & 1.18@xmath199 0.01 & 0.70@xmath199 0.02 & 29.3 & 6.5 & 11.1 & & 0.3 & 1 + 641 & 224.09785 & -1.11317 & 0.9 & 0.25@xmath199 0.03 & 2.46@xmath199 0.11 & 8.56@xmath199 0.44 & 7.15@xmath199 0.01 & 4.52@xmath199 0.03 & 21.6 & 47.2 & 9.5 & 0.04 & 2.8 & 2 + 642 & 224.10175 & -1.79669 & 1.1 & & 1.15@xmath199 0.08 & 5.50@xmath199 0.20 & 0.97@xmath199 0.03 & & 33.4 & 3.3 & 12.8 & 0.01 & 1.6 & 2 + 643 & 224.10960 & -1.96574 & 1.1 & 4.89@xmath199 0.10 & 8.19@xmath199 0.21 & 9.82@xmath199 0.31 & 6.91@xmath199 0.02 & 2.68@xmath199 0.03 & 16.2 & 15.7 & 13.6 & 0.43 & 15.3 & 2 + 644 & 224.11028 & -1.27501 & 0.9 & & & 2.72@xmath199 0.16 & 1.94@xmath199 0.03 & 1.04@xmath199 0.02 & 29.4 & 5.3 & 11.9 & & 0.3 & 1 + 645 & 224.11250 & -1.59808 & 1.5 & & & 0.54@xmath199 0.06 & 0.39@xmath199 0.02 & 0.27@xmath199 0.01 & 15.1 & 4.1 & 10.9 & & 0.2 & 1 + 646 & 224.11566 & -1.52120 & 1.5 & & & 0.95@xmath199 0.06 & 0.49@xmath199 0.02 & 0.27@xmath199 0.04 & 23.9 & 2.0 & 14.3 & & 0.2 & 0 + 647 & 224.11861 & -1.74453 & 1.1 & & & 1.36@xmath199 0.14 & 1.39@xmath199 0.01 & 1.01@xmath199 0.01 & 21.2 & 15.5 & 9.3 & & 0.3 & 1 + 648 & 224.11893 & -1.91625 & 1.1 & & & 2.28@xmath199 0.26 & 1.45@xmath199 0.02 & 0.74@xmath199 0.02 & 28.4 & 4.7 & 13.0 & & 0.4 & 1 + 649 & 224.11919 & -1.10325 & 0.9 & 0.71@xmath199 0.03 & 4.33@xmath199 0.16 & 7.48@xmath199 0.13 & 7.91@xmath199 0.02 & 4.72@xmath199 0.01 & 20.6 & 33.7 & 10.4 & 0.11 & 3.9 & 2 + 650 & 224.12708 & -1.08521 & 0.9 & & & 2.89@xmath199 0.10 & 3.41@xmath199 0.01 & 2.70@xmath199 0.02 & 20.2 & 36.4 & 8.6 & & 0.6 & 1 + 651 & 224.13164 & -1.10598 & 0.9 & 1.55@xmath199 0.06 & 1.62@xmath199 0.11 & 3.50@xmath199 0.11 & 2.63@xmath199 0.01 & & 19.9 & 10.3 & 10.8 & 0.34 & 3.5 & 2 + 652 & 224.13194 & -1.09787 & 0.9 & & & 2.21@xmath199 0.10 & 2.68@xmath199 0.01 & 1.74@xmath199 0.01 & 17.3 & 21.1 & 9.0 & & 0.4 & 1 + 653 & 224.13362 & -1.75298 & 1.5 & & 1.22@xmath199 0.11 & 3.16@xmath199 0.23 & 2.45@xmath199 0.02 & 1.90@xmath199 0.01 & 19.5 & 34.7 & 10.1 & 0.01 & 3.0 & 2 + 654 & 224.13771 & -1.51432 & 1.5 & & & 0.61@xmath199 0.06 & 0.31@xmath199 0.01 & 0.16@xmath199 0.01 & 17.8 & 1.2 & 14.7 & & 0.2 & 0 + 655 & 224.14061 & -1.96140 & 1.1 & & 0.42@xmath199 0.14 & 3.12@xmath199 0.16 & 2.92@xmath199 0.01 & 1.91@xmath199 0.02 & 26.0 & 59.6 & 8.1 & & 1.1 & 1 + 656 & 224.14182 & -0.85457 & 0.8 & & & 0.81@xmath199 0.27 & 0.78@xmath199 0.02 & 0.69@xmath199 0.01 & 20.1 & 5.4 & 8.9 & & 0.1 & 1 + 657 & 224.14194 & -0.34721 & 0.9 & & & 0.20@xmath199 0.04 & 0.20@xmath199 0.01 & 0.07@xmath199 0.01 & 11.0 & 0.3 & 12.2 & & 0.0 & 0 + 658 & 224.14420 & -1.74118 & 1.5 & 0.56@xmath199 0.04 & 1.56@xmath199 0.12 & 5.13@xmath199 0.14 & 4.70@xmath199 0.02 & 1.97@xmath199 0.01 & 21.6 & 48.7 & 10.1 & 0.01 & 5.0 & 2 + 659 & 224.14648 & -1.21325 & 0.9 & & & 0.40@xmath199 0.04 & 0.38@xmath199 0.01 & 0.20@xmath199 0.02 & 17.7 & 1.6 & 10.4 & & 0.1 & 1 + 660 & 224.14891 & -1.80812 & 1.1 & & & 6.32@xmath199 0.25 & 4.34@xmath199 0.02 & 2.46@xmath199 0.04 & 30.5 & 19.2 & 11.9 & & 1.2 & 1 + 661 & 224.15317 & -1.74151 & 1.5 & & & 4.72@xmath199 0.19 & 5.86@xmath199 0.02 & 3.42@xmath199 0.02 & 26.6 & 99.9 & 9.3 & & 2.3 & 1 + 662 & 224.15817 & -1.74984 & 1.5 & & & 2.83@xmath199 0.13 & 4.73@xmath199 0.01 & 2.66@xmath199 0.01 & 19.8 & 109.3 & 8.5 & & 1.7 & 1 + 663 & 224.16173 & -1.07454 & 0.9 & & & 1.63@xmath199 0.16 & 1.63@xmath199 0.01 & 1.36@xmath199 0.02 & 20.2 & 13.3 & 8.9 & & 0.3 & 1 + 664 & 224.16476 & -1.55668 & 1.5 & & & 1.15@xmath199 0.07 & 0.46@xmath199 0.02 & 0.31@xmath199 0.02 & 22.5 & 1.8 & 15.0 & & 0.3 & 0 + 665 & 224.16719 & -1.76209 & 1.5 & 0.11@xmath199 0.01 & 1.33@xmath199 0.12 & 4.96@xmath199 0.25 & 7.55@xmath199 0.01 & 4.53@xmath199 0.01 & 23.1 & 198.2 & 8.4 & & 4.6 & 2 + 666 & 224.17250 & -0.87692 & 0.8 & & & 1.20@xmath199 0.19 & 0.90@xmath199 0.01 & 0.59@xmath199 0.06 & 31.1 & 2.7 & 10.8 & & 0.1 & 1 + 667 & 224.17378 & -1.97505 & 1.1 & & & 4.61@xmath199 0.42 & 2.63@xmath199 0.02 & 1.06@xmath199 0.04 & 33.1 & 4.5 & 15.6 & & 0.7 & 1 + 668 & 224.17648 & -1.63467 & 1.5 & & 1.56@xmath199 0.06 & 1.03@xmath199 0.06 & 0.48@xmath199 0.02 & & 21.0 & 0.6 & 20.5 & 0.04 & 2.6 & 2 + 669 & 224.17860 & -1.76362 & 1.5 & & & 5.52@xmath199 0.22 & 6.50@xmath199 0.02 & 4.06@xmath199 0.02 & 34.6 & 117.0 & 9.2 & & 2.6 & 1 + 670 & 224.18361 & -0.27953 & 0.9 & & & 0.45@xmath199 0.04 & 0.40@xmath199 0.01 & 0.22@xmath199 0.01 & 21.6 & 1.4 & 10.6 & & 0.1 & 1 + 671 & 224.18558 & -1.05020 & 0.9 & 5.79@xmath199 0.10 & 17.20@xmath199 0.24 & 21.34@xmath199 0.28 & 13.30@xmath199 0.01 & 4.86@xmath199 0.02 & 16.9 & 15.7 & 14.1 & 0.31 & 12.7 & 2 + 672 & 224.18906 & -1.78121 & 1.5 & 0.49@xmath199 0.07 & 1.18@xmath199 0.08 & 3.05@xmath199 0.23 & 2.56@xmath199 0.01 & 2.00@xmath199 0.04 & 21.9 & 41.5 & 9.8 & 0.03 & 3.9 & 2 + 673 & 224.20534 & -1.78594 & 1.5 & & & 2.84@xmath199 0.12 & 2.17@xmath199 0.01 & 1.21@xmath199 0.03 & 24.1 & 17.8 & 11.4 & & 0.9 & 1 + 674 & 224.20605 & -1.50343 & 1.5 & & & 0.45@xmath199 0.07 & 0.54@xmath199 0.01 & 0.30@xmath199 0.01 & 19.0 & 8.1 & 9.5 & & 0.2 & 1 + 675 & 224.21378 & -1.53713 & 1.5 & & & 1.09@xmath199 0.08 & 0.64@xmath199 0.01 & 0.30@xmath199 0.02 & 29.3 & 2.6 & 14.0 & & 0.3 & 0 + 676 & 224.21519 & -1.96772 & 1.1 & 0.70@xmath199 0.05 & 1.37@xmath199 0.13 & 2.39@xmath199 0.19 & 1.92@xmath199 0.01 & 1.27@xmath199 0.03 & 18.4 & 11.8 & 10.9 & 0.12 & 2.9 & 2 + 677 & 224.22595 & -1.93188 & 1.1 & & & 3.77@xmath199 0.20 & 2.68@xmath199 0.02 & 1.61@xmath199 0.03 & 31.7 & 13.9 & 11.4 & & 0.7 & 1 + 678 & 224.22614 & -0.86249 & 0.8 & 1.54@xmath199 0.06 & 1.15@xmath199 0.10 & 2.05@xmath199 0.47 & 1.32@xmath199 0.02 & & 19.8 & 2.6 & 11.9 & 0.16 & 1.8 & 2 + 679 & 224.22694 & -1.05284 & 0.9 & 11.98@xmath199 0.14 & 26.76@xmath199 0.24 & 28.09@xmath199 0.34 & 14.35@xmath199 0.01 & 6.00@xmath199 0.01 & 19.2 & 14.8 & 15.2 & 0.77 & 21.8 & 2 + 680 & 224.22977 & -1.95120 & 1.1 & & & 5.91@xmath199 0.20 & 3.15@xmath199 0.02 & 2.43@xmath199 0.03 & 31.4 & 15.6 & 11.9 & & 1.0 & 1 + 681 & 224.23131 & -1.71965 & 1.5 & & & 3.21@xmath199 0.10 & 1.98@xmath199 0.01 & 1.18@xmath199 0.02 & 25.7 & 13.2 & 12.3 & & 0.9 & 1 + 682 & 224.23840 & -1.09235 & 0.9 & 0.77@xmath199 0.06 & 2.58@xmath199 0.08 & 3.60@xmath199 0.11 & 1.76@xmath199 0.02 & 0.63@xmath199 0.01 & 19.4 & 1.9 & 14.8 & 0.18 & 2.1 & 2 + 683 & 224.23933 & -1.96573 & 1.1 & & 1.35@xmath199 0.09 & 3.25@xmath199 0.11 & 1.78@xmath199 0.01 & & 20.5 & 9.5 & 11.4 & 0.10 & 2.5 & 2 + 684 & 224.24265 & -1.92941 & 1.1 & 0.81@xmath199 0.04 & 4.15@xmath199 0.11 & 8.44@xmath199 0.37 & 5.58@xmath199 0.01 & 2.67@xmath199 0.05 & 28.5 & 24.4 & 11.7 & 0.09 & 5.8 & 2 + 685 & 224.24466 & -1.05043 & 0.9 & 26.62@xmath199 0.19 & 44.46@xmath199 0.40 & 40.70@xmath199 1.02 & 19.12@xmath199 0.02 & 8.66@xmath199 0.01 & 16.4 & 21.9 & 15.2 & 0.42 & 36.0 & 2 + 686 & 224.24527 & -1.68742 & 1.5 & & & 0.61@xmath199 0.05 & 0.43@xmath199 0.01 & 0.27@xmath199 0.01 & 16.6 & 3.8 & 11.3 & & 0.2 & 1 + 687 & 224.24817 & -0.92974 & 0.9 & 11.60@xmath199 0.14 & 23.87@xmath199 0.21 & 24.01@xmath199 0.38 & 13.16@xmath199 0.01 & 5.43@xmath199 0.01 & 16.8 & 13.9 & 15.1 & 0.94 & 20.8 & 2 + 688 & 224.25121 & -1.66769 & 1.5 & & & 0.63@xmath199 0.12 & 0.54@xmath199 0.03 & 0.29@xmath199 0.02 & 21.6 & 4.8 & 11.0 & & 0.2 & 1 + 689 & 224.25194 & -1.81851 & 1.1 & & & 1.45@xmath199 0.09 & 1.52@xmath199 0.01 & 0.91@xmath199 0.02 & 20.2 & 13.1 & 9.8 & & 0.4 & 1 + 690 & 224.25304 & -0.90533 & 0.9 & & & 3.08@xmath199 0.17 & 3.99@xmath199 0.02 & 3.75@xmath199 0.04 & 28.0 & 57.7 & 7.9 & & 0.6 & 1 + 691 & 224.25429 & -0.84701 & 0.8 & 11.27@xmath199 0.12 & 15.20@xmath199 0.23 & 7.27@xmath199 0.09 & 4.52@xmath199 0.00 & & 8.8 & 1.6 & 20.3 & 0.07 & 9.9 & 2 + 692 & 224.25441 & -0.82933 & 0.8 & 1.50@xmath199 0.06 & 6.31@xmath199 0.32 & 11.52@xmath199 0.18 & 7.76@xmath199 0.02 & & 19.2 & 16.0 & 11.7 & 0.06 & 3.8 & 2 + 693 & 224.25697 & -1.89894 & 1.1 & 0.57@xmath199 0.05 & 1.22@xmath199 0.11 & 1.52@xmath199 0.09 & 0.87@xmath199 0.02 & 0.55@xmath199 0.02 & 16.7 & 3.0 & 13.0 & 0.06 & 2.0 & 2 + 694 & 224.25731 & -0.85966 & 0.8 & & & 12.34@xmath199 0.22 & 14.60@xmath199 0.01 & 11.26@xmath199 0.06 & 26.2 & 108.9 & 8.6 & & 1.7 & 1 + 695 & 224.25861 & -1.14742 & 0.9 & & & 1.07@xmath199 0.07 & 0.99@xmath199 0.01 & 0.53@xmath199 0.02 & 17.9 & 3.3 & 10.6 & & 0.1 & 1 + 696 & 224.26027 & -1.84953 & 1.1 & & & 5.15@xmath199 0.15 & 3.81@xmath199 0.02 & 2.51@xmath199 0.05 & 30.4 & 24.2 & 10.9 & & 1.1 & 1 + 697 & 224.26051 & -0.87219 & 0.8 & 0.65@xmath199 0.02 & 1.00@xmath199 0.09 & 3.83@xmath199 0.24 & 4.95@xmath199 0.01 & & 19.8 & 43.3 & 8.3 & 0.06 & 1.4 & 2 + 698 & 224.26173 & -1.08233 & 0.9 & 0.65@xmath199 0.02 & 3.27@xmath199 0.20 & 6.31@xmath199 0.15 & 4.34@xmath199 0.02 & 2.35@xmath199 0.02 & 21.0 & 11.8 & 11.5 & & 2.1 & 2 + 699 & 224.26218 & -0.84695 & 0.8 & 0.84@xmath199 0.04 & 1.29@xmath199 0.11 & 8.89@xmath199 0.14 & 8.81@xmath199 0.01 & & 18.0 & 114.1 & 7.8 & 0.24 & 2.6 & 2 + 700 & 224.26407 & -0.82944 & 0.8 & 2.91@xmath199 0.06 & 7.01@xmath199 0.43 & 8.83@xmath199 0.14 & 7.43@xmath199 0.01 & 5.65@xmath199 0.01 & 13.9 & 18.1 & 11.4 & 0.06 & 4.6 & 2 + 701 & 224.26624 & -0.20833 & 0.9 & & & 0.29@xmath199 0.05 & 0.25@xmath199 0.01 & 0.16@xmath199 0.01 & 19.1 & 1.1 & 10.3 & & 0.0 & 1 + 702 & 224.27173 & -0.80352 & 0.9 & 0.60@xmath199 0.02 & 1.10@xmath199 0.09 & 2.02@xmath199 0.15 & 3.32@xmath199 0.01 & 1.21@xmath199 0.03 & 15.2 & 10.9 & 10.1 & 0.12 & 1.7 & 2 + 703 & 224.27226 & -1.00414 & 0.9 & & & 1.06@xmath199 0.15 & 0.84@xmath199 0.14 & 0.45@xmath199 0.02 & 20.0 & 2.2 & 11.4 & & 0.1 & 1 + 704 & 224.27237 & -1.06358 & 0.9 & & 3.98@xmath199 0.11 & 8.80@xmath199 0.17 & 5.49@xmath199 0.02 & 2.63@xmath199 0.02 & 23.3 & 13.6 & 11.6 & 0.02 & 2.6 & 2 + 705 & 224.27271 & -1.87298 & 1.1 & & & 2.41@xmath199 0.12 & 2.27@xmath199 0.01 & 1.40@xmath199 0.03 & 23.4 & 18.5 & 10.0 & & 0.6 & 1 + 706 & 224.27753 & -0.83360 & 0.8 & 1.43@xmath199 0.04 & 3.25@xmath199 0.08 & 12.29@xmath199 0.20 & 12.73@xmath199 0.01 & & 20.9 & 90.5 & 8.7 & 0.02 & 3.5 & 2 + 707 & 224.27759 & -1.09282 & 0.9 & 0.51@xmath199 0.05 & 2.78@xmath199 0.09 & 5.52@xmath199 0.13 & 3.76@xmath199 0.02 & 2.47@xmath199 0.01 & 17.8 & 12.2 & 11.1 & 0.08 & 2.2 & 2 + 708 & 224.28355 & -1.83463 & 1.1 & & & 1.21@xmath199 0.10 & 1.12@xmath199 0.01 & 0.84@xmath199 0.02 & 17.9 & 12.3 & 9.5 & & 0.3 & 1 + 709 & 224.28413 & -0.82566 & 0.9 & 3.86@xmath199 0.06 & 10.34@xmath199 0.24 & 15.81@xmath199 0.19 & 14.77@xmath199 0.01 & 9.39@xmath199 0.02 & 19.2 & 55.8 & 10.9 & 0.24 & 10.6 & 2 + 710 & 224.28613 & -1.87148 & 1.1 & & & 3.56@xmath199 0.27 & 3.91@xmath199 0.01 & 1.78@xmath199 0.04 & 26.7 & 21.7 & 10.5 & & 0.9 & 1 + 711 & 224.28912 & -1.95968 & 1.1 & 2.16@xmath199 0.06 & 3.70@xmath199 0.13 & 3.98@xmath199 0.22 & 2.58@xmath199 0.01 & 1.16@xmath199 0.02 & 17.8 & 5.7 & 13.9 & 0.16 & 5.8 & 2 + 712 & 224.28986 & -0.78891 & 0.9 & 0.50@xmath199 0.01 & 0.53@xmath199 0.08 & 1.36@xmath199 0.15 & 1.09@xmath199 0.01 & & 10.6 & 5.6 & 10.1 & 0.01 & 0.8 & 2 + 713 & 224.29196 & -1.10102 & 0.9 & & 3.40@xmath199 0.12 & 6.72@xmath199 0.17 & 5.14@xmath199 0.01 & 2.45@xmath199 0.01 & 24.9 & 13.2 & 11.4 & 0.05 & 2.5 & 2 + 714 & 224.29425 & -1.90722 & 1.1 & & & 2.03@xmath199 0.19 & 1.94@xmath199 0.03 & 1.19@xmath199 0.03 & 23.8 & 14.0 & 10.0 & & 0.4 & 1 + 715 & 224.29572 & -0.87371 & 0.8 & 0.32@xmath199 0.03 & 0.80@xmath199 0.09 & 1.20@xmath199 0.18 & 0.91@xmath199 0.01 & & 11.8 & 1.7 & 12.0 & & 0.4 & 2 + 716 & 224.29828 & -1.07302 & 0.9 & 1.12@xmath199 0.04 & 5.65@xmath199 0.13 & 10.23@xmath199 0.20 & 6.18@xmath199 0.01 & 3.63@xmath199 0.02 & 28.7 & 15.5 & 11.9 & 0.12 & 4.1 & 2 + 717 & 224.30237 & -1.09298 & 0.9 & & 3.06@xmath199 0.09 & 5.01@xmath199 0.25 & 2.35@xmath199 0.01 & & 30.9 & 3.9 & 13.4 & 0.21 & 2.7 & 2 + 718 & 224.30237 & -0.78053 & 0.9 & 0.59@xmath199 0.03 & 1.64@xmath199 0.09 & 2.99@xmath199 0.19 & 3.65@xmath199 0.02 & & 17.4 & 15.4 & 10.2 & 0.02 & 1.8 & 2 + 719 & 224.30763 & -1.03771 & 0.9 & 0.70@xmath199 0.02 & 1.33@xmath199 0.10 & 2.16@xmath199 0.23 & 2.32@xmath199 0.01 & 1.10@xmath199 0.01 & 14.8 & 6.5 & 10.9 & 0.09 & 1.5 & 2 + 720 & 224.30861 & -1.85701 & 1.1 & & & 2.13@xmath199 0.58 & 2.84@xmath199 0.59 & 1.43@xmath199 0.05 & 33.0 & 22.1 & 9.4 & & 0.6 & 1 + 721 & 224.30872 & -0.81765 & 0.9 & 7.02@xmath199 0.11 & 7.37@xmath199 0.18 & 9.68@xmath199 0.28 & 8.55@xmath199 0.01 & & 19.7 & 19.8 & 12.0 & 0.59 & 11.8 & 2 + 722 & 224.31158 & -1.06185 & 0.9 & & 1.07@xmath199 0.12 & 3.29@xmath199 0.20 & 3.18@xmath199 0.01 & 1.25@xmath199 0.01 & 17.1 & 10.8 & 10.1 & & 0.7 & 1 + 723 & 224.31836 & -1.04701 & 0.9 & & & 6.12@xmath199 0.19 & 6.56@xmath199 0.01 & 4.61@xmath199 0.02 & 30.0 & 44.0 & 9.2 & & 0.9 & 1 + 724 & 224.32008 & -1.97435 & 1.1 & 1.19@xmath199 0.06 & 2.18@xmath199 0.12 & 3.06@xmath199 0.16 & 2.04@xmath199 0.01 & & 15.5 & 6.1 & 12.7 & 0.25 & 4.1 & 2 + 725 & 224.32224 & -0.84937 & 0.9 & 4.29@xmath199 0.08 & 6.19@xmath199 0.16 & 8.26@xmath199 0.14 & 5.65@xmath199 0.01 & & 20.2 & 11.6 & 12.8 & 3.13 & 13.5 & 2 + 726 & 224.32617 & -0.66555 & 0.9 & & & 0.64@xmath199 0.10 & 0.50@xmath199 0.02 & 0.21@xmath199 0.03 & 21.6 & 1.0 & 12.7 & & 0.1 & 0 + 727 & 224.32619 & -1.06262 & 0.9 & & 1.33@xmath199 0.08 & 5.75@xmath199 0.16 & 4.11@xmath199 0.01 & 2.08@xmath199 0.02 & 26.2 & 19.6 & 9.7 & 0.05 & 1.6 & 2 + 728 & 224.33194 & -0.84212 & 0.9 & & & 2.19@xmath199 0.16 & 2.82@xmath199 0.01 & 2.28@xmath199 0.04 & 15.5 & 35.8 & 8.3 & & 0.5 & 1 + 729 & 224.33252 & -1.97626 & 1.1 & 2.20@xmath199 0.07 & 4.65@xmath199 0.16 & 6.36@xmath199 0.20 & 4.20@xmath199 0.02 & 2.02@xmath199 0.04 & 19.0 & 15.2 & 12.3 & 0.14 & 6.8 & 2 + 730 & 224.33328 & -0.54790 & 0.7 & & & 3.25@xmath199 0.38 & 2.79@xmath199 0.03 & 2.29@xmath199 0.05 & 35.3 & 12.4 & 9.5 & & 0.3 & 1 + 731 & 224.33340 & -0.74788 & 0.9 & 9.19@xmath199 0.12 & 13.25@xmath199 0.21 & 11.07@xmath199 0.28 & 5.30@xmath199 0.02 & 1.74@xmath199 0.01 & 11.0 & 4.0 & 17.7 & 1.04 & 16.7 & 2 + 732 & 224.33589 & -0.64171 & 0.9 & & & 0.72@xmath199 0.11 & 0.67@xmath199 0.02 & 0.36@xmath199 0.03 & 25.0 & 2.5 & 10.7 & & 0.1 & 1 + 733 & 224.33864 & -0.58645 & 0.9 & & & 0.87@xmath199 0.06 & 0.97@xmath199 0.01 & 0.63@xmath199 0.01 & 17.3 & 6.8 & 9.3 & & 0.2 & 1 + 734 & 224.33873 & -0.56758 & 0.7 & & & 1.85@xmath199 0.08 & 1.78@xmath199 0.01 & 1.21@xmath199 0.02 & 24.5 & 6.7 & 9.7 & & 0.2 & 1 + 735 & 224.34319 & -1.02459 & 0.9 & 2.36@xmath199 0.05 & 7.07@xmath199 0.14 & 12.90@xmath199 0.12 & 9.17@xmath199 0.01 & 5.28@xmath199 0.05 & 18.7 & 26.1 & 11.4 & 0.36 & 6.6 & 2 + 736 & 224.34431 & -1.03567 & 0.9 & & 0.70@xmath199 0.25 & 7.02@xmath199 0.15 & 6.78@xmath199 0.01 & 3.61@xmath199 0.04 & 22.6 & 48.8 & 8.3 & & 1.3 & 1 + 737 & 224.34552 & -0.60799 & 0.9 & & & 0.94@xmath199 0.17 & 1.29@xmath199 0.01 & 0.48@xmath199 0.50 & 18.3 & 3.8 & 10.5 & & 0.2 & 1 + 738 & 224.34567 & -0.60120 & 0.9 & & & 1.21@xmath199 0.12 & 0.60@xmath199 0.01 & 0.50@xmath199 0.67 & 23.7 & 1.8 & 12.2 & & 0.1 & 1 + 739 & 224.34605 & -0.73986 & 0.9 & 1.50@xmath199 0.04 & 2.42@xmath199 0.12 & 3.83@xmath199 0.23 & 2.82@xmath199 0.01 & 1.45@xmath199 0.01 & 16.2 & 7.6 & 11.9 & 0.11 & 3.1 & 2 + 740 & 224.34805 & -0.51413 & 0.7 & & & 0.84@xmath199 0.06 & 0.83@xmath199 0.01 & 0.69@xmath199 0.02 & 23.9 & 4.3 & 9.1 & & 0.1 & 1 + 741 & 224.34972 & -2.01393 & 1.1 & 32.38@xmath199 0.36 & 118.75@xmath199 0.48 & 108.75@xmath199 0.76 & 57.16@xmath199 0.00 & 22.34@xmath199 0.02 & 16.5 & 83.6 & 15.9 & 6.40 & 137.8 & 2 + 742 & 224.35103 & -0.97620 & 0.9 & & & 1.12@xmath199 0.11 & 1.19@xmath199 0.01 & 1.04@xmath199 0.02 & 26.4 & 11.3 & 8.6 & & 0.2 & 1 + 743 & 224.35159 & -0.49287 & 0.7 & & & 1.89@xmath199 0.14 & 2.40@xmath199 0.01 & 1.41@xmath199 0.03 & 26.8 & 10.4 & 9.1 & & 0.2 & 1 + 744 & 224.35844 & -2.02430 & 1.1 & 5.99@xmath199 0.13 & 11.92@xmath199 0.29 & 13.82@xmath199 0.33 & 11.56@xmath199 0.00 & 5.86@xmath199 0.01 & 10.1 & 35.8 & 12.4 & 0.80 & 19.6 & 2 + 745 & 224.36429 & -0.57453 & 0.9 & & & 0.41@xmath199 0.08 & 0.49@xmath199 0.01 & 0.30@xmath199 0.02 & 13.0 & 3.3 & 9.3 & & 0.1 & 1 + 746 & 224.36475 & -2.05118 & 1.1 & 5.41@xmath199 0.05 & 9.82@xmath199 0.23 & 9.97@xmath199 0.33 & 7.22@xmath199 0.02 & & 21.7 & 15.2 & 13.6 & 0.51 & 14.5 & 2 + 747 & 224.36488 & -2.07441 & 1.1 & & 3.25@xmath199 0.11 & 7.06@xmath199 0.19 & 3.65@xmath199 0.01 & 2.76@xmath199 0.09 & 25.7 & 18.4 & 11.4 & 0.08 & 4.0 & 2 + 748 & 224.36807 & -2.08531 & 1.1 & & 7.08@xmath199 0.34 & 7.74@xmath199 0.46 & 5.82@xmath199 0.02 & & 31.1 & 12.3 & 13.6 & 0.13 & 7.0 & 2 + 749 & 224.36986 & -0.72598 & 0.9 & 3.91@xmath199 0.07 & 4.59@xmath199 0.14 & 4.32@xmath199 0.15 & 2.35@xmath199 0.01 & 0.94@xmath199 0.02 & 12.2 & 3.2 & 15.2 & 0.06 & 5.2 & 2 + 750 & 224.36990 & -2.02941 & 1.1 & 1.05@xmath199 0.05 & 3.91@xmath199 0.25 & 8.63@xmath199 0.20 & 11.05@xmath199 0.01 & & 17.1 & 90.1 & 9.5 & 0.07 & 6.2 & 2 + 751 & 224.37210 & -1.93678 & 1.1 & & & 2.75@xmath199 1.69 & 2.58@xmath199 0.02 & 1.89@xmath199 0.03 & 27.9 & 24.7 & 9.5 & & 0.6 & 1 + 752 & 224.37325 & -0.47919 & 0.7 & & & 0.79@xmath199 0.07 & 0.83@xmath199 0.01 & 0.54@xmath199 0.02 & 25.8 & 3.3 & 9.5 & & 0.1 & 1 + 753 & 224.37488 & -2.04791 & 1.1 & 2.99@xmath199 0.07 & 6.37@xmath199 0.19 & 7.57@xmath199 0.26 & 5.13@xmath199 0.02 & 2.52@xmath199 0.01 & 16.1 & 13.9 & 13.1 & 0.24 & 9.3 & 2 + 754 & 224.37862 & -1.92086 & 1.1 & & & 3.75@xmath199 0.16 & 3.10@xmath199 0.01 & 1.85@xmath199 0.04 & 30.6 & 15.7 & 11.1 & & 0.7 & 1 + 755 & 224.38319 & -1.01548 & 0.9 & 2.92@xmath199 0.07 & 5.15@xmath199 0.15 & 5.41@xmath199 0.17 & 2.74@xmath199 0.01 & 1.27@xmath199 0.02 & 12.7 & 3.2 & 15.0 & 0.16 & 4.7 & 2 + 756 & 224.38483 & -0.84225 & 0.9 & & & 1.48@xmath199 0.08 & 1.34@xmath199 0.01 & 0.73@xmath199 0.03 & 23.0 & 5.2 & 10.7 & & 0.2 & 1 + 757 & 224.39241 & -1.98039 & 1.1 & & & 1.75@xmath199 0.20 & 1.08@xmath199 0.01 & 0.57@xmath199 0.03 & 21.7 & 3.2 & 12.9 & & 0.3 & 1 + 758 & 224.39384 & -0.70904 & 0.9 & 0.86@xmath199 0.04 & 4.91@xmath199 0.14 & 8.59@xmath199 0.22 & 7.04@xmath199 0.01 & 4.33@xmath199 0.03 & 18.1 & 26.4 & 11.1 & 0.03 & 4.0 & 2 + 759 & 224.39561 & -1.95779 & 1.1 & & & 1.97@xmath199 0.11 & 1.97@xmath199 0.01 & 1.17@xmath199 0.01 & 18.5 & 14.4 & 9.9 & & 0.4 & 1 + 760 & 224.39606 & -0.72510 & 0.9 & 4.27@xmath199 0.08 & 5.76@xmath199 0.14 & 6.03@xmath199 0.23 & 3.18@xmath199 0.01 & 1.77@xmath199 0.02 & 12.4 & 5.2 & 14.3 & 0.16 & 6.7 & 2 + 761 & 224.40646 & -0.93984 & 0.9 & 0.25@xmath199 0.01 & & 0.81@xmath199 0.07 & 0.38@xmath199 0.02 & & 23.4 & 0.2 & 19.3 & & 0.3 & 2 + 762 & 224.40698 & -1.04630 & 0.9 & & & 1.56@xmath199 0.14 & 0.93@xmath199 0.02 & 0.56@xmath199 0.04 & 24.4 & 2.0 & 12.5 & & 0.1 & 1 + 763 & 224.40718 & -0.69299 & 0.9 & 14.90@xmath199 0.15 & 52.22@xmath199 0.57 & 50.50@xmath199 0.40 & 31.53@xmath199 0.03 & 12.87@xmath199 0.09 & 17.7 & 40.4 & 14.7 & 0.15 & 35.6 & 2 + 764 & 224.40807 & -0.73863 & 0.9 & 1.13@xmath199 0.04 & 2.77@xmath199 0.11 & 3.25@xmath199 0.06 & 1.55@xmath199 0.01 & 1.00@xmath199 0.01 & 6.9 & 2.8 & 14.1 & 0.01 & 2.2 & 2 + 765 & 224.41164 & -2.09036 & 1.1 & 1.73@xmath199 0.08 & 2.18@xmath199 0.11 & 1.53@xmath199 0.11 & 0.75@xmath199 0.02 & & 13.4 & 0.8 & 17.9 & 0.07 & 3.5 & 2 + 766 & 224.41360 & -1.95621 & 1.1 & & & 2.20@xmath199 0.20 & 1.91@xmath199 0.02 & 0.90@xmath199 0.01 & 24.2 & 7.2 & 11.6 & & 0.4 & 1 + 767 & 224.41409 & -1.27196 & 0.9 & & & 0.81@xmath199 0.06 & 0.66@xmath199 0.02 & 0.39@xmath199 0.03 & 20.6 & 2.6 & 10.9 & & 0.1 & 1 + 768 & 224.42081 & -0.69778 & 0.9 & 8.15@xmath199 0.11 & 35.00@xmath199 0.25 & 43.61@xmath199 0.39 & 25.01@xmath199 0.03 & 11.42@xmath199 0.08 & 18.0 & 41.7 & 13.6 & 0.71 & 26.7 & 2 + 769 & 224.42430 & -0.55163 & 0.9 & & & 2.05@xmath199 0.35 & 2.96@xmath199 0.02 & 2.07@xmath199 0.04 & 32.9 & 33.2 & 8.3 & & 0.5 & 1 + 770 & 224.42996 & -2.03730 & 1.1 & 0.41@xmath199 0.04 & 1.44@xmath199 0.23 & 2.81@xmath199 0.24 & 3.17@xmath199 0.03 & & 25.1 & 20.5 & 10.1 & 0.07 & 2.3 & 2 + 771 & 224.43195 & -0.94217 & 1.0 & & & 0.89@xmath199 0.09 & 0.71@xmath199 0.01 & 0.42@xmath199 0.02 & 27.3 & 3.4 & 10.9 & & 0.2 & 1 + 772 & 224.43466 & -0.78151 & 0.9 & & & 1.26@xmath199 0.13 & 1.26@xmath199 0.01 & 0.95@xmath199 0.04 & 23.7 & 9.9 & 9.2 & & 0.2 & 1 + 773 & 224.43645 & -0.28002 & 0.9 & & & 0.66@xmath199 0.05 & 0.58@xmath199 0.02 & 0.28@xmath199 0.02 & 21.6 & 1.8 & 11.4 & & 0.1 & 1 + 774 & 224.44051 & -0.67038 & 0.9 & 0.32@xmath199 0.02 & 0.76@xmath199 0.07 & 2.76@xmath199 0.21 & 1.96@xmath199 0.01 & 1.06@xmath199 0.03 & 26.2 & 10.2 & 9.9 & 0.13 & 1.3 & 2 + 775 & 224.44063 & -0.57954 & 0.9 & & & 1.04@xmath199 0.15 & 1.23@xmath199 0.03 & 0.91@xmath199 0.04 & 27.3 & 11.7 & 8.7 & & 0.2 & 1 + 776 & 224.44110 & -0.60414 & 0.9 & 0.39@xmath199 0.01 & 1.24@xmath199 0.08 & 2.66@xmath199 0.16 & 2.41@xmath199 0.02 & 1.03@xmath199 0.03 & 14.9 & 6.8 & 11.1 & & 1.1 & 2 + 777 & 224.44452 & -2.04066 & 1.1 & 3.36@xmath199 0.06 & 3.88@xmath199 0.12 & 8.00@xmath199 0.39 & 7.09@xmath199 0.05 & 4.39@xmath199 0.01 & 18.2 & 44.4 & 10.4 & 0.40 & 9.7 & 2 + 778 & 224.44487 & -2.10022 & 0.9 & & & 1.32@xmath199 0.16 & 1.23@xmath199 0.02 & 0.61@xmath199 0.02 & 24.5 & 5.4 & 10.1 & & 0.2 & 1 + 779 & 224.44861 & -0.31915 & 1.0 & & & 0.35@xmath199 0.05 & 0.21@xmath199 0.01 & 0.11@xmath199 0.04 & 21.6 & 0.5 & 13.2 & & 0.0 & 0 + 780 & 224.45079 & -0.70513 & 0.9 & 0.65@xmath199 0.03 & 1.08@xmath199 0.09 & 2.29@xmath199 0.23 & 1.76@xmath199 0.02 & 1.40@xmath199 0.03 & 20.8 & 8.7 & 10.4 & 0.31 & 2.1 & 2 + 781 & 224.45212 & -2.03425 & 1.1 & & 0.92@xmath199 0.10 & 2.36@xmath199 0.20 & 4.57@xmath199 0.04 & & 15.5 & 53.9 & 8.5 & 0.08 & 2.2 & 2 + 782 & 224.45479 & -1.99367 & 1.1 & & & 1.28@xmath199 0.08 & 0.83@xmath199 0.01 & 0.39@xmath199 0.01 & 20.2 & 1.8 & 13.6 & & 0.2 & 1 + 783 & 224.45528 & -1.09919 & 1.0 & & & 0.75@xmath199 0.09 & 0.33@xmath199 0.04 & 0.20@xmath199 0.01 & 19.8 & 0.6 & 14.8 & & 0.1 & 0 + 784 & 224.45647 & -0.67829 & 0.9 & 3.88@xmath199 0.08 & 5.36@xmath199 0.14 & 3.41@xmath199 0.19 & 1.50@xmath199 0.01 & 0.59@xmath199 0.02 & 6.6 & 1.0 & 19.2 & 0.85 & 7.8 & 2 + 785 & 224.45924 & -0.59018 & 0.9 & & & 1.43@xmath199 0.10 & 1.43@xmath199 0.01 & 0.81@xmath199 0.02 & 29.9 & 7.0 & 10.1 & & 0.2 & 1 + 786 & 224.46394 & -0.49888 & 1.0 & & & 0.70@xmath199 0.07 & 1.18@xmath199 0.01 & 0.92@xmath199 0.02 & 19.7 & 21.4 & 7.8 & & 0.2 & 1 + 787 & 224.46959 & -2.03874 & 1.1 & & & 4.00@xmath199 0.20 & 9.95@xmath199 0.07 & 6.52@xmath199 0.02 & 25.8 & 226.6 & 7.4 & & 1.9 & 1 + 788 & 224.47148 & -0.63852 & 0.9 & 0.65@xmath199 0.04 & & 4.83@xmath199 0.26 & 5.55@xmath199 0.03 & 2.31@xmath199 0.02 & 27.8 & 15.7 & 11.1 & 0.04 & 2.5 & 2 + 789 & 224.47331 & -0.69497 & 0.9 & 1.95@xmath199 0.05 & 6.00@xmath199 0.14 & 4.68@xmath199 0.10 & 3.83@xmath199 0.02 & 1.09@xmath199 0.02 & 11.7 & 3.1 & 15.7 & & 3.5 & 2 + 790 & 224.47818 & -0.30780 & 1.0 & & 0.20@xmath199 0.05 & 0.50@xmath199 0.05 & 0.35@xmath199 0.01 & 0.17@xmath199 0.03 & 18.6 & 1.5 & 10.9 & 0.01 & 0.2 & 2 + 791 & 224.48149 & -0.62037 & 0.9 & 3.71@xmath199 0.07 & 5.65@xmath199 0.15 & 5.61@xmath199 0.23 & 3.30@xmath199 0.01 & 1.99@xmath199 0.01 & 11.4 & 5.8 & 13.9 & 0.12 & 6.1 & 2 + 792 & 224.48274 & -0.63637 & 0.9 & 7.04@xmath199 0.11 & 17.11@xmath199 0.21 & 20.96@xmath199 0.19 & 14.92@xmath199 0.07 & 7.45@xmath199 0.01 & 19.3 & 30.0 & 12.9 & 0.84 & 17.7 & 2 + 793 & 224.48524 & -0.57152 & 0.9 & & & 1.69@xmath199 0.13 & 1.63@xmath199 0.01 & 1.32@xmath199 0.02 & 19.0 & 13.7 & 9.1 & & 0.3 & 1 + 794 & 224.48857 & -2.04304 & 1.1 & 0.61@xmath199 0.03 & 0.82@xmath199 0.08 & 3.29@xmath199 0.22 & 3.20@xmath199 0.02 & 2.51@xmath199 0.02 & 22.0 & 45.5 & 8.7 & & 1.8 & 2 + 795 & 224.49171 & -0.66233 & 0.9 & 1.03@xmath199 0.04 & 1.02@xmath199 0.09 & 3.38@xmath199 0.14 & 1.32@xmath199 0.06 & & 18.2 & 4.5 & 11.5 & 0.31 & 2.5 & 2 + 796 & 224.49431 & -0.66956 & 0.9 & & & 2.86@xmath199 0.17 & 3.83@xmath199 0.07 & 3.48@xmath199 0.01 & 21.6 & 48.8 & 8.3 & & 0.7 & 1 + 797 & 224.50319 & -1.00619 & 1.0 & 3.05@xmath199 0.07 & 3.65@xmath199 0.13 & 3.85@xmath199 0.19 & 2.15@xmath199 0.01 & 0.98@xmath199 0.01 & 16.1 & 3.9 & 14.5 & 0.65 & 7.8 & 2 + 798 & 224.50848 & -2.03491 & 0.9 & & & 3.44@xmath199 0.21 & 3.13@xmath199 0.01 & 0.98@xmath199 0.01 & 17.0 & 4.0 & 13.6 & & 0.4 & 1 + 799 & 224.50963 & -2.02642 & 1.1 & 5.42@xmath199 0.07 & 10.45@xmath199 0.19 & 11.49@xmath199 0.55 & 5.90@xmath199 0.01 & 2.74@xmath199 0.01 & 13.0 & 11.6 & 14.7 & 0.08 & 13.2 & 2 + 800 & 224.51379 & -0.90886 & 1.0 & 0.46@xmath199 0.02 & 0.58@xmath199 0.07 & 1.44@xmath199 0.11 & 1.18@xmath199 0.01 & 0.66@xmath199 0.02 & 18.5 & 6.7 & 10.4 & & 0.8 & 2 + 801 & 224.52808 & -2.03471 & 1.2 & 1.13@xmath199 0.04 & 2.19@xmath199 0.06 & 8.87@xmath199 0.28 & 7.95@xmath199 0.04 & 5.23@xmath199 0.03 & 29.5 & 104.3 & 9.1 & 0.14 & 6.5 & 2 + 802 & 224.52951 & -0.49673 & 1.0 & & & 0.58@xmath199 0.06 & 0.57@xmath199 0.01 & 0.31@xmath199 0.02 & 17.1 & 2.6 & 10.5 & & 0.1 & 1 + 803 & 224.53252 & -1.20667 & 0.9 & & & 0.84@xmath199 0.07 & 0.60@xmath199 0.02 & 0.34@xmath199 0.02 & 26.0 & 1.8 & 11.7 & & 0.1 & 1 + 804 & 224.53329 & -0.93453 & 1.0 & & & 0.48@xmath199 0.09 & 0.36@xmath199 0.01 & 0.16@xmath199 0.02 & 17.0 & 1.5 & 11.1 & & 0.1 & 1 + 805 & 224.53442 & -1.94454 & 1.2 & & & 1.12@xmath199 0.15 & 0.56@xmath199 0.02 & 0.31@xmath199 0.04 & 28.4 & 1.6 & 14.3 & & 0.2 & 0 + 806 & 224.53529 & -0.29120 & 1.0 & & & 0.78@xmath199 0.07 & 0.59@xmath199 0.02 & 0.44@xmath199 0.02 & 25.5 & 3.6 & 10.3 & & 0.1 & 1 + 807 & 224.54027 & -0.85986 & 1.0 & & & 1.23@xmath199 0.10 & 1.03@xmath199 0.02 & 0.52@xmath199 0.03 & 24.6 & 3.8 & 11.6 & & 0.2 & 1 + 808 & 224.54694 & -0.58687 & 0.9 & & & 1.94@xmath199 0.09 & 2.50@xmath199 0.06 & 1.65@xmath199 0.05 & 23.0 & 21.7 & 8.8 & & 0.4 & 1 + 809 & 224.55096 & -0.44592 & 1.0 & & & 0.44@xmath199 0.06 & 0.44@xmath199 0.01 & 0.26@xmath199 0.02 & 17.6 & 2.5 & 10.1 & & 0.1 & 1 + 810 & 224.55194 & -0.57854 & 0.9 & & & 0.95@xmath199 0.07 & 1.83@xmath199 0.05 & 0.85@xmath199 0.03 & 19.1 & 13.0 & 8.6 & & 0.2 & 1 + 811 & 224.55296 & -0.98574 & 1.0 & 1.07@xmath199 0.03 & 1.37@xmath199 0.10 & 1.53@xmath199 0.12 & 0.93@xmath199 0.02 & 0.44@xmath199 0.02 & 13.5 & 1.3 & 14.8 & 0.02 & 2.0 & 2 + 812 & 224.56255 & -0.84683 & 1.0 & & & 0.37@xmath199 0.04 & 0.35@xmath199 0.01 & 0.27@xmath199 0.02 & 15.1 & 3.4 & 9.3 & & 0.1 & 1 + 813 & 224.56438 & -0.88288 & 1.0 & & & 0.42@xmath199 0.05 & 0.36@xmath199 0.01 & 0.17@xmath199 0.03 & 21.6 & 1.2 & 11.6 & & 0.1 & 0 + 814 & 224.56554 & -1.82345 & 1.2 & & & 3.06@xmath199 0.13 & 3.33@xmath199 0.01 & 1.17@xmath199 23.36 & 28.9 & 12.5 & 11.7 & & 0.8 & 1 + 815 & 224.58261 & -0.96758 & 1.0 & & & 1.62@xmath199 0.21 & 1.80@xmath199 0.02 & 0.80@xmath199 0.03 & 29.5 & 7.8 & 10.6 & & 0.3 & 1 + 816 & 224.58485 & -2.01319 & 1.0 & & & 0.79@xmath199 0.16 & 0.83@xmath199 0.02 & 0.59@xmath199 0.02 & 23.5 & 7.6 & 9.2 & & 0.2 & 1 + 817 & 224.58771 & -0.55345 & 1.0 & & & 1.56@xmath199 0.10 & 1.81@xmath199 0.02 & 1.79@xmath199 0.02 & 28.1 & 32.2 & 8.1 & & 0.4 & 1 + 818 & 224.59193 & -1.88837 & 1.2 & & & 0.34@xmath199 0.07 & 0.48@xmath199 0.01 & 0.33@xmath199 0.01 & 21.4 & 8.8 & 8.5 & & 0.1 & 1 + 819 & 224.59361 & -0.62799 & 0.9 & & & 0.47@xmath199 0.07 & 0.44@xmath199 0.02 & 0.22@xmath199 0.03 & 20.6 & 1.4 & 11.0 & & 0.1 & 1 + 820 & 224.59508 & -2.08921 & 1.2 & & & 1.04@xmath199 0.12 & 0.67@xmath199 0.01 & 0.19@xmath199 0.04 & 30.5 & 0.7 & 17.6 & & 0.2 & 0 + 821 & 224.59752 & -0.48059 & 1.0 & & & 1.03@xmath199 0.15 & 1.10@xmath199 0.02 & 0.86@xmath199 0.03 & 28.0 & 11.6 & 8.9 & & 0.2 & 1 + 822 & 224.59766 & -1.18078 & 1.1 & & 0.24@xmath199 0.05 & 1.12@xmath199 0.21 & 0.87@xmath199 0.01 & 0.50@xmath199 0.02 & 18.6 & 8.6 & 9.1 & 0.01 & 0.5 & 2 + 823 & 224.60117 & -1.82321 & 1.2 & & 0.78@xmath199 0.09 & 1.42@xmath199 0.15 & 1.25@xmath199 0.01 & & 26.3 & 7.9 & 10.9 & & 0.7 & 1 + 824 & 224.60210 & -0.92914 & 1.0 & & & 0.94@xmath199 0.07 & 0.89@xmath199 0.05 & 0.53@xmath199 0.01 & 24.2 & 5.5 & 10.1 & & 0.2 & 1 + 825 & 224.60228 & -1.66645 & 0.9 & & & 0.65@xmath199 0.10 & 0.48@xmath199 0.01 & 0.23@xmath199 0.04 & 23.5 & 1.0 & 12.4 & & 0.1 & 0 + 826 & 224.60686 & -0.90836 & 1.0 & & & 0.42@xmath199 0.09 & 0.34@xmath199 0.01 & 0.26@xmath199 0.01 & 17.5 & 3.0 & 9.7 & & 0.1 & 1 + 827 & 224.60736 & -1.00611 & 1.0 & 165.79@xmath199 0.50 & 103.63@xmath199 0.41 & 57.05@xmath199 0.46 & 23.51@xmath199 0.02 & 10.56@xmath199 0.01 & 16.2 & 18.4 & 20.4 & 28.75 & 324.2 & 2 + 828 & 224.61028 & -0.45717 & 1.0 & & & 1.01@xmath199 0.07 & 1.19@xmath199 0.03 & 0.47@xmath199 0.04 & 23.2 & 4.1 & 10.8 & & 0.2 & 1 + 829 & 224.61038 & -1.00273 & 1.0 & 42.50@xmath199 0.25 & 60.76@xmath199 0.24 & 61.95@xmath199 0.62 & 20.80@xmath199 0.02 & & 22.7 & 22.7 & 17.8 & 0.44 & 75.9 & 2 + 830 & 224.61362 & -0.49869 & 1.0 & & & 0.76@xmath199 0.12 & 0.80@xmath199 0.02 & 0.68@xmath199 0.03 & 23.1 & 9.6 & 8.7 & & 0.2 & 1 + 831 & 224.61528 & -1.61775 & 1.0 & & & 1.16@xmath199 0.08 & 1.23@xmath199 0.01 & 0.77@xmath199 0.02 & 26.0 & 8.3 & 9.6 & & 0.2 & 1 + 832 & 224.61574 & -1.33985 & 1.1 & & & 0.32@xmath199 0.06 & 0.26@xmath199 0.01 & 0.18@xmath199 0.01 & 18.2 & 1.8 & 10.3 & & 0.1 & 1 + 833 & 224.61826 & -1.13431 & 1.0 & & & 0.70@xmath199 0.08 & 0.62@xmath199 0.01 & 0.49@xmath199 0.01 & 23.6 & 5.5 & 9.5 & & 0.1 & 1 + 834 & 224.62444 & -2.03529 & 1.0 & & & 0.57@xmath199 0.11 & 0.42@xmath199 0.02 & 0.20@xmath199 0.02 & 20.8 & 1.2 & 12.3 & & 0.1 & 0 + 835 & 224.62511 & -0.48293 & 1.0 & & & 1.18@xmath199 0.10 & 1.42@xmath199 0.02 & 1.24@xmath199 0.05 & 24.3 & 21.8 & 8.3 & & 0.3 & 1 + 836 & 224.62695 & -0.44340 & 1.0 & & & 0.78@xmath199 0.12 & 0.92@xmath199 0.01 & 0.60@xmath199 0.02 & 23.6 & 8.2 & 9.1 & & 0.2 & 1 + 837 & 224.62845 & -1.61949 & 1.0 & & & 0.80@xmath199 0.16 & 1.31@xmath199 0.03 & 0.59@xmath199 0.02 & 28.9 & 8.0 & 9.2 & & 0.2 & 1 + 838 & 224.64122 & -0.79287 & 1.0 & & & 0.52@xmath199 0.04 & 0.42@xmath199 0.01 & 0.24@xmath199 0.09 & 20.2 & 1.6 & 11.6 & & 0.1 & 1 + 839 & 224.64362 & -0.88681 & 1.0 & & & 0.85@xmath199 0.06 & 0.69@xmath199 0.01 & 0.31@xmath199 0.04 & 24.5 & 2.1 & 12.0 & & 0.1 & 1 + 840 & 224.64377 & -1.60225 & 1.0 & & & 0.85@xmath199 0.09 & 0.62@xmath199 0.02 & 0.32@xmath199 0.02 & 20.1 & 1.8 & 12.0 & & 0.1 & 1 + 841 & 224.64528 & -1.02545 & 1.0 & & & 2.19@xmath199 0.11 & 3.06@xmath199 0.02 & 1.11@xmath199 0.04 & 23.8 & 11.6 & 10.5 & & 0.5 & 1 + 842 & 224.64560 & -1.12209 & 1.0 & & & 0.87@xmath199 0.08 & 0.60@xmath199 0.02 & 0.26@xmath199 0.02 & 25.8 & 1.3 & 13.2 & & 0.1 & 0 + 843 & 224.64622 & -0.80120 & 1.0 & & & 0.65@xmath199 0.05 & 0.48@xmath199 0.01 & 0.21@xmath199 0.11 & 23.3 & 1.2 & 12.8 & & 0.1 & 0 + 844 & 224.64824 & -0.43994 & 1.0 & & & 0.68@xmath199 0.11 & 0.83@xmath199 0.01 & 0.67@xmath199 0.01 & 12.1 & 11.2 & 8.4 & & 0.2 & 1 + 845 & 224.64903 & -0.54439 & 1.0 & & & 0.82@xmath199 0.06 & 0.91@xmath199 0.03 & 0.41@xmath199 0.03 & 26.0 & 3.7 & 10.5 & & 0.2 & 1 + 846 & 224.65298 & -1.58222 & 1.0 & & & 1.11@xmath199 0.07 & 0.95@xmath199 0.01 & 0.44@xmath199 0.03 & 24.1 & 2.7 & 11.7 & & 0.2 & 1 + 847 & 224.65338 & -1.76353 & 1.0 & & & 3.16@xmath199 0.11 & 2.76@xmath199 0.01 & 1.60@xmath199 0.03 & 25.1 & 14.5 & 10.6 & & 0.6 & 1 + 848 & 224.65628 & -1.89188 & 1.2 & & & 0.42@xmath199 0.06 & 0.28@xmath199 0.02 & 0.14@xmath199 0.02 & 17.6 & 1.0 & 13.0 & & 0.1 & 0 + 849 & 224.65878 & -1.78735 & 1.2 & & & 1.46@xmath199 0.10 & 0.96@xmath199 0.02 & 0.64@xmath199 0.02 & 27.0 & 6.0 & 11.4 & & 0.3 & 1 + 850 & 224.66942 & -1.56488 & 1.0 & & & 1.00@xmath199 0.17 & 0.91@xmath199 0.02 & 0.48@xmath199 0.03 & 26.2 & 3.6 & 10.8 & & 0.2 & 1 + 851 & 224.67023 & -0.48120 & 1.0 & & & 0.38@xmath199 0.08 & 0.51@xmath199 0.03 & 0.25@xmath199 0.01 & 17.9 & 3.1 & 9.5 & & 0.1 & 1 + 852 & 224.67027 & -0.39621 & 1.0 & & & 1.11@xmath199 0.14 & 1.23@xmath199 0.01 & 0.74@xmath199 0.04 & 24.6 & 8.7 & 9.5 & & 0.2 & 1 + 853 & 224.67284 & -1.03862 & 1.1 & & & 1.23@xmath199 0.07 & 1.06@xmath199 0.01 & 0.65@xmath199 0.02 & 17.9 & 7.1 & 10.5 & & 0.3 & 1 + 854 & 224.68654 & -1.49524 & 1.0 & & & 0.57@xmath199 0.08 & 0.44@xmath199 0.02 & 0.20@xmath199 0.03 & 25.5 & 1.1 & 12.3 & & 0.1 & 0 + 855 & 224.69101 & -0.99227 & 1.1 & & & 2.21@xmath199 0.19 & 1.50@xmath199 0.03 & 0.90@xmath199 0.04 & 34.9 & 6.8 & 11.7 & & 0.4 & 1 + 856 & 224.69597 & -1.66223 & 1.0 & & & 0.93@xmath199 0.07 & 0.49@xmath199 0.01 & 0.37@xmath199 0.02 & 23.3 & 1.9 & 12.0 & & 0.1 & 1 + 857 & 224.71361 & -1.82883 & 1.2 & & & 1.92@xmath199 0.19 & 1.20@xmath199 0.01 & 0.93@xmath199 0.03 & 31.8 & 8.3 & 11.1 & & 0.4 & 1 + 858 & 224.71526 & -0.99539 & 1.1 & 0.78@xmath199 0.03 & 2.68@xmath199 0.14 & 4.98@xmath199 0.17 & 3.61@xmath199 0.03 & 1.59@xmath199 0.02 & 23.4 & 13.8 & 11.8 & 0.13 & 4.2 & 2 + 859 & 224.72194 & -0.99885 & 1.1 & & & 4.05@xmath199 0.11 & 2.74@xmath199 0.01 & 1.56@xmath199 0.02 & 25.6 & 12.6 & 11.7 & & 0.7 & 1 + 860 & 224.72362 & -0.23369 & 0.7 & & & 0.33@xmath199 0.05 & 0.23@xmath199 0.02 & 0.13@xmath199 0.03 & 21.6 & 0.4 & 11.7 & & 0.0 & 0 + 861 & 224.72626 & -1.58336 & 1.0 & & & 0.51@xmath199 0.09 & 0.59@xmath199 0.01 & 0.49@xmath199 0.01 & 14.3 & 8.6 & 8.5 & & 0.1 & 1 + 862 & 224.73573 & -1.82862 & 1.0 & & & 2.03@xmath199 0.10 & 1.87@xmath199 0.05 & 1.02@xmath199 0.04 & 27.0 & 10.5 & 10.1 & & 0.4 & 1 + 863 & 224.73793 & -0.33488 & 1.0 & & & 0.51@xmath199 0.05 & 0.35@xmath199 0.01 & 0.17@xmath199 0.02 & 17.3 & 0.8 & 12.7 & & 0.1 & 0 + 864 & 224.74095 & -1.81590 & 1.0 & & & 2.88@xmath199 0.13 & 2.09@xmath199 0.02 & 1.92@xmath199 0.06 & 28.1 & 18.4 & 9.7 & & 0.5 & 1 + 865 & 224.75027 & -1.56005 & 1.2 & & & 0.93@xmath199 0.14 & 0.68@xmath199 0.02 & 0.57@xmath199 0.02 & 28.0 & 6.9 & 10.1 & & 0.2 & 1 + 866 & 224.75464 & -1.01729 & 1.1 & & & 1.78@xmath199 0.11 & 2.77@xmath199 0.02 & 2.06@xmath199 0.03 & 26.0 & 57.6 & 8.0 & & 0.7 & 1 + 867 & 224.75885 & -1.75132 & 1.0 & & & 3.63@xmath199 0.30 & 5.71@xmath199 0.02 & 3.50@xmath199 0.02 & 28.8 & 71.1 & 8.4 & & 1.1 & 1 + 868 & 224.76343 & -0.20556 & 1.0 & 1.22@xmath199 0.05 & 3.29@xmath199 0.09 & 2.86@xmath199 0.11 & 1.38@xmath199 0.02 & 0.54@xmath199 0.03 & 14.6 & 1.4 & 16.8 & 0.05 & 3.0 & 2 + 869 & 224.76361 & -0.48820 & 1.0 & & & 0.72@xmath199 0.06 & 0.60@xmath199 0.01 & 0.40@xmath199 0.01 & 20.9 & 3.3 & 10.2 & & 0.1 & 1 + 870 & 224.76541 & -1.81522 & 1.0 & 0.65@xmath199 0.03 & 3.32@xmath199 0.14 & 5.37@xmath199 0.17 & 3.38@xmath199 0.01 & 2.00@xmath199 0.01 & 14.3 & 11.9 & 12.1 & & 2.8 & 2 + 871 & 224.76581 & -1.49407 & 1.2 & & & 0.98@xmath199 0.17 & 0.56@xmath199 0.01 & 0.28@xmath199 0.03 & 25.4 & 1.7 & 13.8 & & 0.2 & 0 + 872 & 224.78360 & -1.81372 & 1.0 & & & 7.36@xmath199 0.25 & 4.62@xmath199 0.03 & 2.38@xmath199 0.02 & 35.7 & 12.1 & 13.0 & & 1.0 & 1 + 873 & 224.78464 & -1.57732 & 1.2 & & & 0.95@xmath199 0.05 & 0.69@xmath199 0.02 & 0.35@xmath199 0.02 & 22.1 & 3.1 & 12.1 & & 0.2 & 1 + 874 & 224.79005 & -1.69666 & 1.0 & & & 2.50@xmath199 0.09 & 1.67@xmath199 0.01 & 0.96@xmath199 0.01 & 19.6 & 6.2 & 11.7 & & 0.4 & 1 + 875 & 224.79051 & -1.73547 & 1.0 & 56.16@xmath199 0.28 & 68.84@xmath199 0.48 & 42.87@xmath199 0.26 & 20.46@xmath199 0.02 & 8.27@xmath199 0.02 & 21.0 & 15.7 & 19.3 & 6.45 & 119.7 & 2 + 876 & 224.79312 & -0.32209 & 1.0 & & & 0.58@xmath199 0.07 & 0.54@xmath199 0.04 & 0.28@xmath199 0.02 & 24.8 & 1.8 & 11.1 & & 0.1 & 1 + 877 & 224.79527 & -0.55227 & 1.0 & & & 0.60@xmath199 0.05 & 0.47@xmath199 0.02 & 0.11@xmath199 0.16 & 24.3 & 0.3 & 17.1 & & 0.1 & 0 + 878 & 224.79610 & -0.39721 & 1.0 & & & 0.41@xmath199 0.05 & 0.32@xmath199 0.01 & 0.17@xmath199 0.02 & 19.8 & 1.1 & 11.3 & & 0.1 & 1 + 879 & 224.79659 & -1.26037 & 0.7 & & & 0.23@xmath199 0.41 & 0.21@xmath199 0.01 & 0.13@xmath199 0.02 & 16.1 & 0.7 & 10.1 & & 0.0 & 1 + 880 & 224.80194 & -1.53094 & 1.2 & & & 0.96@xmath199 0.09 & 0.72@xmath199 0.02 & 0.40@xmath199 0.01 & 23.3 & 4.0 & 11.4 & & 0.2 & 1 + 881 & 224.80695 & -1.21035 & 0.7 & & & 0.45@xmath199 0.03 & 0.30@xmath199 0.01 & 0.17@xmath199 0.03 & 20.9 & 0.5 & 12.1 & & 0.0 & 0 + 882 & 224.80702 & -0.62801 & 1.0 & & & 0.44@xmath199 0.04 & 0.32@xmath199 0.02 & 0.17@xmath199 0.02 & 23.6 & 1.0 & 11.9 & & 0.1 & 0 + 883 & 224.81694 & -1.80619 & 1.0 & & & 1.07@xmath199 0.08 & 0.78@xmath199 0.01 & 0.61@xmath199 0.01 & 17.3 & 5.3 & 10.3 & & 0.2 & 1 + 884 & 224.82162 & -0.18720 & 1.0 & & & 0.30@xmath199 0.04 & 0.19@xmath199 0.02 & 0.12@xmath199 0.02 & 19.2 & 0.6 & 12.0 & & 0.0 & 0 + 885 & 224.83861 & -1.79430 & 1.0 & & & 1.16@xmath199 0.08 & 0.82@xmath199 0.02 & 0.60@xmath199 0.03 & 25.0 & 4.3 & 11.1 & & 0.2 & 1 + 886 & 224.87282 & -1.50561 & 1.2 & & & 0.37@xmath199 0.04 & 0.31@xmath199 0.01 & 0.25@xmath199 0.04 & 17.1 & 3.9 & 9.6 & & 0.1 & 1 + 887 & 224.88478 & -1.56939 & 1.2 & & & 0.19@xmath199 0.06 & 0.21@xmath199 0.02 & 0.17@xmath199 0.03 & 17.4 & 3.9 & 8.6 & & 0.1 & 1 + 888 & 224.89590 & -0.47621 & 1.0 & & & 0.20@xmath199 0.08 & 0.22@xmath199 0.01 & 0.15@xmath199 0.02 & 21.9 & 2.0 & 9.1 & & 0.0 & 1 + 889 & 224.89636 & -1.25150 & 0.5 & 5.43@xmath199 0.09 & 11.80@xmath199 0.19 & 10.17@xmath199 0.22 & 5.14@xmath199 0.01 & 2.02@xmath199 0.01 & 10.4 & 1.4 & 16.5 & 0.03 & 2.6 & 2 + 890 & 224.90585 & -1.06103 & 0.7 & 0.24@xmath199 0.01 & 0.38@xmath199 0.06 & 0.21@xmath199 0.05 & 0.09@xmath199 0.01 & & 10.8 & 0.0 & 21.9 & 0.01 & 0.2 & 2 + 891 & 224.93352 & -0.36715 & 1.0 & & & 0.49@xmath199 0.17 & 0.35@xmath199 0.01 & 0.22@xmath199 0.02 & 26.9 & 1.4 & 11.1 & & 0.1 & 1 + 892 & 224.96158 & -1.17452 & 0.5 & & & 1.09@xmath199 0.07 & 0.76@xmath199 0.02 & 0.46@xmath199 0.04 & 23.6 & 0.7 & 11.5 & & 0.0 & 1 + 893 & 224.98445 & -0.56287 & 0.9 & & & 0.94@xmath199 0.07 & 0.75@xmath199 0.03 & 0.64@xmath199 0.02 & 27.7 & 5.7 & 9.6 & & 0.2 & 1 + 894 & 225.00194 & -0.57490 & 0.9 & & & 0.64@xmath199 0.10 & 0.54@xmath199 0.04 & 0.49@xmath199 0.02 & 24.1 & 4.8 & 9.3 & & 0.1 & 1 + 895 & 225.01230 & -1.59009 & 1.2 & & & 0.34@xmath199 0.03 & 0.27@xmath199 0.01 & 0.18@xmath199 0.01 & 17.1 & 2.4 & 10.1 & & 0.1 & 1 + 896 & 225.02200 & -0.29787 & 1.1 & & & 0.30@xmath199 0.07 & 0.29@xmath199 0.01 & 0.17@xmath199 0.01 & 16.8 & 1.8 & 10.2 & & 0.1 & 1 + 897 & 225.02840 & -1.06934 & 0.7 & & & 0.28@xmath199 0.07 & 0.25@xmath199 0.01 & 0.16@xmath199 0.02 & 12.7 & 0.7 & 10.3 & & 0.0 & 1 + 898 & 225.03256 & -1.09176 & 0.5 & & & 1.48@xmath199 0.08 & 1.08@xmath199 0.01 & 0.69@xmath199 0.02 & 25.0 & 1.2 & 11.0 & & 0.1 & 1 + 899 & 225.06194 & -1.09620 & 0.5 & & & 1.20@xmath199 0.09 & 1.03@xmath199 0.01 & 0.45@xmath199 0.04 & 23.1 & 0.7 & 11.9 & & 0.0 & 1 + 900 & 225.08028 & -1.06736 & 0.8 & & & 0.29@xmath199 0.06 & 0.23@xmath199 0.01 & 0.12@xmath199 0.02 & 13.4 & 0.5 & 11.6 & & 0.0 & 0 + 901 & 225.08528 & -1.03807 & 0.8 & & & 1.55@xmath199 0.08 & 0.96@xmath199 0.03 & 0.55@xmath199 0.04 & 34.3 & 1.8 & 12.5 & & 0.1 & 1 + 902 & 225.08731 & -0.60464 & 0.9 & & & 0.73@xmath199 0.11 & 0.78@xmath199 0.01 & 0.70@xmath199 0.03 & 24.2 & 9.8 & 8.5 & & 0.1 & 1 + 903 & 225.12480 & -0.97596 & 0.8 & & & 0.55@xmath199 0.07 & 0.44@xmath199 0.03 & 0.28@xmath199 0.02 & 24.1 & 1.5 & 10.7 & & 0.1 & 1 + 904 & 225.14886 & -0.97551 & 0.9 & & & 0.34@xmath199 0.07 & 0.57@xmath199 0.01 & 0.23@xmath199 0.01 & 13.4 & 2.4 & 9.5 & & 0.1 & 1 + 905 & 225.15724 & -0.98628 & 0.8 & & & 0.96@xmath199 0.10 & 1.15@xmath199 0.01 & 0.64@xmath199 0.02 & 21.2 & 5.1 & 9.5 & & 0.1 & 1 + 906 & 225.16586 & -0.30381 & 0.9 & & & 1.13@xmath199 0.22 & 0.72@xmath199 0.03 & 0.38@xmath199 0.02 & 37.8 & 1.5 & 12.8 & & 0.1 & 0 + 907 & 225.17249 & -0.89954 & 0.9 & & & 1.36@xmath199 0.23 & 1.17@xmath199 0.01 & 0.62@xmath199 0.04 & 30.4 & 3.7 & 11.0 & & 0.2 & 1 + 908 & 225.17525 & -0.91053 & 0.9 & & & 1.09@xmath199 0.08 & 1.19@xmath199 0.01 & 0.57@xmath199 0.04 & 28.4 & 4.1 & 10.4 & & 0.2 & 1 + 909 & 225.17958 & -0.24130 & 0.9 & & & 0.57@xmath199 0.05 & 0.43@xmath199 0.01 & 0.41@xmath199 0.01 & 17.8 & 3.0 & 9.7 & & 0.1 & 1 + 910 & 225.19249 & -0.84152 & 0.9 & & & 1.11@xmath199 0.17 & 0.72@xmath199 0.02 & 0.52@xmath199 0.02 & 35.7 & 2.6 & 11.1 & & 0.1 & 1 + 911 & 225.21861 & -0.64243 & 1.3 & & & 0.43@xmath199 0.04 & 0.23@xmath199 0.01 & 0.14@xmath199 0.02 & 17.6 & 0.9 & 13.3 & & 0.1 & 0 + 912 & 225.22267 & -0.26993 & 0.9 & & 0.58@xmath199 0.06 & 1.07@xmath199 0.10 & 0.83@xmath199 0.02 & 0.29@xmath199 0.02 & 14.3 & 1.5 & 12.3 & & 0.3 & 1 + 913 & 225.22701 & -0.96323 & 0.9 & & & 0.57@xmath199 0.07 & 0.41@xmath199 0.01 & 0.28@xmath199 0.02 & 25.1 & 1.6 & 10.8 & & 0.1 & 1 + 914 & 225.24666 & -0.23131 & 1.1 & & & 1.89@xmath199 0.10 & 1.78@xmath199 0.01 & 0.82@xmath199 0.09 & 26.0 & 7.2 & 11.2 & & 0.4 & 1 + 915 & 225.25230 & -0.55593 & 1.1 & & & 1.31@xmath199 0.17 & 0.92@xmath199 0.05 & 0.47@xmath199 0.03 & 26.4 & 3.2 & 12.2 & & 0.2 & 1 + 916 & 225.25284 & -1.15143 & 0.8 & 0.34@xmath199 0.01 & 0.42@xmath199 0.03 & 1.31@xmath199 0.06 & 1.14@xmath199 0.01 & 0.69@xmath199 0.02 & 16.7 & 5.4 & 9.7 & 0.06 & 0.7 & 2 + 917 & 225.27676 & -0.59287 & 1.1 & & & 0.86@xmath199 0.07 & 0.59@xmath199 0.01 & 0.36@xmath199 0.02 & 26.8 & 2.8 & 11.5 & & 0.1 & 1 + 918 & 225.29135 & -0.58120 & 1.1 & & & 5.40@xmath199 1.27 & 1.52@xmath199 0.06 & 1.31@xmath199 0.06 & 79.0 & 1.8 & 19.4 & & 0.6 & 0 + 919 & 225.29695 & -0.55935 & 1.1 & & & 0.84@xmath199 0.07 & 0.62@xmath199 0.01 & 0.32@xmath199 0.03 & 19.9 & 2.3 & 12.0 & & 0.1 & 1 + 920 & 225.31511 & -0.27625 & 1.1 & & & 0.57@xmath199 0.11 & 0.79@xmath199 0.01 & 0.49@xmath199 0.02 & 15.1 & 9.3 & 8.7 & & 0.2 & 1 + 921 & 225.31541 & -0.57747 & 1.1 & & & 0.65@xmath199 0.12 & 0.39@xmath199 0.01 & 0.17@xmath199 0.01 & 21.0 & 0.8 & 14.2 & & 0.1 & 0 + 922 & 225.31949 & -1.14295 & 0.8 & & & 0.99@xmath199 0.07 & 0.69@xmath199 0.02 & 0.41@xmath199 0.03 & 26.0 & 1.7 & 11.6 & & 0.1 & 1 + 923 & 225.32629 & -0.53244 & 1.1 & 16.54@xmath199 0.15 & 17.71@xmath199 0.25 & 14.84@xmath199 0.07 & 7.83@xmath199 0.02 & 2.82@xmath199 0.03 & 19.1 & 9.9 & 16.7 & 3.43 & 44.2 & 2 + 924 & 225.33392 & -1.11957 & 0.8 & & & 0.77@xmath199 0.10 & 0.58@xmath199 0.03 & 0.51@xmath199 0.01 & 25.6 & 4.1 & 9.1 & & 0.1 & 1 + 925 & 225.34178 & -1.07988 & 0.8 & & & 0.84@xmath199 0.06 & 0.75@xmath199 0.01 & 0.36@xmath199 0.03 & 23.9 & 1.9 & 11.1 & & 0.1 & 1 + 926 & 225.34790 & -0.27405 & 1.0 & & & 2.15@xmath199 0.10 & 2.04@xmath199 0.02 & 1.27@xmath199 0.03 & 29.9 & 11.7 & 10.0 & & 0.4 & 1 + 927 & 225.36028 & -0.20453 & 1.0 & & & 0.34@xmath199 0.05 & 0.24@xmath199 0.01 & 0.20@xmath199 0.01 & 15.1 & 1.5 & 10.1 & & 0.1 & 1 + 928 & 225.36090 & -0.40918 & 1.1 & & & 0.78@xmath199 0.18 & 0.98@xmath199 0.01 & 0.60@xmath199 0.02 & 23.7 & 10.4 & 9.1 & & 0.2 & 1 + 929 & 225.36206 & -0.22237 & 1.0 & & & 0.54@xmath199 0.10 & 0.36@xmath199 0.01 & 0.20@xmath199 0.02 & 22.1 & 1.1 & 12.0 & & 0.1 & 0 + 930 & 225.39027 & -0.25121 & 1.0 & & & 1.88@xmath199 37.70 & 1.63@xmath199 0.02 & 0.90@xmath199 0.03 & 40.8 & 6.6 & 10.8 & & 0.3 & 1 + 931 & 225.39120 & -0.98573 & 0.6 & & & 0.25@xmath199 0.05 & 0.34@xmath199 0.01 & 0.16@xmath199 0.01 & 12.0 & 0.6 & 9.6 & & 0.0 & 1 + 932 & 225.39833 & -1.02287 & 0.6 & & & 1.26@xmath199 0.06 & 1.06@xmath199 0.02 & 0.25@xmath199 0.05 & 22.9 & 0.2 & 17.0 & & 0.1 & 0 + 933 & 225.40114 & -0.56134 & 0.6 & 4.26@xmath199 0.08 & 4.61@xmath199 0.13 & 4.70@xmath199 0.20 & 2.61@xmath199 0.05 & 1.50@xmath199 0.01 & 15.3 & 2.1 & 14.2 & 0.19 & 3.1 & 2 + 934 & 225.41173 & -0.35974 & 1.1 & & & 0.34@xmath199 0.06 & 0.44@xmath199 0.01 & 0.27@xmath199 0.01 & 17.9 & 4.7 & 9.1 & & 0.1 & 1 + 935 & 225.42535 & -0.31206 & 1.0 & & & 0.38@xmath199 0.18 & 0.60@xmath199 0.06 & 0.35@xmath199 0.02 & 20.2 & 5.8 & 8.6 & & 0.1 & 1 + 936 & 225.43135 & -0.39845 & 1.1 & & 0.54@xmath199 0.07 & 1.25@xmath199 0.13 & 0.98@xmath199 0.01 & 0.61@xmath199 0.03 & 20.0 & 6.6 & 10.4 & 0.08 & 1.2 & 2 + 937 & 225.43605 & -0.38709 & 1.1 & & & 0.89@xmath199 0.09 & 0.81@xmath199 0.02 & 0.66@xmath199 0.04 & 23.6 & 9.1 & 9.3 & & 0.2 & 1 + 938 & 225.44194 & -0.24260 & 1.0 & & & 0.79@xmath199 0.07 & 0.83@xmath199 0.01 & 0.62@xmath199 0.03 & 22.7 & 7.6 & 9.1 & & 0.2 & 1 + 939 & 225.45551 & -0.40621 & 1.1 & & & 0.85@xmath199 0.12 & 1.18@xmath199 0.01 & 0.77@xmath199 0.05 & 24.5 & 16.3 & 8.6 & & 0.3 & 1 + 940 & 225.47861 & -0.23307 & 1.0 & & & 2.50@xmath199 0.11 & 2.86@xmath199 0.01 & 1.40@xmath199 0.04 & 30.4 & 13.4 & 10.1 & & 0.5 & 1 + 941 & 225.48128 & -0.24487 & 1.0 & & & 1.44@xmath199 0.17 & 1.79@xmath199 0.01 & 1.33@xmath199 0.05 & 31.0 & 20.2 & 8.6 & & 0.3 & 1 + 942 & 225.48631 & -0.32224 & 1.1 & & & 0.77@xmath199 0.13 & 1.00@xmath199 0.01 & 0.67@xmath199 0.02 & 18.4 & 13.2 & 8.7 & & 0.2 & 1 + 943 & 225.48694 & -0.40187 & 1.1 & & & 1.07@xmath199 0.07 & 0.92@xmath199 0.01 & 0.47@xmath199 0.02 & 17.3 & 4.4 & 11.1 & & 0.2 & 1 +	we present the first herschel pacs and spire photometric observations in a portion of the outer galaxy ( @xmath0 and @xmath1 ) as a part of the hi - gal survey . the maps between 70 and 500 @xmath2 m , the derived column density and temperature maps , and the compact source catalog are presented . nanten co(1 - 0 ) line observations are used to derive cloud kinematics and distances , so that we can estimate distance - dependent physical parameters of the compact sources ( cores and clumps ) having a reliable spectral energy distribution , that we separate in 255 proto - stellar and 688 starless . both typologies are found in association with all the distance components observed in the field , up to @xmath3 kpc , testifying the presence of star formation beyond the perseus arm at these longitudes . selecting the starless gravitationally bound sources we identify 590 pre - stellar candidates . several sources of both proto- and pre - stellar nature are found to exceed the minimum requirement for being compatible with massive star formation , based on the mass - radius relation . for the pre - stellar sources belonging to the local arm ( @xmath4 kpc ) we study the mass function , whose high - mass end shows a power - law @xmath5 . finally , we use a luminosity vs mass diagram to infer the evolutionary status of the sources , finding that most of the proto - stellar are in the early accretion phase ( with some cases compatible with a class i stage ) , while for pre - stellar sources , in general , accretion has not started yet .
in this paper we identify hyperbolic surfaces with quotients of the poincar upper halfplane @xmath3 by discrete torsion free subgroups of psl@xmath4 called _ fuchsian groups_. the _ laplacian _ on @xmath5 is the differential operator @xmath6 which associates to a @xmath7- function @xmath8 the function @xmath9 for any fuchsian group @xmath10 , the induced differential operator on @xmath11 , @xmath12 is called the laplacian on @xmath13 . it is a non - positive operator whose spectrum spec@xmath14 is contained in a smallest interval @xmath15 \subset { \mathbb r^{- } } \cup \{0\}$ ] with @xmath16 . points in the discrete spectrum will be referred to as an _ eigenvalue_. in particular this means @xmath17 is an eigenvalue if there exists a non - zero @xmath7-function @xmath18 , called a @xmath19-_eigenfunction _ , such that @xmath20 when @xmath21 , @xmath19 is called a _ small eigenvalue _ and @xmath8 is called a _ small eigenfunction_. we shall restrict ourselves to hyperbolic surfaces with finite area . any such surface @xmath13 is homeomorphic to a closed riemann surface @xmath22 of certain genus @xmath23 from which some @xmath24 many points are removed . in that case @xmath13 is called a finite area hyperbolic surface of type @xmath25 . each of these @xmath24 points is called a _ puncture _ of @xmath13 . the laplace spectrum of a closed hyperbolic surface @xmath13 consists of a discrete set : @xmath26 such that @xmath27 as @xmath28 . each number in the above sequence is repeated according to its multiplicity as eigenvalue . the number @xmath29 is called the @xmath30-th eigenvalue of @xmath13 . it is known that the map @xmath31 that assigns a surface @xmath32 to its @xmath30-th eigenvalue @xmath29 is continuous and bounded @xcite . hence @xmath33 for non - compact hyperbolic surfaces of finite area the spectrum of the laplacian is more complicated . it consists of both continuous and discrete components ( see @xcite for detail ) . however , the part of the spectrum lying in @xmath34 is discrete . keeping resemblance to above definition for any hyperbolic surface @xmath13 let us define @xmath35 to be the smallest positive number in spec@xmath14 . in particular , if @xmath36 then it is an eigenvalue . the function @xmath0 , so defined , is bounded by @xmath37 because @xmath13 has a continuous spectrum on @xmath38 . as before we consider the quantity @xmath39 in @xcite atle selberg proved that for any congruence subgroup @xmath10 of sl@xmath40 @xmath41 recall that a congruence subgroup is a discrete subgroup of sl@xmath40 that contains one of the @xmath42 where @xmath43 is the principal congruence subgroup of level @xmath24 . moreover he conjectured [ sel ] for any congruence subgroup @xmath10 , @xmath44 . m. n. huxley @xcite proved this conjecture for @xmath42 with @xmath45 . several attempts have been made to prove it ( see ( * ? ? ? * chapter 11 ) for details ) in the general case . the best known bound is @xmath46 due to kim and sarnak @xcite . this conjecture motivated , in particular , the question of our interest : [ ques ] given any genus @xmath47 does there exist a closed hyperbolic surface of genus @xmath23 with @xmath0 at least @xmath37 ? a slightly weaker question than the above one would be : is @xmath48 ? this question is studied in @xcite by p. buser , m. burger and j. dodziuk and in @xcite by r. brooks and e. makover . the ideas in @xcite and @xcite , in the light of the bound of kim and sarnak in @xcite , provide the following . [ aprox ] given any @xmath49 , there exists @xmath50 such that for any @xmath51 there exist closed hyperbolic surfaces of genus @xmath23 with @xmath52 . the constant @xmath46 in the above theorem can be replaced by @xmath37 if conjecture [ sel ] is true . hence it is tempting to conjecture : [ conj ] for every @xmath53 there exists a closed hyperbolic surface of genus @xmath23 whose @xmath0 is at least @xmath37 . observe that even if selberg s conjecture ( conjecture [ sel ] ) is true , theorem [ aprox ] do not provide a positive answer to conjecture [ conj ] . however it would answer positively the weaker version of our question i.e. it would imply @xmath54 , for large values of @xmath23 . the existence of genus two hyperbolic surfaces with @xmath55 has been known in the literature for sometime @xcite . it is known that the _ bolza _ surface has @xmath0 approximately @xmath56 ( see @xcite for more details ) . we consider the subset @xmath57 of the moduli space @xmath58 . from the continuity of @xmath0 it is clear that @xmath59 is open . our first result provides better understanding of this set . recall that the moduli space @xmath60 is the quotient of @xmath61 by the _ teichmller modular group _ @xmath62 ( see @xcite ) . we are shifting from the moduli space to the teichmller space mainly because we wish to talk about analytic paths which involves coordinates and on @xmath61 one has the fenchel - nielsen coordinates ( given a pants decomposition ) which is easy to describe . let @xmath63 \to { \mathcal{t}_2}$ ] be an analytic path . since , in this case , @xmath0 is simple as long as small , the function @xmath64 ( @xmath65 ) is also analytic ( see theorem [ b - c ] ) if @xmath66 for all @xmath67 $ ] . for higher genus @xmath0 may not be simple even if small ( see [ mult ] ) . therefore , for an analytic path @xmath68 \to { \mathcal{t}_g}$ ] , @xmath64 is continuous but need not be analytic even if @xmath66 for all @xmath67 $ ] . however we have the following result from ( * ? ? ? * theorem 14.9.3 ) : [ b - c ] let @xmath69 be a real analytic path in @xmath61 . then there exist real analytic functions @xmath70 such that for each @xmath71 the sequence @xmath72 consist of all eigenvalues of @xmath73 ( listed with multiplicities , though not in increasing order ) . each function @xmath74 is called a branch of eigenvalues along @xmath73 . more precisely let @xmath75 \to { \mathcal{t}_g}$ ] be an analytic path . an analytic function @xmath76 \to \mathbb{r}$ ] is called a branch of an eigenvalue along @xmath77 if , for each @xmath78 , @xmath79 is an eigenvalue of @xmath80 . if @xmath81 then we shall say that @xmath82 is a branch of eigenvalues along @xmath77 that starts as @xmath83 . if the underlying path @xmath77 is fixed then we shall skip referring to it . here , instead of considering @xmath0 , we consider branches of eigenvalues that start as @xmath0 and modify question [ ques ] as : for any @xmath47 does there exist branches of eigenvalues in @xmath84 that start as @xmath0 and exceeds @xmath37 eventually ? fortunately this modified question turns out to be much easier than the original one and we have a positive answer to it ( see theorem [ br ] ) . for any eigenvalue @xmath19 of @xmath13 , the dimension of @xmath85 is called the multiplicity of @xmath19 . if the multiplicity of @xmath0 were one for all closed hyperbolic surfaces of genus @xmath23 then theorem [ br ] would have showed the existence of surfaces with @xmath86 implying conjecture [ conj ] . however this is not the case and in fact the following is proved in @xcite : for every @xmath87 and @xmath88 there exists a surface @xmath89 such that @xmath35 is small and has multiplicity equal to the integral part of @xmath90 . for @xmath87 the above bound is more than @xmath91 . hence our methods in theorem [ e2 ] for @xmath92 do not work for @xmath87 . in @xcite the following upper bound on the multiplicity of a small eigenvalue is proved [ otal ] let @xmath13 be a finite area hyperbolic surface of type @xmath25 . then the multiplicity of a small eigenvalue of @xmath13 is at most @xmath93 . our last result is an improvement of this result for hyperbolic surfaces of type @xmath94 ( see theorem [ puncturedsphere ] ) . as mentioned before , it is known that there are closed hyperbolic surfaces of genus two with @xmath55 ( in fact with @xmath95 ) . our first result , in some sense , describes how large is the open subset @xmath96 [ e2 ] @xmath97 is an unbounded set that disconnects @xmath2 . * sketch of the proof of theorem [ e2 ] : * + we first prove that @xmath97 disconnects @xmath2 . we argue by contradiction and assume that @xmath98 is connected . now for any @xmath99 , @xmath35 is small and hence has multiplicity exactly one by @xcite . we shall see that , in fact , the _ nodal set _ of the @xmath35-eigenfunction ( see 3 ) consists of simple closed curves . with the help of this property we shall deduce that the nodal set of the first eigenfunction is constant , up to isotopy , on @xmath98 . finally , using an argument involving geodesic pinching we shall show that there exist surfaces @xmath100 and @xmath101 in @xmath102 such the nodal sets of the @xmath103-eigenfunction is not isotopic to the nodal set of the @xmath104-eigenfunction . this provides the desired contradiction . the rest of the theorem i.e. @xmath97 is unbounded is deduced from a description of the components of @xmath98 . for finite area hyperbolic surfaces with euler characteristic two the ideas in the above proof carries over to provide the following . [ e2 ] for any @xmath25 with @xmath105 ( i.e. @xmath106 or @xmath107 ) the set @xmath108disconnects @xmath109 . moreover for @xmath110 and @xmath111 it is unbounded . our next result is on the existence of branches of eigenvalues in @xmath61 , for any @xmath112 , that start as @xmath0 and eventually becomes larger than @xmath37 . [ br ] there are branches of eigenvalues in @xmath61 that start as @xmath0 and take values strictly bigger than @xmath37 . recall that @xmath113 can be embedded in @xmath61 as an analytic subset containing surfaces with certain symmetries ( see [ brnch ] ) . the branches in theorem [ br ] will be obtained by composing the branches in @xmath113 by the above embedding @xmath114 . we shall use a _ geodesic pinching _ argument to prove that among these branches there are ones that start as @xmath0 . our last result is on the multiplicity of @xmath0 of genus zero hyperbolic surfaces , punctured spheres . [ puncturedsphere ] let @xmath13 be a hyperbolic surface of genus @xmath115 . if @xmath116 is an eigenvalue then the multiplicity of @xmath35 is at most three . * sketch of proof : * let @xmath13 be a hyperbolic surface of genus @xmath115 with @xmath24 punctures . let @xmath22 denote the closed surface obtained by filling in the punctures of @xmath13 . assume that @xmath116 is an eigenvalue . let @xmath117 be a @xmath35-eigenfunction with nodal set @xmath118 ( 2 ) which is a finite graph by @xcite ( see lemma [ o ] ) . using * jordan curve theorem * and * courant s nodal domain theorem * we shall deduce the simple description of @xmath119 as a simple closed curve in @xmath22 . in particular , if one of the punctures @xmath120 of @xmath13 lies on @xmath119 then the number of arcs in @xmath119 emanating from @xmath120 is at most two . let @xmath120 be one of the punctures of @xmath13 . it is a standard fact that in any cusp around @xmath120 any @xmath35-eigenfunction @xmath117 has a fourier development of the form : @xmath121 where @xmath122 with @xmath123 $ ] and @xmath124 is the modified bessel function of exponential decay ( see 2 ) . denote the vector space generated by @xmath35-eigenfunctions by @xmath125 and consider the map @xmath126 given by @xmath127 . this is a linear map and so if @xmath128 then @xmath129 is non - empty . let @xmath130 i.e. @xmath131 . then by the result @xcite of judge , the number of arcs in @xmath132 emanating from @xmath120 is at least four , a contradiction to the above description of @xmath119 at @xmath120 . in this section we recall some definitions and results that will be necessary in later sections . let @xmath13 be a finite area hyperbolic surface . then @xmath13 is homeomorphic to a closed surface with finitely many points removed . each of these point , called punctures , has special neighborhoods in @xmath13 called _ cusp_s . denote by @xmath133 the parabolic isometry @xmath134 . for a choice of @xmath135 , a cusp @xmath136 is the half - infinite cylinder @xmath137 . the boundary curve @xmath138 is a _ horocycle _ of length @xmath78 . the hyperbolic metric on @xmath136 has the form : @xmath139 any function @xmath140 has a fourier development in the @xmath141 variable of the form @xmath142 if @xmath8 satisfy the equation @xmath143 then the above expression can be simplified as @xmath144 where @xmath145 is the modified _ bessel function _ ( see @xcite ) and @xmath146 the function @xmath8 is called _ cuspidal _ if @xmath147 . for any function @xmath148 , the set @xmath149 is called the _ nodal set _ @xmath150 of @xmath8 . each component of @xmath151 is called a _ nodal domain _ of @xmath8 . in a neighborhood of a regular point @xmath152 ( @xmath153 ) the implicit function theorem implies that @xmath150 is a smooth curve . in a neighborhood of a critical point @xmath152 ( @xmath154 ) , it is not so simple to describe @xmath150 . when @xmath8 is an eigenfunction of the laplacian we have the following description due to s. y. cheng @xcite : [ cheng ] let @xmath13 be a surface with a @xmath155 metric . then , for any solution of the equation @xmath156 , @xmath157 , one has : + @xmath158 critical points on the nodal set @xmath118 are isolated . + @xmath159 any critical point in @xmath118 has a neighborhood @xmath160 in @xmath13 which is diffeomorphic to the disc @xmath161 by a @xmath162-diffeomorphism that sends @xmath163 to an equiangular system of rays . [ rcheng ] in particular , if @xmath164 is a critical point of @xmath117 then the degree of the graph @xmath118 at @xmath120 is at least @xmath165 . hence if a component of @xmath118 is a simple closed loop then it is automatically smooth . when @xmath13 is closed theorem [ cheng ] implies that @xmath166 is a finite graph . when @xmath13 is non - compact with finite area it implies _ local _ finiteness of @xmath166 but not _ global_. in this particular case we have the following lemma due to jean - pierre otal ( * ? ? ? * lemma 6 ) ( the second part is ( * ? ? ? * lemma 1 ) ) [ o ] let @xmath13 be a hyperbolic surface with finite area and let @xmath167 be a @xmath19-eigenfunction with @xmath168 . then the closure of @xmath118 in @xmath22 is a finite graph . moreover , each nodal domain of @xmath117 has negative euler characteristic . in particular , @xmath169 is a union ( not necessarily disjoint ) of finitely many _ cycles _ in @xmath22 that may contain some of the punctures of @xmath13 . next we recall courant s nodal domain theorem [ courant ] let @xmath13 be a closed hyperbolic surface . then the number of nodal domains of a @xmath29-eigenfunction can be at most @xmath170 . the proof ( see @xcite or @xcite ) of this theorem works also for finite area hyperbolic surfaces if @xmath171 . in particular , for a hyperbolic surface @xmath13 with finite area if @xmath172 then the number of nodal domains of a @xmath35-eigenfunction is at most two . since any @xmath173-eigenfunction @xmath117 has mean zero , @xmath166 must disconnect @xmath13 . hence any @xmath0-eigenfunction has exactly two nodal domains . we begin by proving that @xmath97 disconnects @xmath2 . we argue by contradiction and assume that @xmath98 is connected . now , for any @xmath99 : @xmath174 and so @xmath35 is simple by @xcite . hence to a surface @xmath99 one can assign its first non - constant eigenfunction @xmath175 without any ambiguity . we assume that @xmath175 is normalized i.e. @xmath176 let @xmath177 denote the _ nodal set _ of @xmath175 . since @xmath175 is the first eigenfunction , by courant s nodal domain theorem , @xmath178 has exactly two components . denote by @xmath179 ( resp . @xmath180 ) the component of @xmath178 where @xmath175 is positive ( resp . negative ) . by euler - poicar formula applied to the cell decomposition of @xmath13 consisting of nodal domains of @xmath175 as the two skeleton and the nodal set @xmath177 as the one skeleton we have the following equality : @xmath181 since @xmath182 and both @xmath183 and @xmath184 are negative by lemma [ o ] , we conclude from that @xmath185 . this means that @xmath177 consists of simple closed curve(s ) that divide @xmath13 into exactly two components . moreover , since no nodal domain of @xmath186 is a disc or an annulus by lemma [ o ] , each curve in @xmath177 is _ essential _ ( homotopically non - trivial in @xmath13 ) and no two curves in @xmath177 are homotopic . in particular , [ description ] for any @xmath99 , the nodal set @xmath177 of @xmath175 consists either of tree smooth simple closed curves that divide @xmath13 into two pair of pants ( the first picture below ) or of a unique smooth simple closed curve that divides @xmath13 into two tori with one hole ( the second picture below ) . now we have the following : [ isotopy ] let @xmath187 such that @xmath35 is simple and the nodal set @xmath177 of the @xmath35-eigenfunction @xmath175 is also simple . then @xmath13 has a neighborhood @xmath188 in @xmath2 such that for any @xmath189 the nodal set @xmath190 is isotopic to @xmath177 . first observe that @xmath35 being simple we have a neighborhood @xmath191 in @xmath2 such that for any @xmath192 @xmath193 is simple . hence @xmath194 is well defined too . now @xmath175 is the @xmath35-eigenfunction , so @xmath195 has exactly two connected components @xmath196 and @xmath197 such that @xmath175 has positive sign on @xmath196 . so necessarily @xmath175 has negative sign on @xmath197 . now consider a tubular neighborhood @xmath198 of @xmath177 . by * theorem 3.36)(see also @xcite , @xcite ) we have a neighborhood @xmath199 of @xmath13 such that for any @xmath192 , @xmath200 has positive sign on @xmath201 and negative sign on @xmath202 . in particular , @xmath203 . hence by the description of @xmath204 as in claim [ description ] the proof follows . therefore , there exists @xmath205 such that @xmath177 consists of only one curve if and only if for all @xmath206 , @xmath190 consists of only one curve . this is a contradiction to proposition [ diffisotopy ] . the _ systole @xmath207 _ of a surface @xmath13 is the minimum of the lengths of closed geodesics on @xmath13 . injectivity radius _ of @xmath13 at a point @xmath120 is the maximum of the radius of the geodesic discs with center @xmath120 that embed in @xmath13 . for any @xmath49 the set of points of @xmath13 with injectivity radius at least @xmath208 is denoted by @xmath209 . each point in the complement @xmath210 has injectivity radius at most @xmath208 . @xmath209 and @xmath211 are respectively called _ @xmath208-thick part _ and _ @xmath208-thin part _ of @xmath13 . [ diffisotopy ] let @xmath13 be a finite area hyperbolic surface of type @xmath25 . let @xmath212 be a collection of smooth , mutually non - intersection simple closed curves on @xmath13 that separates @xmath13 in exactly two components . assume that @xmath213 is minimal in the sense that no proper subset of @xmath213 can separate @xmath13 . then given any @xmath214 there exists a finite area hyperbolic surface @xmath215 of type @xmath25 with @xmath216 such that @xmath217 is simple and the nodal set of the @xmath218-eigenfunction is isotopic to @xmath213 . it is not very difficult to construct two collections of curves on @xmath13 , as in the above lemma , that are not isotopic . in particular for @xmath219 claim [ description ] provides two such collections . therefore the above lemma indeed provide two surfaces @xmath100 and @xmath101 in @xmath58 such that @xmath100 , @xmath220 and @xmath221 is not isotopic to @xmath222 . proof of proposition [ diffisotopy ] uses the behavior of sequences of small eigenpairs over degenerating sequences of hyperbolic surfaces . for precise definitions of these concepts we refer the reader to @xcite . without loss of generality we may assume that each curve in @xmath213 is a geodesic . extend @xmath213 to a pants decomposition @xmath223 of @xmath13 . let @xmath224 denote the fenchel - nielsen coordinates on @xmath225 with respect to @xmath226 . here @xmath227 denotes the length parameter and @xmath228 denotes the twist parameter along @xmath229 . now consider the sequence of surfaces @xmath230 in @xmath231 such that @xmath232 for @xmath233 , @xmath234 for @xmath235 and @xmath236 for @xmath237 . then , up to extracting a subsequence , @xmath230 converges to a finite area hyperbolic surface @xmath238 . let us denote the extracted subsequence by @xmath239 it self . observe that @xmath240 is obtained from @xmath13 by pinching the geodesics in @xmath213 . namely , for each @xmath241 there is a geodesic @xmath242 in @xmath243 , in the homotopy class of @xmath229 , whose length tends to zero as @xmath244 . the number of components of @xmath245 is exactly two . hence by @xcite , @xmath246 and all other eigenvalues of @xmath243 stay away from zero . in particular @xmath247 is simple for @xmath248 sufficiently large . let @xmath249 be the @xmath250-eigenfunction with @xmath251-norm @xmath252 . recall that we want to prove that for any @xmath253 there exists a @xmath215 with @xmath216 such that @xmath254 is simple and the nodal set of the @xmath218-eigenfunction is isotopic to @xmath213 . since @xmath255 by construction and @xmath246 by above it suffices to prove that @xmath256 is isotopic to @xmath213 for sufficiently large @xmath248 . now we apply ( * ? ? ? * theorem 3.34 ) to extract a subsequence of @xmath249 that converges uniformly over compacta to a @xmath115-eigenfunction @xmath257 of @xmath240 with @xmath251-norm @xmath252 . let us denote the extracted subsequence by @xmath239 itself . since @xmath115-eigenfunctions are constant functions , @xmath257 is constant on each components of @xmath240 . [ sign ] the two constant values of @xmath257 on the two components of @xmath240 are non - zero and have opposite sign . for @xmath49 let us denote the @xmath251-norm of @xmath249 restricted to @xmath258 by @xmath259 . by the uniform convergence of @xmath249 to @xmath257 over compacta we have @xmath260 since @xmath261 we obtain that for any @xmath262 there exists @xmath49 such that @xmath263 . now @xmath264 here @xmath265 denotes the area of @xmath258 . recall that , for any @xmath266 , @xmath267 . so for @xmath268 and @xmath208 sufficiently small : @xmath269 finally , taking @xmath208 to be sufficiently small , we calculate : @xmath270 since @xmath271 since @xmath272 is arbitrary we conclude that @xmath273 . hence @xmath274 has @xmath251-norm @xmath252 and mean zero . since @xmath257 has @xmath251-norm @xmath252 at least one of the two constant values of @xmath257 on the two components of @xmath240 is non - zero . since @xmath257 has mean zero both of these values are non - zero have opposite sign . as the length of @xmath242 tends to zero , we may assume that the collar neighborhood @xmath275 of @xmath242 with two boundary components of length @xmath252 embeds in @xmath243 and @xmath276 are mutually disjoint . at this point we recall that @xmath213 is minimal in the sense that no proper subset of @xmath213 can separate @xmath13 . hence not only @xmath277 separates @xmath13 in exactly two components but also no proper sub - collection of @xmath276 can separate @xmath243 . in particular , for each @xmath30 , the limits of the two components of @xmath278 belong to two different components of @xmath240 . using claim [ sign ] let us denote the limits of these two boundary sets by @xmath279 and @xmath280 such that @xmath281 and @xmath282 . correspondingly denote the two components of @xmath278 by @xmath283 and @xmath284 such that @xmath285 is the limit of @xmath286 respectively . by the uniform convergence of @xmath249 to @xmath257 over compacta we conclude that , for sufficiently large @xmath248 , @xmath287 and @xmath288 . hence , for @xmath248 sufficiently large , at least one component of @xmath289 is contained in @xmath275 . let @xmath290 denote the union of the components of @xmath289 that are contained in @xmath275 . let @xmath77 be a simple closed loop in @xmath290 . since @xmath291 is @xmath292 there are only two possibilities for @xmath77 . either it bounds a disc in @xmath275 or it is homotopic to @xmath242 . since @xmath250 is small , each component of @xmath293 has negative euler characteristic by lemma [ o ] . this discards the possibility that @xmath77 bounds a disc in @xmath275 . hence @xmath77 is homotopic to @xmath242 . let @xmath294 be another simple closed loop in @xmath290 . then @xmath294 is also homotopic to @xmath242 implying that one of the components of @xmath293 has non - negative euler characteristic . this leaves us with the observation that each @xmath275 contains exactly one loop @xmath295 from @xmath289 . by remark [ rcheng ] @xmath295 is in fact smooth . therefore we have an isotopy of @xmath13 that sends @xmath295 to @xmath242 . combining these isotopies we obtain that @xmath289 is isotopic to @xmath296 . it remains to show that @xmath97 is unbounded . we argue by contradiction and assume that @xmath97 is bounded . then we have @xmath49 such that @xmath97 is contained in the compact set @xmath297 @xcite . now applying lemma [ diffisotopy ] obtain @xmath100 and @xmath101 in @xmath2 such that @xmath298 , @xmath299 is simple and the nodal set of the @xmath300-eigenfunction is not isotopic to @xmath301-eigenfunction . since @xmath302 is path connected ( see lemma [ pconn ] ) we may have a path @xmath294 in @xmath302 that joins @xmath100 and @xmath101 . then lemma [ isotopy ] implies that the nodal set of the @xmath300-eigenfunction is isotopic to @xmath301-eigenfunction . this is a contradiction to our choice of @xmath100 and @xmath101 . the case @xmath219 follows from the above theorem . it remains to show theorem [ e2 ] for @xmath303 and @xmath107 . * for the rest of the proof we refer to the pair @xmath25 for only these two cases . * we argue by contradiction and assume that @xmath304 is connected . by definition @xmath305 for any @xmath306 . hence @xmath307 is an eigenvalue and by @xcite it is the only non - zero small eigenvalue of @xmath13 . so we can consider the first non - constant eigenfunction @xmath175 of @xmath13 . as before let @xmath308 be the nodal set of @xmath175 . denote by @xmath22 the surface obtained from @xmath13 by filling in its punctures and by @xmath309 the closure of @xmath308 in @xmath22 . by lemma [ o ] @xmath309 is a finite graph . now apply euler - poincar formula to the cell decomposition of @xmath22 defined as follows : the punctures on @xmath13 that do not lie on @xmath309 is the zero skeleton , @xmath309 is the one skeleton and @xmath310 is the two skeleton . if @xmath311 is the number of punctures of @xmath13 that do not lie on @xmath312 then @xmath313 by lemma [ o ] each component of @xmath310 has negative euler characteristic and so @xmath314 . for @xmath315 , @xmath316 and so we have the only possibility @xmath317 and @xmath318 . for @xmath319 , @xmath320 leaving us with the only possibility @xmath321 and @xmath318 . hence none of the punctures of @xmath13 lie on the closure of the nodal set @xmath312 i.e. @xmath322 is a compact subset of @xmath13 . since @xmath323 we conclude that @xmath324 is a union of simple closed curves . also by lemma [ o ] we know that no loop in @xmath324 can bound a disc and no two components of @xmath324 can be homotopic . summarizing these observations we get : [ ne2 ] let @xmath306 . + @xmath158 if @xmath315 then @xmath308 consists of either exactly one simple closed curve or two simple closed curves . in the first case @xmath308 divides @xmath13 into two components one of which is a surface of genus one with a copy of @xmath308 as its boundary and the other one is a twice punctured sphere with a copy of @xmath308 as its boundary . in the last case @xmath308 divides @xmath13 into two components each of which is a once punctured sphere with two boundary components coming from @xmath308 . + @xmath159 if @xmath325 then @xmath308 consists of exactly one simple closed curve ( there are two possibilities for this up to isotopy ) that separates @xmath13 into two components each of which is a twice punctured sphere with one boundary component coming from @xmath308 . next we have the following modified version of claim [ isotopy ] . let @xmath306 with @xmath175 the @xmath307-eigenfunction . [ isotopy1 ] there exists a neighborhood @xmath188 of @xmath13 in @xmath326 such that for any @xmath192 : @xmath327 is simple and the nodal set @xmath204 of the @xmath327-eigenfunction @xmath200 is isotopic to @xmath308 . since @xmath307 is @xmath328 , @xmath0 is a continuous function in a neighborhood of @xmath13 by @xcite(see also @xcite , @xcite ) and so we have a neighborhood @xmath191 of @xmath13 which is contained in @xmath304 . in particular , @xmath200 is well - defined for @xmath329 and @xmath204 has the description in claim [ ne2 ] . now consider a tubular neighborhood @xmath198 of @xmath308 in @xmath13 such that @xmath330 has two components @xmath331 and @xmath332 each of which is a simple closed curve with @xmath333 and @xmath334 . now @xmath335 and simple on @xmath191 . hence by @xcite for any compact subset @xmath124 of @xmath13 the map : @xmath336 given by @xmath337 is continuous . considering @xmath338 we obtain @xmath199 such that for any @xmath192 : @xmath339 and @xmath340 . in particular , for any @xmath192 : @xmath204 has a component inside @xmath198 . hence by the description of @xmath204 in claim [ ne2 ] we obtain the claim . since by our assumption @xmath304 is connected the above claim implies that only one of the two possibilities in claim [ ne2 ] can actually occur . this is a contradiction to proposition [ diffisotopy ] . now we show that @xmath341 is unbounded . we argue by contradiction and assume that @xmath341 is bounded . then we have @xmath49 such that @xmath341 is contained in the compact set @xmath342 @xcite . applying lemma [ diffisotopy ] we obtain @xmath100 and @xmath101 in @xmath343 such that @xmath298 , @xmath299 is simple and the nodal set of the @xmath300-eigenfunction is not isotopic to @xmath301-eigenfunction . since @xmath344 is path connected ( see lemma [ pconn ] ) we may have a path @xmath294 in @xmath344 that joins @xmath100 and @xmath101 . then lemma [ isotopy1 ] implies that the nodal set of the @xmath300-eigenfunction is isotopic to @xmath301-eigenfunction . this is a contradiction to our choice of @xmath100 and @xmath101 . in this section we consider branches of eigenvalues along paths in @xmath61 . main purpose of doing so is that the multiplicity of @xmath83 , in particular @xmath0 is not one in general . therefore along nice paths in @xmath61 the functions @xmath83 may not be nice enough ( see introduction ) . however , theorem [ b - c ] shows that up to certain choice at points of multiplicity @xmath83 s are in fact nice. this nice choice makes @xmath83 into a branch of eigenvalues . theorem [ br ] says that if we restrict ourselves to branches of eigenvalues then we have a positive answer to conjecture [ conj ] , namely there are branches of eigenvalues that start as @xmath0 and becomes more than @xmath37 . we begin by explaining the the embedding @xmath345 ( see the next figure ) . let @xmath13 be the closed hyperbolic surface of genus two and @xmath346 , @xmath272 are four geodesics on @xmath13 as in the following picture . now cut @xmath13 along @xmath272 to obtain a hyperbolic surface @xmath347 with genus one and two geodesic boundaries ( each a copy of @xmath272 ) . consider @xmath348 many copies of @xmath347 and glue them along their consecutive boundaries after arranging them along a circle as in the picture below . let @xmath349 denote the resulting hyperbolic surface . now take a geodesic pants decomposition @xmath350 of @xmath13 involving @xmath351 and consider the fenchel - nielsen coordinates @xmath352 on @xmath113 with respect to this pants decomposition . here @xmath353 is the length of the closed geodesic @xmath354 and @xmath355 is the twist parameter at @xmath354 . the images of @xmath350 in @xmath349 , @xmath356 is a geodesic pants decomposition of @xmath349 . consider the the fenchel - nielsen coordinates @xmath357 on @xmath61 with respect to this pants decomposition . as before , @xmath358 is the length of the closed geodesic @xmath359 and @xmath360 is the twist parameter at @xmath359 . with respect to these pants decompositions @xmath361 is expressed as @xmath362 this is an analytic map and the image @xmath349 of any @xmath363 has an isometry @xmath364 of order @xmath365 that sends one @xmath366-tuple @xmath367 to the next one . also @xmath368 is isometric to @xmath13 i.e. @xmath349 is a @xmath365 sheeted covering of @xmath13 . hence each eigenvalue of @xmath13 is also an eigenvalue of @xmath349 . in particular , a branch @xmath79 of eigenvalues in @xmath369 along @xmath370 is a a branch of eigenvalues in @xmath84 along @xmath371 . to finish the proof we need only to find @xmath363 such that @xmath372 . once we find such a @xmath13 , we can consider any analytic path @xmath373 in @xmath369 such that @xmath374 and @xmath375 . then the branch of eigenvalues @xmath376 along @xmath371 would be a branch that we seek . to show this we employ the technique in claim [ diffisotopy ] . let @xmath377 be a sequence of surfaces of genus two on which the lengths of the geodesics @xmath378 and @xmath379 tends to zero . in particular , @xmath380 implying @xmath381 and @xmath382 . the sequence @xmath383 converges to a surface in @xmath384 and so @xmath385 and @xmath386 . so for large @xmath24 , @xmath387 implying @xmath388 . we begin this section by recapitulating the ideas in @xcite . by purely number theoretic methods atle selberg showed that for any congruence subgroup @xmath10 of sl@xmath40 , @xmath389 . the purpose in @xcite was to construct explicit closed hyperbolic surfaces with @xmath0 close to @xmath390 . to achieve this goal the authors of @xcite considered principal congruence subgroups @xmath42 ( see introduction ) and corresponding finite area hyperbolic surfaces @xmath391 . then they replaced the cusps in @xmath391 , which is even in number , by closed geodesics of small length @xmath78 and glued them in pairs ( see @xcite for details ) . the surface @xmath392 obtained in this way is closed , their genus @xmath23 is independent of @xmath78 and as @xmath393 , @xmath394 in the compactification of the moduli space @xmath60 . rest of the proof showed that @xmath0 is lower semi - continuous over the family @xmath392 . this approach together with the result of kim and sarnak provides theorem [ aprox ] . limiting properties of eigenvalues over degenerating family of hyperbolic metrics have been studied well in the literature ( to name a few denis hejhal @xcite , gilles courtois - bruno colbois @xcite , lizhen ji @xcite , scott wolpert @xcite , chris judge @xcite ) ( see also ( * ? ? ? * theorem 2 ) ) . these limiting results can be summarized as : let @xmath239 be a sequence of hyperbolic surfaces in @xmath326 that converges to a finite area hyperbolic surface @xmath395 . let @xmath396 be an eigenpair of @xmath243 such that @xmath397 . then , up to extracting a subsequence and up to rescalling , the sequence @xmath398 converges to a generalized eigenfunction over compacta if one of the following is true + @xmath158 @xmath399 ( @xcite ) @xmath159 @xmath400 and @xmath401 ( @xcite , @xcite ) @xmath402 @xmath400 and @xmath403 ( @xcite ) @xmath402 @xmath400 , @xmath404 and @xmath405 is cuspidal ( @xcite ) . recall that there is a copy of @xmath406 in the compactification @xmath407 of @xmath109 . the ideas in @xcite along with above limiting results imply [ bbd ] for any pair @xmath25 , @xmath408 motivated by this we focus on @xmath409 . although we would not be able to prove conjecture [ conj ] we have theorem [ puncturedsphere ] on the multiplicity of @xmath0 which we prove now . let @xmath13 be a hyperbolic surface of genus @xmath115 and assume that @xmath116 is an eigenvalue . let @xmath117 be a @xmath35-eigenfunction . then the closure @xmath119 of the nodal set @xmath118 of @xmath117 is a finite graph in @xmath22 by theorem [ otal ] . in particular , @xmath119 is a union of closed loops in @xmath22 . observe also that the number of components of @xmath410 is same as that of @xmath411 . now let @xmath119 consists of more than one closed loop . then by jordan curve theorem the number of components of @xmath410 is at least three . this is a contradiction to courant s nodal domain theorem [ courant ] which says that a @xmath35-eigenfunction can have at most two nodal domains . hence we conclude that @xmath119 is a simple closed curve in @xmath22 . in particular , we have the following description of @xmath118 at any puncture . [ arcs ] if one of the punctures @xmath120 of @xmath13 is a vertex of @xmath119 then the number of arcs in @xmath119 emanating from @xmath120 is at most two . let @xmath122 with @xmath123 $ ] . let @xmath120 be one of the punctures of @xmath13 . let @xmath136 be a cusp around @xmath120 ( see 1 ) . recall that @xmath13 being a punctured sphere , does not have any cuspidal eigenvalue @xcite , @xcite . thus any @xmath35-eigenfunction @xmath117 is a linear combination of residues of eisenstein series ( see @xcite ) . it follows from ( * ? ? ? * thorem 6.9 ) that the @xmath412 term can not occur in the fourier development ( see ) of these residues in @xmath136 . hence @xmath117 has a fourier development in @xmath136 of the form ( see 1 ) : @xmath413 now we consider the space @xmath125 generated by @xmath35-eigenfunctions . the map @xmath126 given by @xmath127 is linear and so if @xmath414 then @xmath129 is non - empty . let @xmath130 i.e. @xmath415 . then by the result @xcite of judge , the number of arcs in @xmath416 emanating from @xmath120 is at least four , a contradiction to claim [ arcs ] . i would like to thank my advisor jean - pierre otal for all his help starting from suggesting the problem to me . i am thankful to peter buser and werner ballmann for the discussions that i had with them on this problem . i would like to thank the max planck institute for mathematics in bonn for its support and hospitality . let @xmath100 and @xmath101 be two surfaces in @xmath109 such that @xmath421 . so we have simple closed geodesics @xmath422 on @xmath100 and @xmath423 on @xmath101 such that the length @xmath424 of @xmath229 is @xmath425 . recall that it has always been our practise to treat @xmath326 as a subset of all possible metrics on a fixed surface @xmath13 and the geodesics are understood to be parametric curves on @xmath13 that satisfy certain differential equations provided by the metric . with this understanding let us first assume that @xmath422 does not intersect @xmath423 . so we may consider a pants decomposition @xmath426 of @xmath13 containing both @xmath422 and @xmath423 . let the fenchel - nielsen coordinates of @xmath427 be given by @xmath428 . here @xmath429 , @xmath430 are the length parameters along @xmath431 and @xmath432 are twist parameters along @xmath431 . then consider the path @xmath433 \to \mathcal{t}_2 $ ] given by : @xmath434 $ } , \\ 2(1-t){l_1}(s_1 ) + ( 2t-1){l_1}(s_2 ) & \text{if $ t \in [ \frac{1}{2 } , 1]$ } \end{cases}\ ] ] @xmath435 $ } , \\ { l_2}(s_2 ) & \text{if $ t \in [ \frac{1}{2 } , 1]$ } \end{cases}\ ] ] @xmath436 and @xmath437 . since @xmath438 for @xmath439 $ ] and @xmath440 for @xmath441 $ ] we observe that @xmath442 for all @xmath78 . the image of @xmath294 under the quotient map @xmath443 produces the required path joining @xmath100 and @xmath101 . now let us assume that @xmath422 intersects @xmath423 . let @xmath379 be a simple closed geodesic that does not intersect @xmath422 and @xmath423 . by our assumption i.e. @xmath417 such a geodesic exists . then by the procedure described above both @xmath100 and @xmath101 can be joined by a path in @xmath418 to a surface on which @xmath379 has length @xmath425 . this finishes the proof . bers , l. ; a remark on mumford s compactness theorem , israel j. math . 12 ( 1972 ) , 400 - 407 . buser , peter ; cubic graphs and the first eigenvalue of a riemann surface , math . z. 162 ( 1978 ) , 87 - 99 . buser , peter ; on the bipartition of graphs , discrete applied mathematics , 9 ( 1984 ) , 105 - 109 . buser , peter ; geometry and spectra of compact riemann surfaces . progress in mathematics , 106 . birkhuser boston , inc . , boston , ma , 1992 . burger , m. , buser , p. , dodziuk , j. ; riemann surfaces of large genus and large @xmath0 . geometry and analysis on manifolds ( t. sunada , ed . ) , lecture notes in math . 1339 , springer - verlag , berlin , 1988 , 54 - 63 . brooks , r. , makover e. , riemann surfaces with large first eigenvalue . 83 ( 2001 ) , 243 - 258 . chavel , isaac ; eigenvalues in riemannian geometry . pure and applied mathematics , 115 . academic press , 1984 . cheng , s. y. ; eigenfunctions and nodal sets , comment . helvetici 51 ( 1976 ) , 43 - 55 . colbois , b. , courtois , g. , les valeurs propres infrieures 1/4 des surfaces de riemann de petit rayon dinjectivit . ( 1989 ) , no . 3 , 349- 362 . colbois , b. ; colin de verdire , y. ; sur la multiplicit de la premire valeur propre dune surface de riemann courbure constante . ( french ) [ multiplicity of the first eigenvalue of a riemann surface with constant curvature ] comment . 63 ( 1988 ) , no . 2 , 194208 . hejhal , d. ; regular b - groups , degenerating riemann surfaces and spectral theory , memoires of amer . math . 88 , no . 437 , 1990 . huxley , m. n. ; cheeger s inequality with a boundary term , commentarii mathematici helvetici 58 ( 1983 ) . iwaniec , h. , introduction to the spectral theory of automorphic forms , bibl . iberoamericana , revista matemtica iberoamericana , madrid , 1995 . jenni , f. ; uber den ersten eigenwert des laplace - operators auf ausgewhlten beispielen kompakter riemannscher flchen . ( german ) [ on the first eigenvalue of the laplace operator on selected examples of compact riemann surfaces ] comment . 59 ( 1984 ) , no . 2 , 193 - 203 . ji , lizhen ; spectral degeneration of hyperbolic riemann surfaces . j. differential geom . 38 ( 1993 ) , no . 2 , 263 - 313 . judge , chris ; the nodal set of a finite sum of maass cusp forms is a graph . proceedings of symposia in pure mathematics , volume 84 , 2012 kim , henry h. ; functoriality for the exterior square of @xmath444 and symmetric fourth of @xmath445 . ( 2003 ) , no . 1 , 139 - 183 . mondal , sugata ; topological bounds on the number of cuspidal eigenvalues of finite area hyperbolic surfaces ( preprint ) otal , jean - pierre ; three topological properties of small eigenfunctions on hyperbolic surfaces . geometry and dynamics of groups and spaces , progr . 265 , birkhuser , bassel , 2008 . otal , jean - pierre ; rosas , eulalio ; pour toute surface hyperbolique de genre g , @xmath446 . duke math . j. 150 ( 2009 ) , no . 1 , 101 - 115 . selberg , a. ; on the estimation of fourier coefficients of modular forms . pure math . vii , amer . ( 1965 ) , 1 - 15 . strohmaier , a. , uski , v. ; an algorithm for the computation of eigenvalues , spectral zeta functions and zeta - determinants on hyperbolic surfaces . 317 , ( 2013 ) , no . 3 , 827869 . wolpert , s. a. ; spectral limits for hyperbolic surface , i , invent . ( 1992 ) , 67 - 89 .	we apply topological methods to study the smallest non - zero number @xmath0 in the spectrum of the laplacian on finite area hyperbolic surfaces . for closed hyperbolic surfaces of genus two we show that the set @xmath1 is unbounded and disconnects the moduli space @xmath2 .
the notion of a recognizable language is a fundamental concept in formal language theory , which has been clearly identified since the 1950 s . it is important because of its numerous applications , in particular for the construction of compilers , and also for the development of the theory : indeed , these languages can be specified in several very different ways , by means of _ automata _ , _ congruences _ , _ regular expressions _ and _ logical formulas_. this multiplicity of quite different definitions is a clear indication that the notion is central since one arrives at it in a natural way from different approaches . the equivalence of definitions is proved in fundamental results by kleene , myhill and nerode , elgot and bchi . the notion of a recognizable set has been extended in the 1960 s to trees ( actually to trees representing finite algebraic terms ) , to infinite words and to infinite trees . in the present article we discuss its extension to sets of finite graphs . the recognizability of a set of finite words or trees can be defined in several ways , as mentioned above , and in particular by finite _ deterministic _ automata . this definition ( together with the related effective translations from other definitions ) provides linear - time recognition algorithms , which are essential for compiler construction , coding , text processing , and in other situations . recognizable sets of words can also be defined in an algebraic way by finite saturating congruences relative to the monoid structure . these definitions , by automata and congruences , extend smoothly to the case of finite trees ( i.e. , algebraic terms ) , using the natural algebra structure . the notion of recognizability in a general algebra is due to mezei and wright @xcite . we will not discuss here the extensions to infinite words and trees , which raise specific problems surveyed by thomas @xcite and perrin and pin @xcite . our aim will be to consider sets of finite graphs . for finite graphs , there is no automaton model , except in very special cases , and in particular in the case of graphs representing certain _ labelled partially ordered sets _ and _ traces _ ( a trace is a directed acyclic graph , representing the equivalence class of a word w.r.t . a partial commutation relation ) , see the volume edited by diekert @xcite and the papers by lodaya and weil @xcite and sik and nmeth @xcite . algebraic definitions via finite congruences can be given because the set of finite graphs can be equipped with an algebraic structure , based on graph operations like the concatenation of words . however , many operations on graphs can be defined , and there is no prominent choice for a standard algebraic structure like in the case of words where a unique associative binary operation is sufficient . several algebraic structures on graphs can be defined , and distinct notions of recognizability follow from these possible choices . it appears nevertheless that two graph algebras , called the _ @xmath0-algebra _ and the _ @xmath1-algebra _ for reasons explained below , emerge and provide robust notions of recognizability . the main purpose of this paper is to demonstrate the robustness of these notions . by robustness , we mean that taking variants of the basic definitions does not modify the corresponding classes of recognizable sets of graphs . in any algebra , one can define two family of sets , the recognizable sets and the equational sets . the equational sets are defined as the components of the least solutions of certain systems of recursive set equations , written with set union and the operations of the algebra , extended to sets in the standard way . equational sets can be considered as the natural extension of context - free languages in a general algebraic framework ( mezei and wright @xcite , courcelle @xcite for a thorough development ) . the two graph algebras introduced above , the @xmath0- and the @xmath1__-algebra _ _ , are familiar to readers interested in graph grammars , because their equational sets are the ( context - free ) _ hyperedge replacement _ ( @xmath0 ) sets of graphs on the one hand , and the ( context - free ) _ vertex replacement _ ( @xmath1 ) sets on the other . both classes of context - free sets of graphs can be defined in alternative , more complicated ways in terms of graph rewritings , and are robust in the sense that they are closed under certain transformations expressible in monadic second - order logic ( courcelle @xcite ) . the main results of this paper , described below in more detail , are : \1 ) the robustness of the classes of @xmath1- and @xmath0-recognizable sets of graphs , \2 ) the robustness of the class of recognizable sets of finite relational structures ( equivalently of simple directed ranked hypergraphs ) , which extends the two previous classes , \3 ) the exhibition of structural conditions on sets of graphs implying that @xmath0-recognizability and @xmath1-recognizability coincide , \4 ) the comparison of the recognizable sets of the @xmath1-algebra and those of a closely related algebra representing _ modular decompositions _ ( modular decomposition is another useful notion for graph algorithms ) . the notion of recognizability of a set of finite graphs is important for several reasons . first , because recognizability yields linear - time algorithms for the verification of a wide class of graph properties on graphs belonging to certain finitely generated graph algebras . these classes consist of graphs of bounded tree - width and of bounded clique - width . these two notions of graph complexity are important for constructions of polynomial graph algorithms , see downey and fellows @xcite and courcelle et al . furthermore , these graph properties are not very difficult to identify because monadic second - order ( ms ) logic can specify them in a formalized and uniform way . ( in many cases , an ms formula can be obtained from the graph theoretical expression of a property ) . more precisely , a central result @xcite says that every set of graphs ( or graph property ) definable by an ms formula is recognizable ( respectively admits such algorithms ) , for appropriate graph algebras . this general statement covers actually several distinct situations . another reason comes from the theory of graph grammars . the intersection of a context - free set of graphs and of a recognizable set is context - free ( in the appropriate algebraic framework ) . this gives immediately many closure properties for context - free sets of graphs , via the use of ms logic as a specification language for graph properties . recognizability also makes it possible to construct terminating and ( in a certain sense ) confluent graph rewriting rules by which one can recognize sets of graphs of bounded tree - width by graph reduction in linear time , see arnborg _ _ @xcite . finally , recognizability is a basic notion for dealing with languages and sets of terms , and on this ground , its extension to sets of graphs is worth investigating . logical characterizations of recognizability can be given using ms logic , extending many results in language theory @xcite . several questions remain open in this research field . we have noted above that defining recognizability for sets of graphs can not be done in terms of finite automata , so that the algebraic definition in terms of finite congruences has no alternative . another advantage of the algebraic definition is that it is given at the level of universal algebra ( mezei and wright @xcite ) , and thus applies to objects other than graphs . however , even in the case of graphs , the algebraic setting is useful because it hides ( temporarily ) the complexities of operations on graphs and makes it possible to understand what is going on at a structural level . we now present the main results of this article more in detail . the two main algebraic structures on graphs called @xmath1 and @xmath0 , originate from algebraic descriptions of context - free graph grammars . definitions will be given in the body of the text . it is enough for this introduction to retain that the operations of @xmath1 are more powerful than those of @xmath0 . hence every @xmath0-context - free set of graphs ( i.e. , defined by a grammar based on the operations of @xmath0 ) is @xmath1-context - free , but not vice - versa . for recognizability , the inclusion goes in the opposite direction : every @xmath1-recognizable set is @xmath0-recognizable but the converse is not true . however , if the graphs of a set @xmath2 have no subgraph of the form @xmath3 ( the complete bipartite graph on @xmath4 vertices ) for some @xmath5 , then @xmath2 is @xmath0-recognizable if and only if it is @xmath1-recognizable ( this is the main theorem of section [ sec knn ] ) . a similar statement is known to hold under the same hypothesis for context - free sets : if @xmath2 is without @xmath3 ( _ i.e. _ , no graph in @xmath2 contains a subgraph isomorphic to @xmath3 ) , then it is @xmath0-context - free if and only if it is @xmath1-context - free ( courcelle , @xcite ) . the proofs of the two statements are however different ( and both difficult ) . up to now we have only discussed graphs , but our approach , which extends the approach developped by courcelle in @xcite , also works for hypergraphs and for relational structures . the operations on graphs , hypergraphs and structures are basically of three types defined in section [ sec : algrelstr ] : we use only one binary operation , the disjoint union ; we use unary operations defined by quantifier - free first - order formulas ; and basic graphs and structures corresponding to nullary operations . in this way we can generate graphs and structures by finite algebraic terms . the quantifier - free definable operations can modify vertex and edge labels , add or delete edges . this notion is thus quite flexible . what is remarkable is that these numerous operations can be added without altering the notion of recognizability . the main result of section [ sec gp ] states that the same recognizable sets of graphs are obtained if one uses the basic @xmath1-algebra ( closely connected to the definition of _ clique - width _ ) , the same algebra enriched with quantifier - free definable operations , and even the larger algebra dealing with relational structures . variants of the @xmath1-algebra which are useful , in particular for algorithmic applications , are also considered , and they are proved to yield the same class of recognizable sets . in section [ sec gs ] , we discuss similarly the @xmath0-algebra which is very important because of its relation with _ tree - width _ and with context - free graph grammars . we prove a robustness result relative to the subclass such that the distinguished vertices denoted by distinct labels ( nullary operations ) are different . the @xmath0-operations are appropriate to handle graphs and hypergraphs with multiple edges and hyperedges ( whereas the @xmath1-operations are not ) . the original definitions ( see courcelle @xcite ) were given for graphs with multiple edges and hyperedges . in section [ sec simple multiple ] , we prove that for a set of simple graphs , @xmath0-recognizability is the same in the @xmath0-algebra of simple graphs and in the larger @xmath0-algebra of graphs with multiple edges . without being extremely difficult , the proof is not just a routine verification . in section [ sec : modular ] , we consider an algebra arising from the theory of modular decomposition of graphs . we show that under a natural finiteness condition , the corresponding class of recognizable sets is equal to that of @xmath1-recognizable ones . in an appendix , we clarify the definitions of certain equivalences of logical formulas , focusing on cases where they are decidable , and we give upper bounds to the cardinalities of the quotient sets for these equivalences . these results yield upper bounds to the number of equivalence classes in logically based congruences . they are thus useful for the investigation of recognizability in view of the cases where the sets under consideration are defined by logical formulas . they also provide elements to appreciate ( an upper bound of ) the complexity of the algorithms underlying a number of the effective proofs in the main body of the paper . this work has been presented in invited lectures by b. courcelle @xcite and p. weil @xcite . the notion of a recognizable set is due to mezei and wright @xcite . it was originally defined for one - sort structures , and we adapt it to many - sorted ones with infinitely many sorts . we begin with definitions concerning many - sorted algebras . we follow essentially the notation and definitions from @xcite , see also @xcite . let @xmath6 be a set called the set of _ sorts_. an @xmath6-_signature _ is a set @xmath7 given with two mappings @xmath8 ( the set of finite sequences of elements of @xmath6 ) , called the _ arity _ mapping , and @xmath9 , called the _ sort _ mapping . we denote by @xmath10 the length of the sequence @xmath11 , which we call also _ arity_. the _ type _ of @xmath12 in @xmath7 is the pair @xmath13 that we shall rather write @xmath14 , or @xmath15 if @xmath16 and @xmath17 . the sequence @xmath11 may be empty ( that is , @xmath18 ) , in which case @xmath12 is called a _ constant of type @xmath17_. an _ @xmath7-algebra _ is an object @xmath19 , where for each @xmath20 , @xmath21 is a non - empty set , called the _ domain of sort @xmath22 _ of @xmath23 . for a nonempty sequence of sorts @xmath24 , we denote by @xmath25 the product @xmath26 . if @xmath27 , then @xmath28 is a total mapping from @xmath29 to @xmath30 . if @xmath12 is a constant of type @xmath22 , then @xmath28 is an element of @xmath31 . the objects @xmath28 are called the _ operations _ of @xmath23 . we assume that @xmath32 for @xmath33 . we also let @xmath23 denote the union of the @xmath21 ( @xmath34 ) . for @xmath35 , we let @xmath36 denote the unique @xmath37 such that @xmath38 . a mapping @xmath39 between @xmath7-algebras is a _ homomorphism _ ( or _ @xmath7-homomorphism _ if it is useful to specify the signature ) if it maps @xmath21 into @xmath40 for each sort @xmath22 and it commutes with the operations of @xmath7 . we denote by @xmath41 the set of finite well - formed terms built with @xmath42 ( we will call them _ @xmath7-terms _ ) , and by @xmath43 the set of those terms of sort @xmath22 ( the sort of a term is that of its leading symbol ) . if @xmath7 has no constant the set @xmath44 is empty . there is a standard structure of @xmath7-algebra on @xmath44 . its domain of sort @xmath22 is @xmath43 , and @xmath44 can be characterized as the _ initial @xmath7-algebra_. this means that for every @xmath7-algebra @xmath23 , there is a unique homomorphism @xmath45 . if @xmath46 , the image of @xmath47 under @xmath48 is an element of @xmath21 , also denoted by @xmath49 . it is nothing but the evaluation of @xmath47 in @xmath23 , where the function symbols are interpreted by the corresponding functions of @xmath23 . one can consider @xmath47 as a term _ denoting _ @xmath49 , and @xmath49 as the _ value _ of @xmath47 in @xmath23 . the set of values in @xmath23 of the terms in @xmath50 is called the _ subset generated _ by @xmath7 . we say that a subset of @xmath23 is _ finitely generated _ if it is the set of values of terms in @xmath51 for some finite subset @xmath52 of @xmath7 . let @xmath7 be an @xmath53-signature , @xmath54 be an @xmath55-signature where @xmath56 . we say that @xmath57 is a _ subsignature _ of @xmath7 , written @xmath58 , if @xmath57 is a subset of @xmath7 and the types of every @xmath12 in @xmath57 are the same with respect to @xmath7 and to @xmath57 . we say then that an @xmath57-algebra @xmath59 is a _ subalgebra _ of an @xmath7-algebra @xmath23 if @xmath60 for every @xmath61 , and every operation of @xmath59 coincides with the restriction to the domains of @xmath59 of the corresponding operation of @xmath23 . we will often encounter the case where an @xmath7-algebra @xmath23 is also the carrier of a @xmath62-algebra , and the @xmath62-operations of @xmath23 can be expressed as @xmath7-terms : in that case , we say that the @xmath62-operations of @xmath23 are _ @xmath7-derived _ , and the @xmath62-algebra @xmath23 is an _ @xmath7-derived algebra _ ( or it is _ derived from _ @xmath23 ) . more formally , an @xmath6-_sorted set of variables _ is a pair @xmath63 consisting of a set @xmath64 and a _ sort mapping _ @xmath65 ( usually denoted simply by @xmath64 ) . we let @xmath66 be the set of @xmath67-terms written with @xmath68 , where it is understood that the variables are among the nullary symbols ( constants ) of @xmath68 . @xmath69 denotes the subset of those terms of sort @xmath22 . now if @xmath70 is a finite sequence of pairwise distinct variables from @xmath64 and @xmath71 , we denote by @xmath72 the mapping from @xmath73 to @xmath21 associated with @xmath47 in the obvious way ( @xmath74 denotes the sequence of sorts of the elements of @xmath70 ) . we call @xmath72 a _ derived operation of the algebra _ @xmath23 . if @xmath70 is known from the context , we write @xmath49 instead of @xmath72 . this is the case in particular if @xmath47 is defined as a member of @xmath75 : the sequence @xmath70 is implicitly @xmath76 . let @xmath7 be an @xmath6-signature . an @xmath7-algebra @xmath23 is _ locally finite _ if each domain @xmath21 is finite . if @xmath23 is an @xmath7-algebra and @xmath34 is a sort , a subset @xmath2 of @xmath21 is @xmath23-_recognizable _ if there exists a locally finite @xmath42-algebra @xmath77 , a homomorphism @xmath78 , and a ( finite ) subset @xmath79 of @xmath80 such that @xmath81 . we denote by @xmath82 the family of @xmath23-recognizable subsets of @xmath21 . in some cases it will be useful to stress the relevant signature and we will talk of @xmath7-recognizable sets instead of @xmath23-recognizable sets . an equivalent definition can be given in terms of finite congruences . a _ congruence _ on @xmath23 is an equivalence relation @xmath83 on @xmath84 , such that each set @xmath31 is a union of equivalence classes , and which is stable under the operations of @xmath23 . it is _ locally finite _ if for each sort @xmath85 , the restriction @xmath86 of @xmath83 to @xmath21 has finite index . a congruence _ saturates a set _ if this set is a union of classes . a subset @xmath2 of @xmath21 is @xmath23-recognizable if and only if it is saturated by a locally finite congruence on @xmath23 . the following facts are easily verified from the definition of recognizability or its characterization in terms of congruences ( see @xcite ) , and will be used freely in the sequel . [ easy recog facts ] let @xmath23 be an @xmath7-algebra . * for each sort @xmath22 , the family @xmath82 contains @xmath21 and the empty set , and it is closed under union , intersection and difference . * if @xmath87 is a unary derived operation of @xmath23 or a homomorphism of @xmath59 into @xmath23 , ( where @xmath59 is another @xmath7-algebra ) , then the inverse image under @xmath87 of an @xmath23-recognizable set is recognizable . * if @xmath88 is a @xmath62-algebra with the same domain as @xmath23 , and if every @xmath62-congruence of @xmath88 is an @xmath7-congruence of @xmath23 ( e.g. @xmath88 is derived from @xmath23 , or @xmath62 is obtained from @xmath7 by adding constants ) , then every @xmath23-recognizable set is @xmath88-recognizable . if in addition @xmath62 contains @xmath7 , then @xmath23 and @xmath88 have the same recognizable subsets . * if @xmath59 is a subalgebra of @xmath23 and @xmath2 is an @xmath23-recognizable set , then @xmath89 is @xmath59-recognizable . this includes the case where @xmath59 has the same domain as @xmath23 , and is an @xmath57-algebra for some subsignature @xmath57 of @xmath7 . * suppose that @xmath23 is generated by @xmath7 and let @xmath48 be the evaluation homomorphism from @xmath44 onto @xmath23 . a subset @xmath2 of @xmath21 is @xmath7-recognizable if and only if @xmath90 is a recognizable subset of @xmath44 . if in addition @xmath7 is finite , then this is equivalent to the existence of a finite tree - automaton recognizing @xmath90 . on the set of all words over a finite alphabet @xmath77 , let us consider the binary operation of the concatenation product , and the unary operation @xmath91 , which is derived from the concatenation product . then the 3rd statement in proposition [ easy recog facts ] shows that we have the same recognizable subsets as if we considered only the concatenation product . it is interesting to note that , in contrast , adding the operation @xmath91 to the signature adds new equational languages , _ e.g. _ the set of all squares . we will see more technical conditions that guarantee the transfer of recognizability between algebras in section [ sec technical lemmas ] below . we gather here some observations on the significance of recognizability . first , we note that if @xmath12 is an operation of an @xmath7-algebra @xmath23 , with arity @xmath92 , and if @xmath93 are @xmath23-recognizable , then @xmath94 is not necessarily recognizable . this is discussed for instance in @xcite , where sufficient conditions are given to ensure that @xmath94 is recognizable . it is well - known for instance that the product of two recognizable subsets of the free monoid ( word languages ) or of the trace monoid is recognizable ; a similar result holds for recognizable sets of trees . now , let @xmath23 be an @xmath7-algebra and let @xmath57 be a signature which differs from @xmath7 only by the choice of constants and their values . in particular , @xmath57 may be obtained from @xmath7 by the addition of countably many new constants . then the congruences on @xmath23 are the same with respect to @xmath7 and to @xmath57 and it follows that a subset of @xmath23 is @xmath7-recognizable if and only if it is @xmath57-recognizable . it is customary to assume that the @xmath7-algebra @xmath23 is generated by the signature @xmath7 . if @xmath23 is a countable @xmath7-algebra that is not generated by @xmath7 , we can enrich @xmath7 to @xmath57 by adding to @xmath7 one constant of the appropriate sort for each element of @xmath23 . then @xmath57 generates @xmath23 ( in a trivial way ) . as noted above , @xmath23 has the same @xmath7- and @xmath57-recognizable subsets . if @xmath2 is one of these subsets , the set @xmath90 of @xmath57-terms is recognizable but we can not do much with it , because we lack the notion of a finite tree - automaton . see the conclusion of the paper for a further discussion of this point . finally , we can question the interest of the notion of a recognizable set . is it interesting in every algebra ? the answer is clearly no . let us explain why . if the algebraic structure over the considered set @xmath23 is poor , for example in the absence of non - nullary functions , then every set @xmath2 is recognizable , by a congruence with two classes , namely @xmath2 and its complement . the notion of recognizability becomes void . another extreme case is when the algebraic structure is so rich that there are very few recognizable sets . for an example , consider the set @xmath95 of natural integers equipped with the successor and the predecessor functions ( predecessor is defined by @xmath96 , @xmath97 ) . the only recognizable sets are @xmath95 and the empty set . indeed , if @xmath98 is a congruence and if @xmath99 for some @xmath100 , @xmath101 , then by using the function @xmath102 times , we find that @xmath103 . it follows ( using the successor function repeatedly ) that any two integers are equivalent . intuitively , if one enriches an algebraic structure by adding new operations , one gets fewer recognizable sets . for another example , let us consider the monoid @xmath104 of words over two letters . let us add a unary operation , the _ circular shift _ , defined by : @xmath105 and @xmath106 , @xmath107 , for every word @xmath108 . the language @xmath109 is no longer recognizable w.r.t . this new structure , however recognizability does not degenerate completely since every commutative language that is recognizable in the usual sense remains recognizable in the enriched algebraic structure . it is not completely clear yet which algebraic condition makes recognizability `` interesting '' . the statements in this section explain how to transfer a locally finite congruence from one algebra to another , possibly with a different signature , and hence how to transfer recognizability properties between algebras . proposition [ easy recog facts ] above contains examples of such results . the statements that follow will be used in the proof of some of our main results , in section [ sec gp ] . they are , unfortunately , heavily technical in their statements ( but not in their proofs ) [ fgh - congruence ] let @xmath7 be an @xmath53-signature and let @xmath62 be a @xmath110-signature . let @xmath111 be an @xmath7-algebra and let @xmath112 be a @xmath62-algebra . let also @xmath113 be a collection @xmath114 such that , for each @xmath115 and @xmath116 , @xmath117 consists of mappings from @xmath118 into @xmath119 with the following property : for each operation @xmath120 of type @xmath121 and for each @xmath122 , there exist sorts @xmath123 , mappings @xmath124 ( @xmath125 ) and an @xmath7-derived operation @xmath12 of type @xmath126 such that , for every @xmath127 , , @xmath128 , @xmath129 . finally , let @xmath130 be an @xmath7-congruence on @xmath111 and let @xmath83 be the equivalence relation defined , on each @xmath118 , by @xmath131 then @xmath83 is a @xmath62-congruence on @xmath112 . let @xmath132 be an operation in @xmath62 , of type @xmath121 , and let @xmath133 such that @xmath134 for each @xmath135 . let also @xmath136 with @xmath116 . by hypothesis , there exist sorts @xmath123 , mappings @xmath124 ( for @xmath137 ) and an @xmath7-derived operation @xmath12 of type @xmath126 such that @xmath138 since @xmath134 for each @xmath139 , we have @xmath140 ; and since @xmath130 is an @xmath7-congruence , it follows that @xmath141 . thus we have @xmath142 , which concludes the proof . with the notation of lemma [ fgh - congruence ] , for each sort @xmath115 , let @xmath143 be the quasi - order relation defined on @xmath144 by @xmath145 [ fgh - finiteness ] with the notation of lemma [ fgh - congruence ] , if for each @xmath146 the order relation associated with @xmath143 has a finite number of minimal elements , and if the @xmath7-congruence @xmath130 on @xmath111 is locally finite , then the @xmath62-congruence @xmath83 on @xmath112 is locally finite . let @xmath147 . we want to show that there are only finitely many @xmath83-classes in @xmath148 . by assumption , there exist elements @xmath149 such that every mapping of @xmath150 is of the form @xmath151 for some @xmath152 and some @xmath7-derived operation @xmath12 . for each @xmath139 , let @xmath153 be the range of @xmath154 and let @xmath155 be the number of @xmath130-classes in @xmath153 . it is immediately verified from the definition of @xmath143 that if @xmath156 , then @xmath157 if and only if @xmath158 for each @xmath152 . in particular , @xmath118 has at most @xmath159 @xmath83-classes , which concludes the proof . we will actually need even more technical versions of these lemmas . [ fgh - congruence+zeta(t ) ] let @xmath111 , @xmath112 , @xmath7 , @xmath62 and @xmath113 be as in lemma [ fgh - congruence ] , and let @xmath160 be a @xmath62-congruence on @xmath112 such that : for each operation @xmath120 of type @xmath121 , for each @xmath122 and for each @xmath161 where each @xmath162 is a @xmath160-class of @xmath163 , there exist sorts @xmath164 , mappings @xmath165 ( @xmath125 ) and an @xmath7-derived operation @xmath166 of type @xmath167 such that , in @xmath112 , @xmath168 if each @xmath169 is in @xmath162 . finally , let @xmath130 be an @xmath7-congruence on @xmath111 and let @xmath83 be the equivalence relation defined , on each @xmath118 , by @xmath170 then @xmath83 is a @xmath62-congruence on @xmath112 . moreover , if @xmath113 satisfies the hypothesis of lemma [ fgh - finiteness ] and @xmath130 and @xmath160 are locally finite , then @xmath83 is locally finite as well . the proof is the same as for lemmas [ fgh - congruence ] and [ fgh - finiteness ] . even though we are ultimately interested in studying sets of graphs , it will be convenient to handle the more general case of relational structures . furthermore , relational structures can be identified with simple directed hypergraphs . such hypergraphs form a natural representation of terms . see for instance the chapter on hypergraphs in @xcite for applications . in this paper , all graphs and structures are finite or countable . our proofs will not usually depend on cardinality assumptions on the graphs or structures , and hence our results will hold for finite as well as for infinite graphs or structures . however , recognizability in the algebraic sense we defined , is really interesting only for dealing with finitely generated objects , and hence for finite graphs and structures . for dealing with infinite words , trees and graphs , other tools are necessary , see for instance @xcite . let @xmath171 be a finite set of relation symbols , and @xmath79 be a finite set of nullary symbols . each symbol @xmath172 has an associated positive integer called its _ rank _ , denoted by @xmath173 . an @xmath174-structure is a tuple @xmath175 such that @xmath176 is a ( possibly empty ) set called the _ domain _ of @xmath111 , each @xmath177 is a @xmath173-ary relation on @xmath176 , i.e. , a subset of @xmath178 , and each @xmath179 is an element of @xmath176 , called the _ @xmath180-source _ of @xmath111 . we denote by @xmath181 the class of ( finite or countable ) @xmath174-structures , and we sometimes write @xmath182 for @xmath183 . by convention , isomorphic structures will be considered as equal . in the notation @xmath184 , @xmath185 stands for _ structures _ , while the second @xmath186 stands for _ sources_. a structure @xmath187 is _ source - separated _ if @xmath188 for @xmath189 . we will denote by @xmath190 the class of source - separated structures in @xmath181 . see corollary [ stssep recognizable ] and section [ sec source separation ] below . in order to handle graphs , we will consider particular kinds of structures in the sequel . we let @xmath191 be the set of relation symbols consisting of a single binary relation @xmath192 , intended to represent directed edges . thus graphs can be seen as the elements of @xmath193 , also written @xmath194 . clearly these graphs are directed , simple ( we can not represent multiple edges ) and they may have loops . for a discussion of graphs with multiple edges , see section [ sec simple multiple ] . we let @xmath195 denote the set @xmath196 . these structures are called _ graphs with sources_. we let @xmath197 denote the intersection @xmath198 . we will discuss also _ graphs with ports _ ( section [ sec gp ] ) : if @xmath199 is a finite set of unary relation symbols called _ port labels _ , then we denote by @xmath200 the set of relational symbols @xmath201 and by @xmath202 the class @xmath203 . port labels are useful for studying the clique - width of graphs , see @xcite and remark [ rk clique - width ] below . we first define some operations on structures . [ [ disjoint - union ] ] disjoint union + + + + + + + + + + + + + + let @xmath79 and @xmath204 be disjoint sets of constants and let @xmath187 and @xmath205 . let us also assume that @xmath111 and @xmath206 have disjoint domains . we denote by @xmath207 the union of @xmath111 and @xmath206 , which is naturally a structure in @xmath208 . if @xmath111 and @xmath206 are not disjoint , we replace @xmath206 by a disjoint copy . we need not be very precise on how to choose this copy because different choices will yield isomorphic @xmath209-sums , and we are interested in structures up to isomorphism . it is also possible to define a similar operation , without the restriction that @xmath79 and @xmath204 are disjoint ( as in , say , @xcite ) . see section [ sec sts - parallel ] below for a discussion . [ [ quantifier - free - definable - operations ] ] quantifier - free definable operations + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + our purpose is now to define functions from @xmath181 to @xmath210 by quantifier - free formulas . we denote by @xmath211 the set of quantifier - free formulas on @xmath174-structures with variables in @xmath212 . a _ qfd operation scheme _ from @xmath181 to @xmath210 is a tuple @xmath213 where @xmath214 , @xmath215 if @xmath216 is a @xmath173-ary relation symbol , @xmath217 , such that the following formulas are valid in every structure in @xmath181 , for all @xmath218 , @xmath219 and @xmath220 of arity @xmath173 : * @xmath221 ; * @xmath222 ; * @xmath223 ; * @xmath224 . the reason for these conditions becomes apparent with the following definition of the _ qfd operation _ @xmath225 defined by such a scheme . let @xmath187 . the domain of the structure @xmath226 is the subset of the domain of @xmath111 defined by formula @xmath227 and the relation @xmath216 ( @xmath220 ) on @xmath226 is described by formula @xmath228 . finally , if @xmath219 , then @xmath229 if @xmath230 and @xmath111 satisfies @xmath231 . the first two conditions imposed above assert that relative to @xmath111 , @xmath180 is uniquely defined for each @xmath232 , the third condition asserts that @xmath233 always lies in the domain of @xmath226 , and the fourth condition asserts that the relation @xmath228 ( @xmath234 ) can only relate elements of the domain of @xmath226 . note that in the first condition , @xmath235 does not mean that @xmath180 and @xmath236 are the same constant , but that they have the same value in the considered structure . the conditions to be verified by a qfd operation scheme are decidable . it follows that the notion of a qfd operation scheme is effective . see the appendix ( remark [ remark a4 ] in particular ) for a discussion of this decidability result . [ def hr ops ] let @xmath171 be a finite set of relational symbols , @xmath79 be a finite set of source labels and let @xmath237 be source labels . we define the following operations . * if @xmath238 and @xmath239 , @xmath240 is the unary operation of type @xmath241 which renames the @xmath242-source of a structure to a @xmath243-source ; * if @xmath238 , @xmath244 is the unary operation of type @xmath245 which forgets the @xmath242-source of a structure ; * if @xmath246 , @xmath247 is the unary operation of type @xmath248 which identifies the @xmath242-source and the @xmath243-source of a structure ( so the resulting domain element is both the @xmath242-source and the @xmath243-source ) , and reorganizes the tuples of the relational structure accordingly . note that the operation names @xmath240 , @xmath244 and @xmath247 are overloaded : they denote different operations when the sets @xmath171 and @xmath79 are allowed to vary . a completely formal definition would use operation names such as @xmath249 , which would be inconvenient . it is immediately verified that the operations of the form @xmath240 and @xmath244 are qfd . it is probably worth showing explicitly a qfd operation scheme defining the operation @xmath247 . let @xmath250 be the formula @xmath251 . if @xmath172 has arity @xmath252 , let @xmath253 be the formula @xmath254 where for each @xmath255 , @xmath256 if @xmath257 and @xmath258 otherwise . for each @xmath259 such that @xmath260 and for each @xmath261 , let @xmath231 be the formula @xmath262 ; let @xmath263 be the formula @xmath264 , and let @xmath265 be the formula @xmath266 for each @xmath267 . it is now routine to verify that the scheme @xmath268 defines @xmath247 . there is no qfd operation from @xmath182 into @xmath269 if @xmath270 , because in the absence of constants in the input structure , we can not define constants in the output structure . [ example silent ] the natural inclusion of @xmath181 into @xmath271 when @xmath272 contains @xmath171 is a qfd operation in natural way : the formulas intended to define relations in @xmath273 are taken to be identically @xmath274 . [ [ the - signature - cals ] ] the signature @xmath275 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we define the algebra @xmath184 of _ structures with sources _ as follows . first , let us fix once and for all a countable set of relation symbols containing @xmath192 and countably many relation symbols of each arity , and a countable set of constants . in the sequel , finite sets of relation symbols @xmath171 and finite sets of constants @xmath79 will be taken in these fixed sets . the set of sorts consists of all such pairs @xmath174 . the set of elements of @xmath184 of sort @xmath174 is @xmath181 . the signature @xmath275 consists of the following operations ( interpreted in @xmath184 ) . first , for each pair of sorts @xmath174 and @xmath276 such that @xmath277 , the disjoint union @xmath209 is an operation of type @xmath278 . note that we overload the symbol @xmath209 , that is , we denote in the same way an infinite number of operations on @xmath184 . next , every qfd operation is a ( unary ) operation in @xmath275 . finally , we observe that the signature @xmath275 contains the natural inclusions of @xmath181 into @xmath271 when @xmath272 contains @xmath171 , which are qfd ( example [ example silent ] ) . as for constants in @xmath275 , one can pick a single source label @xmath242 , and consider a single constant @xmath279 , denoting the structure with a single element , which is an @xmath242-source , and no relations . together with the operations in @xmath275 , this constant suffices to generate all finite relational structures . as noted in section [ significance ] , the choice of constants does not affect recognizability . it only affects the generating power of the signature , but this is not our point in this paper . we first consider the composition of qfd operations . [ composition of qfd ] qfd operations in @xmath184 are closed under composition ( whenever types fit for defining meaningful composition ) . let @xmath280 and @xmath281 be qfd operations , given respectively by the schemes @xmath282 and @xmath283 . the composite @xmath284 turns an @xmath174-structure into an @xmath285-structure . let @xmath286 , @xmath287 ( @xmath288 ) and @xmath289 ( @xmath290 , @xmath291 ) be obtained from @xmath292 , @xmath293 and @xmath294 by replacing every occurrence of @xmath295 ( @xmath220 ) by @xmath296 ; our formulas are now in the language of @xmath297-structures and we need to `` translate '' the constants @xmath219 into elements of @xmath79 . however , this translation , a mapping from @xmath204 to @xmath79 , depends on the structure in which we operate . to reflect this observation , for each mapping @xmath298 , we let @xmath299 be the conjunction of the formulas @xmath300 ( @xmath219 ) and the formula obtained from @xmath286 by replacing each occurrence of @xmath232 ( @xmath219 ) by @xmath301 . finally , we let @xmath302 be the disjunction of the @xmath303 when @xmath87 runs over all mappings from @xmath204 to @xmath79 . we proceed in the same fashion to define @xmath304 and @xmath305 for each @xmath288 and each @xmath306 , @xmath291 . finally , if @xmath307 and @xmath291 , we let @xmath308 . it is a routine verification that @xmath309 is a qfd operation scheme , which defines the composite operation @xmath284 . this completes the proof . for each @xmath187 , we define the _ type of @xmath111 _ , written @xmath310 , to be the restriction of @xmath111 to its set of sources . that is : the domain of @xmath310 is the set of @xmath79-sources of @xmath111 , and the relations of @xmath310 are those tuples of @xmath79-sources that are relations in @xmath111 . in order to simplify notation , we also denote by @xmath160 the equivalence relation on @xmath184 given by @xmath311 [ type equivalence ] let @xmath312 . then @xmath313 if and only if @xmath111 and @xmath112 satisfy the same formulas in @xmath314 . a formula in @xmath314 is a boolean combination of atoms of the form @xmath262 where @xmath315 , or @xmath316 where @xmath172 has arity @xmath5 and the @xmath169 are in @xmath79 . it is immediate that such an atom is true in @xmath111 if and only if it is true in @xmath310 . thus @xmath111 and @xmath310 satisfy the same formulas in @xmath314 : in particular , @xmath160-equivalent structures satisfy the same formulas in @xmath314 . thus , if we denote by @xmath317 the set of formulas in @xmath314 that are satisfied by @xmath111 ( see section [ sec ffvv ] ) , we find that @xmath318 . conversely , we observe that if @xmath111 is a structure in @xmath181 , which consists only of its @xmath79-sources ( that is , @xmath319 ) , then @xmath111 is entirely described by some formula in @xmath314 . thus , if @xmath320 , then @xmath321 . this suffices to conclude the proof . the type relation @xmath160 has the following important property . [ ffvv light ] the type relation @xmath160 is a locally finite congruence on @xmath184 . the verification that @xmath322 ( @xmath187 , @xmath205 and @xmath277 ) is immediate . let us now consider a qfd operation @xmath280 , specified by the qfd operation scheme @xmath323 . by lemma [ type equivalence ] , @xmath111 and @xmath310 satisfy the same formulas of @xmath314 . in particular , for each @xmath230 and @xmath219 , @xmath111 and @xmath310 both satisfy @xmath231 , or both satisfy its negation . thus @xmath226 and @xmath324 have the same sources , and hence @xmath325 . we have just shown that the type relation is a congruence . to complete the proof , it suffices to show that for each sort @xmath174 , the set of types of sort @xmath174 , that is , the set @xmath326 is finite . note that if @xmath187 , then @xmath310 has cardinality at most @xmath327 ( and also at most @xmath328 ) . it follows that @xmath329 . proposition [ ffvv light ] can be seen as a particular case of a result of feferman and vaught @xcite , theorem [ ffvv ] below , which will be used in section [ sec knn ] . the simple formulation above will be very useful . note that the knowledge of @xmath310 is sufficient to determine whether @xmath111 is a source - separated structure . this observation is used to prove the following corollary . [ stssep recognizable ] let @xmath174 be a sort in @xmath184 . then @xmath190 is a recognizable subset of @xmath181 . whether a structure @xmath111 is source - separated depends only on its type @xmath310 : in particular , the type congruence @xmath160 saturates @xmath190 . by proposition [ ffvv light ] , this relation is a locally finite congruence , and hence @xmath190 is recognizable . if @xmath330 is a sort of @xmath184 , we denote by @xmath331 the set of closed first - order formulas over @xmath171 and @xmath79 . for each integer @xmath232 , we denote by @xmath332 the set of those formulas of quantifier - depth at most @xmath232 . up to a decidable syntactic equivalence ( taking into account boolean laws , properties of equality , renaming of quantified variables , see appendix [ appendix ] ) , there are only finitely many formulas in each set @xmath332 . thus , we can reason as if @xmath332 was actually finite . for an @xmath174-structure @xmath111 , we let its _ @xmath333-theory _ be the set @xmath334 of formulas in @xmath332 that are valid in @xmath111 . it is finite since it is a subset of the finite set @xmath332 . [ ffvv ] let @xmath335 . * for every qfd operation @xmath12 of type @xmath336 , there exists a mapping @xmath337 such that , for every @xmath174-structure @xmath111 @xmath338 * for every @xmath174 and @xmath276 , where @xmath79 and @xmath204 are disjoint , there exists a binary function @xmath339 such that , for every @xmath174-structure @xmath111 , and every @xmath276-structure @xmath206 , @xmath340 the second assertion was proved in @xcite for first - order logic , and extended by shelah to monadic second - order logic @xcite . the importance of this result is discussed by makowsky in @xcite . the functions @xmath337 and @xmath339 have finite domains and codomains . however these sets are quite large . these functions can be ( at least in principle ) effectively determined for given @xmath174 , @xmath276 , and @xmath232 . in the literature on recognizable and equational graph languages , several variants of the signature @xmath275 and the algebra @xmath184 are considered , notably a variant where the definition of the disjoint union is replaced by a more general parallel product , and a variant where all structures are assumed to be source - separated . we verify in this section that these variants do not yield different notions of recognizability . in the literature ( e.g. @xcite ) , the operation of disjoint union @xmath209 is sometimes replaced by the so - called _ parallel composition _ ( or _ product _ ) , written @xmath341 , an operation of type @xmath342 for which we do not assume that @xmath79 and @xmath204 are disjoint . if @xmath343 and @xmath205 , the parallel composition @xmath344 is obtained by taking the ( set - theoretic ) disjoint union of @xmath111 and @xmath206 and then identifying the @xmath180-sources of @xmath111 and @xmath206 for each @xmath345 . let @xmath346 denote the signature obtained from @xmath275 by substituting @xmath341 for @xmath209 . [ prop sts - parallel ] let @xmath2 be a subset of @xmath184 . then @xmath2 is @xmath275-recognizable if and only if it is @xmath346-recognizable . we first observe that the operation @xmath209 is a particular case of @xmath341 . therefore @xmath275 is a sub - signature of @xmath346 and hence , every @xmath346-recognizable set is @xmath275-recognizable . to prove the converse , it suffices to verify that @xmath341 is an @xmath275-derived operation by proposition [ easy recog facts ] . indeed , if @xmath187 and @xmath205 , the parallel composition @xmath344 can be obtained by the following sequence of @xmath275-operations ( see example [ def hr ops ] for their definition ) : - for each @xmath345 , apply the qfd operation @xmath347 which renames the @xmath180-source in @xmath206 with a new source label , say @xmath348 , not in @xmath79 ; let @xmath349 be the resulting structure ; - take the disjoint union @xmath350 ; - for each @xmath345 , apply the operation @xmath351 which identifies the @xmath180-source and the @xmath348-source in @xmath352 ; - apply the source - forgetting operation @xmath353 for each @xmath345 . the property that @xmath354 for @xmath355 is called _ source separation_. this property makes it easier to work with operations on structures and graphs , and hence we discuss a variant of the @xmath275-algebra @xmath184 , which handles source - separated structures . we will also use it in section [ sec knn ] . recall that @xmath190 denotes the set of source - separated structures in @xmath181 . we now define a subsignature @xmath356 of @xmath275 such that @xmath357 is a sub - algebra of @xmath184 . disjoint union @xmath209 clearly preserves source separation , and is part of @xmath356 . next we include in @xmath356 the operations specified by qfd operation schemes such that , for each @xmath230 and @xmath358 ( see the notation in section [ sec algebra sts ] ) , @xmath359 which guarantees that the operation preserves source separation . [ exemple hrsep ] the operations @xmath240 and @xmath244 defined in example [ def hr ops ] are in @xmath356 . the operation @xmath247 defined in the same example is not . in contrast , the operation written @xmath360 , which identifies the @xmath242-source and the @xmath243-source of a structure as in @xmath247 , and makes the resulting element of the domain a @xmath243-source but not an @xmath242-source , preserves source separation . it can be written as @xmath361 . the operation which , given a graph with source labels @xmath242 and @xmath243 , exchanges the source labels @xmath242 and @xmath243 if the corresponding vertices are linked by an edge and does nothing otherwise , is another example of a qfd operation in @xmath356 . regarding the effectiveness of the definition of @xmath356 , we observe the following . given a qfd operation scheme , one can decide whether the corresponding qfd operation preserves source separation . let @xmath132 be the qfd operation specified by the given qfd operation scheme , and let @xmath181 be the domain of @xmath132 . one can effectively construct the images under @xmath132 of every type in @xmath181 , since there are only finitely many of them , and they can all be enumerated . one can then verify whether the operation preserves souce - separation on types . now it follows from the proof of proposition [ ffvv light ] that for each @xmath187 , we have @xmath362 . in particular , @xmath132 preserves source separation if and only if it preserves it for the structures of the form @xmath310 . thus one can effectively decide whether @xmath363 . we now show that the restriction to source - separated structures does not change the notion of recognizability . [ thm ssep - recognizable ] let @xmath2 be a subset of @xmath357 . then @xmath2 is @xmath275-recognizable if and only if it is @xmath356-recognizable . by definition , @xmath356 is a subsignature of @xmath275 , so every @xmath275-recognizable set is @xmath356-recognizable . to prove the converse , we first define a mapping @xmath87 , which maps a structure @xmath343 to a source - separated structure @xmath364 by splitting sources that were identified in @xmath111 . we assume that the countable set of constant symbols ( from which @xmath79 is taken , see section [ sec algebra sts ] ) is linearly ordered . let @xmath365 be given by @xmath366 we let @xmath367 and @xmath368 . the structure @xmath369 has domain set the disjoint union of @xmath111 and @xmath370 . for each @xmath371 , the @xmath180-source of @xmath369 is the @xmath180-source of @xmath111 , and for each @xmath372 , the @xmath180-source of @xmath369 is the element @xmath372 . finally , for each @xmath172 , the relation @xmath373 equals the relation @xmath177 ( so it does not involve the elements of @xmath370 ) . observe that @xmath87 is not a qfd operation , and that @xmath374 , @xmath375 and @xmath370 depend only on @xmath310 . now let @xmath2 be an @xmath356-recognizable subset of @xmath357 and let @xmath130 be a locally finite @xmath356-congruence recognizing it . we need to construct a locally finite @xmath275-congruence @xmath98 on @xmath184 which recognizes @xmath2 . the relation @xmath98 on @xmath184 is defined as follows . if @xmath312 , we say that @xmath376 if @xmath377 and @xmath378 . it is immediately verified that @xmath98 is an equivalence relation . moreover , the @xmath98-class of a structure @xmath111 is determined by its @xmath160-class , and by the @xmath130-class of @xmath369 . since both @xmath160 and @xmath130 are locally finite , @xmath98 also is locally finite . let us now prove that @xmath98 is an @xmath275-congruence . let @xmath379 and @xmath380 , with @xmath381 . by proposition [ ffvv light ] , @xmath382 . it is not difficult to verify that @xmath383 it follows that @xmath384 since @xmath209 is an operation in @xmath356 . thus @xmath385 . next let @xmath132 be a qfd operation from @xmath181 to @xmath386 , given by the qfd operation scheme @xmath387 . let @xmath111 and @xmath112 be @xmath98-equivalent elements of @xmath181 , which will remain fixed for the rest of this proof . we need to show that @xmath388 . we already know from proposition [ ffvv light ] that if @xmath389 , then @xmath390 , and we want to show that @xmath391 . since @xmath390 , the mappings @xmath392 and @xmath393 , from @xmath394 to @xmath394 , coincide . let @xmath395 and @xmath396 . without loss of generality , we may assume that @xmath397 . the domain set of @xmath398 ( resp . @xmath399 ) is the disjoint union of the domain of @xmath226 ( resp . @xmath400 ) and @xmath401 . it suffices to show that there exists a qfd operation @xmath402 , depending on @xmath132 and @xmath310 , such that @xmath403 and @xmath404 ( where @xmath401 is the source - only element of @xmath405 ) . indeed , the fact that @xmath130 is an @xmath356-congruence will then imply that @xmath406 . let @xmath292 be obtained from @xmath227 by replacing every occurrence of @xmath230 by @xmath407 . for each @xmath408 , @xmath230 and @xmath409 , let @xmath410 be obtained from @xmath411 and @xmath412 be obtained from @xmath413 in the same fashion . let now @xmath414 be defined by the scheme @xmath415 it is now a routine verification that ( for our fixed structure @xmath111 ) @xmath416 . since all our definitions depend only on @xmath310 , we also have @xmath417 . one last step is required in this proof as the qfd operation @xmath418 may not preserve source separation for all structures , that is , @xmath418 may not lie in @xmath356 . it does for the particular structures @xmath419 and @xmath420 , but perhaps not for others . actually , structures @xmath421 such that @xmath422 do not matter in this context , so we can replace @xmath418 by the operation @xmath92 , with the same domain and range as @xmath418 , which maps a structure @xmath421 to @xmath423 if @xmath424 , and to the source - only source - separated structure @xmath425 where all relations are empty . this new operation @xmath92 preserves source separation by construction , and it is easily verified to be qfd . this completes ( at last ) the proof . graphs with ports were introduced in section [ sec rel structures ] . recall that if @xmath199 is a set of unary relation symbols , then @xmath200 denotes the set @xmath427 and the class of graphs with ports in @xmath199 , written @xmath202 can be identified with @xmath203 . we observe that a vertex of a graph with ports in @xmath199 can be a @xmath428-port for one or several port labels @xmath429 , or for none at all . for convenience , we will consider that @xmath199 is a finite subset of the set of natural integers . we define the set of sorts of the algebra @xmath426 to be the set of finite subsets of . for each such subset @xmath199 , the set of elements of @xmath426 of sort @xmath199 is the set @xmath202 of graphs with ports in @xmath199 . the signature @xmath1 consists of constants , unary operations and binary operations . these operations ( interpreted in @xmath426 ) are as follows . first , if @xmath430 are finite subsets of , then @xmath209 is as in @xmath184 , and is thus a binary operation of type @xmath431 . in @xmath426 , we consider @xmath209 as an operation of type @xmath432 . next , the unary operations of @xmath1 are the following ( clearly qfd ) operations : * if @xmath433 are distinct integers , @xmath434 is an operation of type @xmath435 for each sort @xmath199 such that @xmath436 : it modifies neither the domain ( the set of vertices ) nor the unary relations @xmath428 ( @xmath429 ) ; the new edge relation has the existing edges , _ plus _ every edge from a @xmath428-port to a @xmath437-port : it is given by @xmath438 * if @xmath439 is a finite subset of @xmath440 , @xmath441 is an operation of type @xmath442 where @xmath199 is any finite set containing the domain of the relation @xmath439 and @xmath443 is any finite set containing the range of @xmath439 ; it modifies neither the domain ( set of vertices ) nor the edge relation ; for each @xmath408 , the @xmath437-ports of the output structure are the vertices of the input structure that are @xmath428-ports for some @xmath428 such that @xmath444 ; that is , @xmath445 is given by @xmath446 . finally , for each integer @xmath428 , we let @xmath447 be the constant of type @xmath448 denoting the graph with a single vertex , no edges , and whose vertex is a @xmath428-port . we also let @xmath449 be the same graph , with a single loop . [ rename , forget ] the following operations on graphs with ports occur in the literature , and are particular cases of @xmath1-operations . let @xmath450 be integers , @xmath199 be a subset of containing @xmath428 and @xmath451 . the operation @xmath452 , of type @xmath442 which _ renames _ every @xmath428-port to a @xmath437-port , is an operation of @xmath1 : it is equal to @xmath441 where @xmath453 . observe that this operation fuses the sets of vertices defined by @xmath428 and @xmath437 . let @xmath428 be an integer , and let @xmath199 be a subset of containing @xmath428 . the operation @xmath454 , of type @xmath455 , which _ forgets _ @xmath428-ports is an operation of @xmath1 : it is equal to @xmath441 where @xmath456 . in our definition of graph with ports , an element of @xmath457 does not need to have @xmath437-ports for each @xmath408 . thus , if @xmath458 , every graph with ports in @xmath199 can also be viewed as a graph with ports in @xmath443 . the natural inclusion of @xmath202 into @xmath457 is part of the signature @xmath1 : it is equal to @xmath441 where @xmath459 . again ( as in example [ def hr ops ] ) , the operations introduced in this section are denoted by overloaded symbols . a formal definition should specify the type of the operation , and would read something like @xmath460 or @xmath461 . we prefer the more concise notation introduced here . the following result describes the action of a qfd operation on a disjoint union of structures . it is the key to the main results of this section , described in section [ vr equivalence ] below . [ unary - parallel ] let @xmath160 be the type congruence ( see section [ sec elementary properties ] ) . let @xmath87 be a unary qfd operation on @xmath184 , from @xmath181 to @xmath462 , let @xmath463 and @xmath464 be sorts of @xmath184 such that @xmath465 , @xmath466 and @xmath467 , and let @xmath468 with @xmath469 a @xmath160-class in @xmath470 and @xmath471 a @xmath160-class in @xmath472 . then there exist quantifier - free definable operations @xmath473 , @xmath474 , and @xmath475 , such that * @xmath166 is a composition of unary operations in @xmath1 ; * for each @xmath476 in class @xmath469 and each @xmath477 in class @xmath471 , @xmath478 . let @xmath479 be the qfd operation scheme defining the operation @xmath87 : here @xmath480 defines the @xmath192 relation , @xmath411 defines the @xmath437-ports ( @xmath408 ) , and there is no formula of the form @xmath231 since the range of @xmath87 is in @xmath481 . the formulas @xmath227 , @xmath480 and @xmath411 , for @xmath408 , are in the language of @xmath174-structures . the atoms of @xmath482 are either of the form @xmath295 ( @xmath172 ) , or @xmath483 , or @xmath484 ( @xmath485 ) . let @xmath486 be the formula obtained from @xmath482 by substituting the boolean value @xmath487 ( false ) for the following atoms , which are certainly false in a disjoint sum @xmath488 , with @xmath476 , @xmath477 and the variable @xmath489 interpreted in @xmath490 : * each @xmath216-atom such that @xmath491 and an argument of @xmath216 is @xmath489 or a constant in @xmath492 ; * each @xmath216-atom such that @xmath493 and an argument of @xmath216 is a constant in @xmath494 ; * each @xmath216-atom such that @xmath495 , an argument of @xmath216 is a constant in @xmath494 , and another argument of @xmath216 is @xmath489 or a constant in @xmath492 ; * each atom of the form @xmath496 such that @xmath497 and @xmath498 is equal to @xmath489 or to a constant in @xmath492 . the remaining atoms in @xmath499 are either in @xmath500 or in @xmath501 . note that the @xmath160-class of an element of @xmath472 determines entirely which formulas in @xmath501 it satisfies . for each @xmath502 as in the statement of the proposition , we let @xmath503 be the formula in @xmath500 obtained from @xmath486 by replacing each atom in @xmath501 by the boolean value 0 or 1 according to the @xmath160-class @xmath471 . we observe that if @xmath489 is a vertex of @xmath488 which happens to be in @xmath490 , then @xmath504 for each @xmath408 , let @xmath505 be defined similarly . then we also have , if @xmath489 is a vertex of @xmath488 in @xmath490 , @xmath506 let also @xmath507 and @xmath508 be defined dually . and again , if @xmath509 , we let @xmath510 be the formula obtained from @xmath480 by substituting the boolean value @xmath487 for the atoms that are certainly false in a disjoint sum @xmath488 for the variable @xmath489 interpreted in @xmath169 and the variable @xmath511 interpreted in @xmath512 : * each @xmath216-atom such that @xmath513 and @xmath489 is an argument of @xmath216 ; * each @xmath216-atom such that @xmath514 and @xmath511 is an argument of @xmath216 ; * each @xmath216-atom such that @xmath491 and a constant in @xmath492 is an argument of @xmath216 ; * each @xmath216-atom such that @xmath493 and a constant in @xmath494 is an argument of @xmath216 ; * each @xmath216-atom such that @xmath495 , an argument of @xmath216 is a constant in @xmath494 , and another argument of @xmath216 is a constant in @xmath492 ; * each @xmath216-atom such that @xmath495 , an argument of @xmath216 is @xmath489 ( resp . @xmath511 ) and another argument of @xmath216 is a constant in @xmath515 ( resp . @xmath516 ) ; * each atom of the form @xmath483 with @xmath517 , @xmath518 with @xmath519 , or @xmath520 with @xmath521 and @xmath522 ; * if @xmath523 , each @xmath216-atom such that @xmath495 , and @xmath489 and @xmath511 are arguments of @xmath216 . as above , the remaining atoms in @xmath524 are in @xmath525 , and for each @xmath502 , we let @xmath526 be obtained from @xmath524 by substituting the boolean values 0 or 1 for the atoms in @xmath501 according to the @xmath160-class @xmath471 . if @xmath527 are vertices of @xmath488 in @xmath490 , and if the @xmath160-class of @xmath528 is @xmath471 , then @xmath529 we define @xmath530 similarly , and get the analogous equivalence . if @xmath523 , the atoms of @xmath531 are in @xmath532 and in @xmath533 which may include atoms in @xmath534 and in @xmath501 . again , we let @xmath535 be obtained from @xmath531 by substituting the boolean values 0 or 1 for the atoms without free variables according to the @xmath160-classes @xmath469 and @xmath471 . and we observe that if @xmath527 are vertices of @xmath488 , @xmath489 is in @xmath169 and in the @xmath160-class @xmath162 , @xmath511 is in @xmath512 and in the @xmath160-class @xmath536 , then @xmath537 now let @xmath538 , let @xmath539 be an enumeration of the subsets of @xmath540 , and let @xmath541 be an enumeration of the subsets of @xmath542 . let us denote by @xmath543 the set @xmath544 and by @xmath545 the set @xmath546 . we define the qfd operation @xmath547 defined by the following operation scheme : @xmath548 where for each @xmath549 , @xmath550 holds if the set of quantifier - free formulas in @xmath540 satisfied by @xmath489 is exactly @xmath551 . similarly , the qfd operation @xmath552 is defined by the operation scheme @xmath553 where for each @xmath554 , @xmath550 holds if the set of quantifier - free formulas in @xmath542 satisfied by @xmath489 is exactly @xmath551 . finally , we consider structures @xmath476 and @xmath477 , with @xmath160-classes respectively @xmath469 and @xmath471 , and we compare the graphs with ports @xmath555 and @xmath556 . the above remarks show that these two graphs have the same set of vertices , the same @xmath437-ports ( @xmath408 ) , and the same edges between two vertices of @xmath490 or two vertices of @xmath528 . on the other hand , @xmath555 misses the edges of @xmath557 that connect a vertex of @xmath490 with a vertex of @xmath528 . these edges are captured by the formulas @xmath558 and @xmath559 . now , if @xmath489 is a vertex of @xmath490 and @xmath511 is a vertex of @xmath528 , we already observed that the truth values of @xmath560 and @xmath561 are entirely determined by the quantifier - free formulas with one free variable satisfied by @xmath489 in @xmath490 and by @xmath511 in @xmath528 : that is , they are entirely determined by the ( unique ) index @xmath562 such that @xmath550 and by the ( unique ) index @xmath554 such that @xmath563 . in other words , @xmath564 and @xmath565 are equivalent to disjunctions of conjunctions of the form @xmath566 thus the edges in @xmath556 from a vertex of @xmath490 to a vertex of @xmath528 can be created from @xmath567 by applying repeatedly the operations ( in @xmath1 ) of the form @xmath568 such that @xmath569 $ ] , @xmath570 is a disjunct of @xmath558 . similarly , the edges in @xmath556 from a vertex of @xmath528 to a vertex of @xmath490 can be created from @xmath567 by applying the appropriate operations of the form @xmath571 . the last operation consists in forgetting the auxiliary ports numbered @xmath572 to @xmath573 , that is , in applying the operation @xmath441 , with @xmath574 . in this section , we consider different notions of recognizability that can be used for sets of graphs with ports . let @xmath575 . then @xmath2 can be @xmath1-recognizable , as a subset of the @xmath1-algebra @xmath426 . it can also be @xmath275-recognizable , as a subset of the @xmath275-algebra @xmath184 since @xmath576 . finally , we introduce another signature , written @xmath577 , on @xmath426 : it is obtained from @xmath1 by adding all the qfd operations between the sorts of @xmath426 . [ thm vr equivalence ] let @xmath199 be a finite subset of @xmath578 and let @xmath2 be a subset of @xmath202 . the following properties are equivalent : 1 : : @xmath2 is @xmath275-recognizable ; 2 : : @xmath2 is @xmath577-recognizable ; 3 : : @xmath2 is @xmath1-recognizable ; since the operations of @xmath1 are operations of @xmath577 , and the operations of @xmath577 are operations of @xmath275 , it follows from proposition [ easy recog facts ] that ( 1 ) implies ( 2 ) , and ( 2 ) implies ( 3 ) . thus , we only need to verify that ( 3 ) implies ( 1 ) . we use lemma [ fgh - congruence+zeta(t ) ] , with @xmath579 , @xmath580 , @xmath581 , @xmath582 , and @xmath160 the type congruence ( see section [ sec elementary properties ] ) , which relates structures with sources of the same sort , provided they satisfy the same quantifier - free formulas . we use the collection @xmath113 of sets @xmath583 of unary qfd operations from @xmath181 to @xmath202 . let @xmath2 be a @xmath1-recognizable subset of @xmath202 and let @xmath130 be a locally finite @xmath1-congruence on @xmath426 such that @xmath2 is a union of @xmath130-classes . since @xmath160 is a locally finite @xmath275-congruence on @xmath184 ( proposition [ ffvv light ] ) , its restriction to @xmath426 is also a locally finite @xmath1-congruence ; and the intersection of @xmath130 and @xmath160 is a locally finite @xmath1-congruence on @xmath426 which saturates @xmath2 . thus we can assume , without loss of generality , that @xmath130-equivalent elements of @xmath426 are also @xmath160-equivalent . next we consider the equivalence relation @xmath83 on @xmath184 defined as in lemma [ fgh - congruence+zeta(t ) ] . note that the identity of @xmath202 belongs to @xmath584 , so that @xmath83-equivalent elements of @xmath585 are also @xmath130-equivalent . in particular , @xmath83 saturates @xmath2 and it suffices to show that @xmath83 is locally finite and is a @xmath275-congruence . in view of lemma [ fgh - congruence+zeta(t ) ] , it is enough to verify that @xmath113 satisfies the assumptions of lemma [ fgh - congruence ] and [ fgh - finiteness ] . we first verify the hypothesis of lemma [ fgh - congruence ] . let @xmath132 be an operation of @xmath275 : either @xmath132 is a unary qfd operation or @xmath586 . in the latter case , proposition [ unary - parallel ] states precisely that the required property holds . if @xmath132 is a qfd operation of type @xmath587 , and @xmath588 , then @xmath589 is a qfd operation ( lemma [ composition of qfd ] ) and hence , @xmath590 . now letting @xmath12 be the identity mapping of @xmath202 , we find that @xmath591 as required . in this case , @xmath592 and @xmath12 do not depend on the @xmath160-class of @xmath593 . next , we turn to the verification of the hypothesis of lemma [ fgh - finiteness ] . let @xmath594 , , @xmath595 be an enumeration of the elements of @xmath596 and let @xmath597 , , @xmath598 be an enumeration of the elements of @xmath599 . thus , a qfd operation scheme from @xmath181 into @xmath457 consists in the choice of a formula @xmath600 ( @xmath601 ) , a formula @xmath602 ( @xmath603 ) , a sequence of formulas @xmath604 ( @xmath605 ) , and a partition of @xmath443 as @xmath606 : if @xmath607 , then @xmath608 . ( if @xmath609 , then @xmath610 . ) let us now consider two unary qfd operations @xmath611 and @xmath612 , associated with the same choice of values @xmath613 , @xmath614 and @xmath615 . let @xmath616 and @xmath617 be the corresponding partitions of @xmath443 and @xmath618 . finally let @xmath619 be the following operations in the signature @xmath1 . these operations have the common particularity to not alter the graph structure , and to modify only the port predicates . the mapping @xmath620 shifts every port index of an element of @xmath457 by @xmath621 , to yield a graph with ports in @xmath622 , whose port names do not intersect @xmath618 . we let @xmath623 for @xmath624 . for @xmath625 , @xmath626 where @xmath627 thus @xmath628 turns a graph with ports in @xmath629 into a graph with ports in @xmath630 , with the same vertex set , the same edge relation , the same @xmath437-ports for each @xmath631 , and with each @xmath216-port ( @xmath632 ) turned into a @xmath437-port for each @xmath633 . it is now an easy verification that , if @xmath634 , then @xmath635 for each @xmath636 . thus the quasi - order @xmath637 defined in lemma [ fgh - finiteness ] is in fact a finite index equivalence relation , and this concludes the proof . this actually proves also that we get the same recognizable sets of graphs with ports , if we consider @xmath457 as a domain of sort @xmath443 in the algebra of structures _ without _ sources which consists of the domains @xmath183 equipped with the operations of @xmath275 between them . if we were only interested in the equivalence of this recognizability with @xmath1- and @xmath577-recognizability ( or just the equivalence between @xmath1- and @xmath577-recognizability ) , we could do with lemmas [ fgh - congruence ] and [ fgh - finiteness ] instead of lemma [ fgh - congruence+zeta(t ) ] , and with a simpler version of proposition [ unary - parallel ] , making no reference to @xmath160 . the first variant considered here replaces the signature @xmath1 by a smaller signature , which we will see is equivalent to @xmath1 in terms of recognizability . the second one concerns a certain class of graphs with ports , and is central in the definition of the clique - width of a finite graph . in section [ vr equivalence ] , we exhibited signatures larger than @xmath1 , for which all the @xmath1-recognizable sets of graphs with ports are recognizable : namely the signature @xmath577 on @xmath426 and the signature @xmath275 on the wider algebra @xmath184 . in contrast , we exhibit in this section a smaller signature ( in fact , a signature consisting of @xmath1-derived operations ) which does not create new recognizable subsets . the basic idea behind the definition of this new signature is the following : when we evaluate a @xmath1-term @xmath47 of the form @xmath638 , then we add edges from each @xmath428-port of @xmath639 , the value of @xmath640 , to each of its @xmath437-ports . it may happen that some edges from a @xmath428-port to a @xmath437-port already exist in @xmath639 . in this case , we do not add a parallel edge since we are dealing with simple graphs . thus the term @xmath47 presents a form of redundancy , since some of its edges may be , in some sense , defined twice . for disjoint sets of port labels @xmath199 and @xmath443 , we denote by @xmath641 the set of @xmath1-derived unary operations defined by terms of the form @xmath642 , where the @xmath643 are of the forms @xmath434 or @xmath644 for @xmath428 in @xmath199 and @xmath437 in @xmath443 . since the operations @xmath434 are idempotent and commute with one another , an operation in @xmath641 is completely described by a subset of @xmath645 . thus @xmath641 is finite , although one can write infinitely many terms specifying its elements . for each element @xmath646 , we let @xmath647 denote the binary operation defined , for @xmath648 and @xmath649 , by @xmath650 . we observe that in the evaluation of a term of the form @xmath651 , the application of @xmath647 does not recreate edges that already exist in @xmath652 , the value of @xmath47 , or in @xmath639 , the value of @xmath640 since the @xmath434 operations forming @xmath647 add edges between the disjoint graphs @xmath652 and @xmath639 ( because @xmath428 and @xmath437 are not port labels of the same argument graphs ) . now the signature @xmath653 consists of the operations @xmath647 as above , the unary qfd operations of the form @xmath454 and @xmath654 as defined in remark [ rename , forget ] , and the constants @xmath447 and @xmath449 as in @xmath1 . we denote by @xmath655 the @xmath653-algebra of graphs with ports . the notation @xmath653 refers to a very similar algebra used by wanke @xcite . [ important example nlc ] we have in fact already encountered @xmath653-operations and @xmath653-derived operations . the @xmath1-derived operation @xmath166 whose existence is proved in proposition [ unary - parallel ] is actually @xmath653-derived . consider indeed the last paragraphs of the proof of that proposition : the operation @xmath166 is obtained by first composing operations of the form @xmath568 and @xmath571 , where the pairs @xmath656 lie in a certain subset of @xmath657 \times [ \ell+1,m]$ ] and the pairs @xmath658 lie in another subset of @xmath659 \times [ k+1,\ell]$ ] , and then composing operations of the form @xmath454 . one can also check that the operations @xmath660 at the end of the proof of theorem [ thm vr equivalence ] are @xmath653-derived . [ prop nlc ] let @xmath199 be a finite subset of @xmath578 and let @xmath2 be a subset of @xmath202 . then @xmath2 is @xmath1-recognizable if and only if @xmath2 is @xmath653-recognizable . the proof is a simple extension of the proof of theorem [ thm vr equivalence ] . since the operations of @xmath653 are @xmath1-derived , every @xmath1-recognizable subset of @xmath426 is @xmath653-recognizable . for the converse , we observe that the proof that ( 1 ) implies ( 3 ) in theorem [ thm vr equivalence ] can be modified to show that an @xmath653-recognizable set of @xmath426 is @xmath275-recognizable . again , we rely on lemma [ fgh - congruence+zeta(t ) ] , but now with @xmath661 , @xmath662 , and @xmath62 , @xmath112 , @xmath160 and @xmath113 as in theorem [ thm vr equivalence ] . in order to justify the fact that the arguments used in the proof of theorem [ thm vr equivalence ] are also valid with these assumptions , we refer to example [ important example nlc ] . indeed this example shows two things : on one hand , the operation @xmath166 in proposition [ unary - parallel ] is in fact @xmath653-derived , so that the first hypothesis of lemma [ fgh - congruence+zeta(t ) ] is satisfied by this new choice of @xmath7 and @xmath111 . on the other hand the finiteness hypothesis of lemma [ fgh - finiteness ] is also satisfied with this new value of @xmath661 . this completes the proof . in certain contexts , and in particular in the definition of the clique - width of a graph ( see remark [ rk clique - width ] below ) , one needs to consider graphs with ports where port labels partition the vertex set . more precisely , for each set of port labels @xmath199 , let @xmath663 be the set of elements of @xmath202 such that each vertex is a port , and no vertex is both a @xmath428-port and a @xmath437-port for @xmath664 . let also @xmath665 . note that @xmath666 is preserved by the operations of the form @xmath209 , @xmath434 and @xmath654 . these operations form the signature @xmath667 , and @xmath666 is a @xmath667-algebra . the operation @xmath434 is written @xmath668 in @xcite . [ rk clique - width ] the _ clique - width _ of a finite graph @xmath652 , denoted by @xmath669 , is defined as the smallest cardinality of a set @xmath199 such that @xmath652 is the value of a ( finite ) @xmath667-term using a set @xmath199 of port labels , see @xcite . for algorithmic applications @xcite , it is useful to have efficient recognition algorithms for classes of graphs of clique - width at most @xmath92 . at the moment we only know that this problem is @xmath670 . it is polynomial for @xmath671 , see @xcite . let @xmath2 be a subset of @xmath663 . then @xmath2 is @xmath667-recognizable if and only if @xmath2 is @xmath1-recognizable . since @xmath667 consists of operations in @xmath1 , every locally finite @xmath1-congruence on @xmath426 induces a locally finite @xmath667-congruence on @xmath666 . in particular , if @xmath2 is @xmath1-recognizable , and hence is saturated by a locally finite @xmath1-congruence on @xmath426 , then @xmath2 is saturated by a locally finite @xmath667-congruence on @xmath666 , and hence @xmath2 is @xmath667-recognizable . to prove the converse , we first introduce the mapping @xmath672 defined as follows . if @xmath648 , then @xmath673 is the graph in @xmath674 with the same set of vertices and the same edge relation as @xmath652 , and such that for each vertex @xmath489 and each @xmath675 , @xmath489 is an @xmath64-port in @xmath673 if and only if @xmath64 is the set of @xmath429 such that @xmath489 is a @xmath428-port in @xmath652 . we say that a port label @xmath428 is _ void in @xmath652 _ if there are no @xmath428-ports in @xmath652 . now let us assume that @xmath2 is @xmath667-recognizable , and let @xmath130 be a locally finite congruence on @xmath666 saturating it . if @xmath676 , we let @xmath677 if @xmath673 and @xmath678 have the same non - void port labels , and @xmath679 . it is immediately verified that @xmath98 is a locally finite equivalence relation . we now verify that @xmath98 is a @xmath1-congruence . if @xmath680 and @xmath649 , it is easily seen that @xmath681 . if @xmath436 , then @xmath682 where @xmath12 is the composition of the operations @xmath683 for each @xmath684 such that @xmath685 and @xmath686 . finally , one can verify that if @xmath687 , then @xmath688 where @xmath132 is the composition of the operations @xmath689 , where @xmath675 , @xmath690 and @xmath691 . it is a routine task to derive from these observations the fact that @xmath98 is a @xmath1-congruence . we now need to verify that @xmath98 saturates @xmath2 . let @xmath692 and @xmath677 . in particular , @xmath693 , so that the non - void port labels of @xmath673 are exactly the sets @xmath448 where @xmath428 is a non - void port label of @xmath652 . since @xmath673 and @xmath678 have the same non - void port labels , @xmath694 is also in @xmath666 . moreover , if @xmath87 is the composition of the operations @xmath695 ( @xmath428 non - void in @xmath652 ) , then @xmath696 and @xmath697 . since @xmath87 is @xmath667-derived , it follows that @xmath698 , and hence @xmath699 . this concludes the proof . recall that we call _ graphs with sources _ the elements of @xmath184 of sort @xmath700 , where @xmath191 and @xmath79 is some finite set of source labels , and that we write @xmath195 for @xmath196 ( see section [ sec rel structures ] ) . the disjoint union @xmath209 and the operations of the form @xmath240 , @xmath244 and @xmath247 ( defined in example [ def hr ops ] ) preserve graphs with sources . we denote by @xmath0 the signature consisting of all these operations , so @xmath701 is an @xmath0-algebra . we note the following properties of @xmath0-recognizability . [ s2hr ] let @xmath79 be a finite set of source labels . every @xmath275-recognizable subset of @xmath196 is @xmath0-recognizable . this is a simple consequence of proposition [ easy recog facts ] and of the observation given above that the operations of @xmath0 are also operations of @xmath275 . note that the class @xmath194 of graphs , defined in section [ sec rel structures ] , is equal to @xmath702 as well as to @xmath703 . thus @xmath1-recognizability and @xmath0-recognizability are properties of subsets of @xmath194 . [ vr2hr ] let @xmath2 be a set of graphs ( a subset of @xmath194 ) . if @xmath2 is @xmath1-recognizable , then it is @xmath0-recognizable . this follows immediately from proposition [ s2hr ] and theorem [ thm vr equivalence ] . intuitively , the @xmath1-operations are more powerful than the @xmath0-operations ( every @xmath0-context - free set of simple graphs is @xmath1-context - free but the converse is not true , courcelle @xcite ) , but the @xmath0-operations are not among the @xmath1-operations , nor are they derived from them . we will see in sections [ proof no kmm ] and [ other finiteness ] sufficient conditions for @xmath0-recognizable sets to be @xmath1-recognizable , and in section [ hr not vr ] , examples of @xmath0-recognizable sets which are not @xmath1-recognizable . we find in the literature a number of variants of the signature @xmath0 or of the algebra @xmath701 . we now discuss these different variants , to verify that they do not introduce artefacts from the point of view of recognizability . let @xmath704 denote the signature on @xmath701 obtained by substituting the parallel composition @xmath341 for @xmath209 ( see section [ sec sts - parallel ] ) . with the same proof as proposition [ prop sts - parallel ] , we get the following result . [ prop hr - parallel ] let @xmath2 be a subset of @xmath701 . then @xmath2 is @xmath0-recognizable if and only if it is @xmath704-recognizable . as in section [ sec source separation ] , we now discuss the class @xmath705 of source separated graphs . the operations of @xmath0 all preserve source separation , except for @xmath247 , but we defined in example [ exemple hrsep ] the operation @xmath706 which does . let @xmath707 be the signature on @xmath705 consisting of @xmath209 and the qfd unary operations of the form @xmath240 , @xmath244 and @xmath360 . [ prop hrsep ] let @xmath2 be a subset of @xmath705 . then @xmath2 is @xmath0-recognizable if and only if it is @xmath707-recognizable . since @xmath707 consists only of @xmath0-derived operations , every @xmath0-recognizable set subset of @xmath705 is also @xmath707-recognizable . the proof of the converse is a variant of the proof of theorem [ thm ssep - recognizable ] . first we note that the type relation @xmath160 ( see section [ sec elementary properties ] ) is also an @xmath0-congruence on @xmath701 . we use the same mapping @xmath87 defined in the proof of theorem [ thm ssep - recognizable ] , that maps a graph with sources @xmath708 to a source - separated graph @xmath709 by splitting sources that were identified in @xmath111 . we refer to that proof for notation used here . if @xmath2 is an @xmath707-recognizable subset of @xmath705 and @xmath130 is a locally finite @xmath707-congruence recognizing it , we define a relation @xmath98 on @xmath701 as follows . if @xmath710 , we say that @xmath376 if @xmath377 and @xmath378 . as in the proof of theorem [ thm ssep - recognizable ] , @xmath98 is easily seen to be a locally finite equivalence relation . it is also easily seen that @xmath98 is preserved under the @xmath707-operation @xmath209 . we now need to verify that if @xmath711 and @xmath132 is one of the unary operations of @xmath707 defined on @xmath712 , then @xmath713 . again , proposition [ ffvv light ] shows that @xmath714 and we want to show that @xmath406 . the graphs @xmath111 and @xmath112 are fixed for the rest of this proof . we write @xmath715 , @xmath716 and @xmath492 for @xmath374 , @xmath375 and @xmath370 . as in the proof of theorem [ thm ssep - recognizable ] , it suffices to construct an @xmath707-derived operation @xmath92 , depending on @xmath132 and @xmath310 , such that @xmath717 and @xmath718 . there is no reason why the operation @xmath92 constructed in the proof of theorem [ thm ssep - recognizable ] should be @xmath707-derived , but the operations @xmath132 considered here , namely @xmath240 , @xmath244 and @xmath360 are simple enough that we can directly construct a suitable @xmath92 in each case . [ [ if - g - srcren_arightarrow - b ] ] if @xmath719 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + then @xmath132 is defined on @xmath195 ( where @xmath238 and @xmath239 ) and its range is @xmath720 . one verifies that @xmath721 is equal to : * @xmath722 if @xmath723 and @xmath724 ; * @xmath725 if @xmath723 and @xmath726 ; * @xmath722 if @xmath727 and @xmath728 for every @xmath729 such that @xmath730 ; * @xmath731 if @xmath727 and @xmath732 . [ [ if - g - srcfg_a ] ] if @xmath733 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + then @xmath132 is defined on @xmath195 ( where @xmath238 ) and its range is @xmath734 . one verifies that @xmath735 is equal to : * @xmath736 if @xmath723 ; * @xmath737 if @xmath727 , @xmath738 and @xmath739 ; * @xmath740 if @xmath727 and @xmath741 . [ [ if - g - fus_arightarrow - b ] ] if @xmath742 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + then @xmath132 is defined on @xmath195 ( where @xmath246 ) and its range is @xmath734 . one verifies that @xmath743 is equal to : * @xmath744 if @xmath723 and @xmath745 ; * @xmath746 if @xmath723 and @xmath747 ; * @xmath748 if @xmath723 and @xmath749 ; * @xmath750 if @xmath727 and @xmath751 ; * @xmath752 if @xmath727 , and @xmath753 , and @xmath754 ; * @xmath755 if @xmath727 , @xmath756 and @xmath757 . this concludes the proof . again with the same proof as for proposition [ prop sts - parallel ] , we can show that the operation @xmath209 can be replaced by @xmath341 in the signature @xmath707 yielding the signature @xmath758 . [ prop hr - sep - parallel ] let @xmath2 be a subset of @xmath705 . then @xmath2 is @xmath707-recognizable if and only if it is @xmath758-recognizable . the equivalence between @xmath758- and @xmath704-recognizability for a set of source - separated graphs a consequence of propositions [ prop hr - parallel ] , [ prop hrsep ] and [ prop hr - sep - parallel ] was already established by courcelle in @xcite for graphs with multi - edges ( see section [ sec simple multiple ] ) . in the same paper , courcelle established the equivalence between @xmath707- and @xmath759-recognizability for several variants @xmath759 of the signature @xmath0 , which we now describe . we refer to @xcite for the proofs . for each finite set @xmath79 of source labels , let @xmath760 be the composition of the operations @xmath761 for each @xmath230 ( in any order ) . let also @xmath762 be the following binary operation on @xmath705 , of type @xmath763 : if @xmath764 , then @xmath765 : @xmath766 is obtained by first taking the parallel composition @xmath767 , and then forgetting all source labels . let @xmath768 be the signature on @xmath705 , which consists only of the @xmath762 operations . let @xmath769 be the derived signature of @xmath704 , which consists of the operations @xmath760 and @xmath341 . let @xmath770 be the subsignature of @xmath704 , which consists of the operations @xmath771 and @xmath341 . let @xmath772 be the subsignature of @xmath758 , which consists of the operations @xmath341 and those operations @xmath771 which preserve source separation . the following result is a compilation of ( * ? ? ? * section 4 ) . [ prop : other variants ] if @xmath773 , then @xmath2 is @xmath0-recognizable if and only if @xmath2 is @xmath770-recognizable . if @xmath774 , then @xmath2 is @xmath707-recognizable if and only if @xmath2 is @xmath772-recognizable . if @xmath775 , the following are equivalent : * @xmath2 is @xmath0-recognizable ; * @xmath2 is @xmath768-recognizable ; * @xmath2 is @xmath769-recognizable . the notation @xmath768 refers to the notion of _ fully cutset - regular _ sets of graphs , introduced by abrahamson and fellows @xcite . full cutset - regularity is equivalent to @xmath768-recognizability . in @xcite , courcelle also shows a number of closure properties of the class of @xmath707-recognizable sets of source - separated graphs with sources . in particular , it is shown that this class contains all singletons and it is closed under the operations of @xmath707 ( * ? ? ? * section 6 ) . finally courcelle shows the following result ( * ? ? ? * theorem 6.7 ) . . then @xmath2 is @xmath0-recognizable if and only if @xmath777 is @xmath0-recognizable . we saw that a @xmath1-recognizable set of graphs is always @xmath0-recognizable ( corollary [ vr2hr ] ) . the converse does not hold in general , as we discuss in section [ hr not vr ] . we first explore structural conditions on graphs , which are sufficient to guarantee that an @xmath0-recognizable set of graphs is also @xmath1-recognizable . let @xmath778 be the directed complete bipartite graph with @xmath4 vertices . a directed graph @xmath779 is _ without @xmath778 _ if it has no subgraph isomorphic to @xmath778 . the main result in this section is the following . [ no kmm ] let @xmath5 be an integer . an @xmath0-recognizable set of graphs without @xmath778 is @xmath1-recognizable . this theorem is proved in section [ proof no kmm ] , and some of its corollaries are discussed in section [ other finiteness ] . note that results similar to corollary [ vr2hr ] and theorem [ no kmm ] hold for @xmath1- and @xmath0-equational sets of graphs . as explained in the introduction , such sets are exactly the context - free sets of graphs , formally specified in terms of recursive sets of equations using the operations of @xmath1 and @xmath0 respectively . specifically , the following results are known to hold : * every @xmath0-equational set of simple directed graphs is @xmath1-equational ( courcelle @xcite ) ; * if a @xmath1-equational set of directed graphs is without @xmath780 for some @xmath5 , then it is @xmath0-equational ( by the main theorem in courcelle @xcite and lemma [ no kpp ] below ) . thus the same combinatorial condition is sufficient to guarantee the equivalence between @xmath1- and @xmath0-recognizability , as well as between @xmath1- and @xmath0-equationality . a further similar result concerning monadic second - order definability and using a stronger combinatorial property will be discussed in section [ sparse mso ] . we first record the following observation . [ lemma yyy ] let @xmath652 be a directed graph and let @xmath781 be two vertices of @xmath652 that are not adjacent , and such that there is no vertex @xmath782 such that both @xmath783 and @xmath784 ( resp . both @xmath785 and @xmath786 ) are edges . let @xmath694 be obtained from @xmath652 by identifying @xmath593 and @xmath498 . if @xmath652 contains @xmath787 as a subgraph , then so does @xmath694 . let @xmath788 be a subgraph of @xmath652 isomorphic to @xmath787 . from the hypothesis , the vertices @xmath593 and @xmath498 are not both in @xmath788 . it follows that @xmath788 is still isomorphic to a subgraph of @xmath694 . the proof of theorem [ no kmm ] will proceed as follows . we consider an @xmath0-recognizable set @xmath2 of finite graphs without @xmath778 and we denote by @xmath573 the largest integer such that @xmath787 is a subgraph of a graph in @xmath2 . such an integer exists by hypothesis . since we are talking about source - less graphs , the set @xmath2 is @xmath707-recognizable by proposition [ prop hrsep ] , and we consider a locally finite @xmath707-congruence @xmath130 saturating @xmath2 . we will define a locally finite @xmath653-congruence @xmath98 on @xmath426 that also saturates @xmath2 . by proposition [ prop nlc ] , this suffices to show that @xmath2 is @xmath1-recognizable . the definition of @xmath98 makes use of the notion of expansion of a graph , defined below . note that the following definitions depend on the integer @xmath573 , even though terminology and notation do not make this dependence explicit . [ [ small - and - large - port - labels - and - formulas ] ] small and large port labels and formulas + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath680 be a graph with ports . if @xmath429 , we denote by @xmath789 the set of @xmath428-ports of @xmath652 . we say that a port label @xmath428 is _ void in @xmath652 _ if @xmath789 is empty , we say that @xmath428 is _ small in @xmath652 _ if @xmath790 and that @xmath428 is _ large in @xmath652 _ if @xmath791 . observe that if the port labels @xmath428 and @xmath437 are both large in @xmath652 , then @xmath792 contains @xmath793 as a subgraph . moreover , if @xmath428 is large in @xmath652 , if @xmath794 are small in @xmath652 , let @xmath795 for @xmath796 , let @xmath797 . if @xmath694 does not contain @xmath793 , then we must have @xmath798 . if @xmath652 already contains edges from the @xmath428-ports to other vertices , then @xmath799 . the notion of expansion below will make it possible to handle this sort of complicated situation ( see example [ ex : expansion ] below ) . let us say that a closed first - order formula is _ small _ if it has quantifier - depth at most @xmath800 . note that the existence of a subgraph isomorphic to @xmath793 can be expressed by a first - order formula of quantifier - depth @xmath800 . [ [ expansions ] ] expansions + + + + + + + + + + we will define supergraphs of @xmath680 called _ expansions _ , that contain information relevant to the distribution of small and large port labels , and where ports are represented by sources . furthermore , it will be possible to simulate an @xmath653-operation on @xmath652 that does not create @xmath793 subgraphs by @xmath0-operations on expansions of @xmath652 . these expansions will then be used to transform the @xmath707-congruence @xmath130 into an @xmath653-congruence @xmath98 . furthermore , we will define @xmath98 in such a way that two equivalent graphs satisfy the same small first - order formulas . we now give formal definitions . for each port label @xmath428 , we define a set @xmath801 of source labels , @xmath802 if @xmath199 is a set of port labels , @xmath803 denotes the union of the @xmath801 , for @xmath428 in @xmath199 . let @xmath648 be a graph with ports , let @xmath804 , and let @xmath805 be a graph in @xmath197 . we say that @xmath805 is an _ expansion _ of @xmath652 if the following conditions hold : * @xmath806 has no subgraph isomorphic to @xmath793 . * except for the labeling of ports and sources , @xmath652 is a subgraph of @xmath805 . the sources of @xmath806 , and its vertices and edges not in @xmath652 , are specified by conditions ( 3 ) and ( 4 ) . * if @xmath428 is small in @xmath652 , then each @xmath428-port of @xmath652 is an @xmath807-source of @xmath805 for some integer @xmath808 . different @xmath428-ports are of course labelled by different source labels . there are no @xmath809- or @xmath810-sources . * if @xmath428 is large in @xmath652 , then there may be vertices of @xmath806 that are not in @xmath652 , with source labels of the form @xmath811 or @xmath812 for some @xmath813 . moreover , there is an edge in @xmath806 from each vertex of @xmath789 to each @xmath811-source , and from each @xmath812-source to each vertex in @xmath789 . there are no @xmath814-sources . in particular , @xmath652 may have several different expansions , but it has only a finite number of expansions ( up to isomorphism ) . this number is bounded by a function depending on @xmath573 and the cardinality of @xmath199 . indeed , for each small port label @xmath428 , there is only a bounded number of ways to make @xmath428-ports into @xmath807-sources ( see ( 3 ) ) , and for each large port label @xmath428 , there is a bounded number of ways to create @xmath811- and @xmath812-sources ( see ( 4 ) ) . [ ex : expansion ] let @xmath815 , and let @xmath652 be a graph with port labels @xmath816 . suppose that @xmath652 has 4 @xmath428-ports , 2 @xmath437-ports and 1 @xmath216-port , so that @xmath428 is large , and @xmath437 , @xmath216 are small in @xmath652 , see figure [ fig : example expansion ] . then in any expansion of @xmath652 , every @xmath437- and @xmath216-port will be a source , say labeled by @xmath817 , @xmath818 and @xmath819 ( there is only one @xmath820-source , but it is not required that these sources should be labeled with consecutive numbers starting at 1 ) . ( 70,55)(0,-50 ) ( 10,-47)(60,35 ) ( 2,-45 ) ( 12,-45 ) ( n0)(12.0,-12.0)@xmath428 ( n1)(22.0,-12.0)@xmath428 ( n2)(32.0,-12.0)@xmath428 ( n3)(42.0,-12.0)@xmath428 ( n4)(70.0,-17.0)@xmath437 ( n5)(70.0,-23.0)@xmath437 ( n6)(70.0,-35.0)@xmath216 ( n4)(70.0,-17.0)@xmath817 ( n5)(70.0,-23.0)@xmath818 ( n6)(70.0,-35.0)@xmath819 ( n7)(7.0,0.0)@xmath821 ( n8)(27.0,0.0)@xmath822 ( n9)(43.0,0.0)@xmath823 ( n10)(37.0,-30.0)@xmath593 ( n10,n0 ) ( n10,n1 ) ( n10,n2 ) ( n10,n3 ) ( n7,n0 ) ( n7,n1 ) ( n7,n2 ) ( n7,n3 ) ( n0,n8 ) ( n1,n8 ) ( n2,n8 ) ( n3,n8 ) ( n0,n9 ) ( n1,n9 ) ( n2,n9 ) ( n3,n9 ) moreover , an expansion of @xmath652 may have up to two new vertices that are @xmath809-sources , and at most one @xmath810-source . say , an expansion @xmath694 could have new vertices as @xmath822- and @xmath823-sources , with edges from each of the 4 @xmath428-ports to each @xmath809-source ; and it could have a new vertex as a , say , @xmath821-source , with edges from that vertex to each of the @xmath428-ports . note that if @xmath652 has a vertex @xmath593 with an edge from @xmath593 to at least 3 @xmath428-sources , then an expansion can not have 2 @xmath810-sources : otherwise it would contain a copy of @xmath824 , which is not allowed for an expansion . it is not always the case that @xmath652 is determined by each of its expansions @xmath806 . if @xmath428 is large in @xmath652 but @xmath806 has no @xmath811- or @xmath812-sources , then it is not possible to determine which of its vertices are @xmath428-ports . [ [ construction - of - an - nlc - congruence - from - an - hr_sep - congruence ] ] construction of an @xmath653-congruence from an @xmath707-congruence + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath130 be a locally finite @xmath707-congruence saturating @xmath2 . we define a relation @xmath98 on @xmath426 as follows . for @xmath652 and @xmath639 in @xmath202 we let @xmath825 if and only if * either @xmath652 and @xmath639 both contain @xmath793 as a subgraph , or neither does and in that case , the following two conditions hold : * @xmath652 and @xmath639 satisfy the same small first - order formulas ( _ i.e. _ , with quantifier - depth at most @xmath800 ) on graphs with ports . * for every expansion @xmath805 of @xmath652 , there exists an expansion @xmath826 of @xmath639 such that @xmath827 and @xmath805 and @xmath828 satisfy the same small first - order formulas on graphs with sources ( we say that @xmath806 and @xmath828 are _ equivalent expansions _ ) ; and conversely , for every expansion @xmath828 of @xmath639 there exists an expansion @xmath805 of @xmath652 equivalent to @xmath828 . note that condition ( b ) implies that @xmath652 and @xmath639 have the same void , small and large port labels , and condition ( c ) implies that @xmath805 and @xmath826 have the same source labels . the relation @xmath98 is clearly an equivalence relation on each set @xmath202 . it has finitely many classes on each @xmath202 since a finite graph has a uniformly bounded number of expansions ( up to isomorphism ) , the @xmath707-congruence @xmath130 is locally finite , and there are finitely many first - order formulas of each quantifier - depth on graphs with sources in a subset of @xmath803 . now a graph without ports and without @xmath829 has a unique expansion : itself . it follows that , for graphs without ports and without @xmath829 , the equivalences @xmath130and @xmath98 coincide . in particular , @xmath98 saturates @xmath2 since @xmath130 does . it remains to prove that @xmath98 is an @xmath653-congruence . recall that the signature @xmath653 consists of the operations of the form @xmath454 , @xmath654 and @xmath647 . [ [ the - port - forgetting - operation ] ] the port forgetting operation + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we first consider the operation @xmath454 . we consider @xmath830 with @xmath831 and we want to prove that @xmath832 , where @xmath833 and @xmath834 . first of all , the underlying graphs of @xmath652 and @xmath694 ( resp . @xmath639 and @xmath835 ) are identical , so that @xmath652 and @xmath639 contain @xmath793 if and only if so do @xmath694 and @xmath835 . if this is the case , then @xmath831 and @xmath832 . we now exclude this case and assume that @xmath652 and @xmath639 are without @xmath793 . note also that if @xmath428 is void in @xmath652 , then it is in @xmath639 as well , and we have @xmath836 , @xmath837 , so that @xmath832 . we now assume that @xmath428 is not void in @xmath652 . it is an immediate consequence of theorem [ ffvv ] that @xmath694 and @xmath835 satisfy the same small first - order formulas on graphs with ports , so condition ( b ) is verified . we now consider condition ( c ) . let @xmath838 be an expansion of @xmath694 . we will show that there exists an expansion @xmath805 of @xmath652 and a unary @xmath707-term @xmath47 such that @xmath839 = @xmath840 . since @xmath841 , there exists an equivalent expansion @xmath842 of @xmath639 , and @xmath843 will be the desired expansion of @xmath835 . using the fact that @xmath130 is an @xmath707-congruence and theorem [ ffvv ] , we will have @xmath832 as expected . if @xmath428 is large in @xmath652 , the situation is particularly simple : @xmath838 is also an expansion of @xmath652 , so we can choose @xmath47 to represent the identity . if @xmath828 is an expansion of @xmath639 , equivalent to @xmath838 , then @xmath828 does not use source labels of the form @xmath807 , @xmath811 or @xmath812 , so @xmath828 is also an expansion of @xmath835 . if @xmath428 is small in @xmath652 , let @xmath806 be a graph with source obtained from @xmath838 by letting each @xmath428-port of @xmath652 be an @xmath807-source ( where distinct source labels are used for distinct @xmath428-ports ) . then @xmath806 is an expansion of @xmath652 , and @xmath844 where @xmath47 is the composition of the operations @xmath845 ( @xmath846 ) . using the definition of @xmath98 , there exists an expansion @xmath828 of @xmath639 which is equivalent to @xmath806 , and we only need to verify that @xmath847 is an expansion of @xmath835 . the only point to check here is the fact that @xmath835 is a subgraph of @xmath848 : this follows from the facts that @xmath652 is a subgraph of @xmath806 and the operations @xmath47 and @xmath454 do not change the underlying graph structures . [ [ the - renaming - operation ] ] the renaming operation + + + + + + + + + + + + + + + + + + + + + + we now consider the operation @xmath654 . let @xmath830 with @xmath831 : as with the port forgetting operation , we want to prove that @xmath832 where @xmath849 and @xmath850 . as above , we can reduce the proof to the case where neither @xmath652 nor @xmath639 contains @xmath793 , and where @xmath428 is not void in @xmath652 ( if @xmath428 is void in @xmath652 , then @xmath836 and @xmath837 ) . moreover , condition ( b ) follows from theorem [ ffvv ] . we consider condition ( c ) , following the same strategy as above . let @xmath838 be an expansion of @xmath694 . if @xmath437 is void in @xmath652 , then the transformation @xmath654 is a reversible renaming , that is , @xmath851 . moreover , if @xmath47 is the composition of the operations of the form @xmath852 , @xmath853 and @xmath854 , and if @xmath640 is the composition of the operations @xmath855 , @xmath856 and @xmath857 , then @xmath858 is an expansion of @xmath652 , @xmath844 . moreover , if @xmath828 is an expansion of @xmath639 , equivalent to @xmath806 , then @xmath847 is an expansion of @xmath835 . we now assume that @xmath437 is not void in @xmath652 . we need to consider several cases . * case 1. @xmath428 and @xmath437 are both large in @xmath652 . then @xmath428 is void and @xmath437 is large in @xmath694 . in order to build the desired @xmath805 , we split each @xmath859-source of @xmath839 into an @xmath811-source and an @xmath859-source . the @xmath811-source is linked by incoming edges to all @xmath428-ports of @xmath652 , and the @xmath859-source is linked similarly to all @xmath437-ports . in the same fashion , we split each @xmath860-source of @xmath839 into an @xmath812-source and an @xmath860-source linked by ougoing edges to all @xmath428-ports of @xmath652 and to all @xmath437-ports respectively . the term @xmath47 such that @xmath844 is the composition of the operations @xmath861 and @xmath862 . the graph @xmath805 does not contain @xmath793 , since @xmath839 does not ( by lemma [ lemma yyy ] ) . hence @xmath805 is an expansion of @xmath652 . let now @xmath842 be an expansion of @xmath639 equivalent to @xmath805 , and let @xmath863 . it is easily verified that @xmath848 is an expansion of @xmath835 , and as above , it follows that @xmath864 . case 2. @xmath428 is small and @xmath437 is large in @xmath652 . in order to build @xmath806 from @xmath839 , we make the @xmath428-ports of @xmath652 into @xmath807-sources , we delete the edges between the @xmath859- and the @xmath860-sources and the @xmath428-ports of @xmath652 . the term @xmath47 which must do the opposite ( that is , construct @xmath839 from @xmath805 ) is a composition of source forgetting operations and of additions of new edges . more precisely , for each @xmath865 such that @xmath807 and @xmath866 are source labels in @xmath806 , we use the operation @xmath867 , where @xmath868 is the 2-vertex , 2-source , 1-edge graph , followed by the operations @xmath869 and @xmath870 . we then apply similar operations to create edges from the @xmath871- to the @xmath807-sources . and we finally apply the operations @xmath845 . the graph @xmath806 is a subgraph of @xmath838 ( up to source labels ) , so @xmath806 does not contain @xmath793 , and hence it is an expansion of @xmath652 . the proof continues as in the previous case . case 3. @xmath437 is small and @xmath428 is large in @xmath652 . to build @xmath806 from @xmath838 , we make the @xmath437-ports of @xmath652 into @xmath872-sources , we delete the edges between the @xmath811-sources or the @xmath812-sources and the @xmath437-ports of @xmath652 . in addition we rename each @xmath811-source to an @xmath859-source , and each @xmath812-source to an @xmath860-source . we can use the same reasoning as in case 2 to conclude in this case . case 4. @xmath428 and @xmath437 are small in @xmath652 , and @xmath873 . to build @xmath806 from @xmath838 , we rename @xmath872 into @xmath807 whenever the @xmath872-source of @xmath839 is a @xmath428-port in @xmath652 . the term @xmath47 which does the opposite is a composition of source renamings . the graph @xmath805 does not contain @xmath793 , otherwise @xmath839 would do , since @xmath805 is equal to @xmath839 up to source labels , and hence @xmath806 is an expansion of @xmath652 . the other parts of the proof are the same . case 5. @xmath428 and @xmath437 are small in @xmath652 , and @xmath874 . to build @xmath806 from @xmath838 , we make the @xmath428-ports ( resp . @xmath437-ports ) of @xmath652 into @xmath807-sources ( resp . @xmath872-sources ) , we delete the edges between the @xmath859- and @xmath860-sources and the @xmath428- and @xmath437-ports of @xmath652 , and we delete the @xmath859- and @xmath860-sources . the term @xmath47 which does the opposite is a composition of additions of new edges and of @xmath875 operations , as in case 2 , see figure [ fig : figure 2 ] . the graph @xmath805 does not contain @xmath876 , otherwise @xmath839 would too , since @xmath805 is a subgraph of @xmath839 ( up to source labels ) , and hence @xmath806 is an expansion of @xmath652 . the proof continues as in the previous cases . ( 88,75)(0,-75 ) ( n0)(20.0,-16.0)@xmath652 ( n1)(66.0,-16.0)@xmath694 ( n2)(7.0,-10.0)@xmath428 ( n3)(16.0,-10.0)@xmath428 ( n4)(25.0,-10.0)@xmath437 ( n5)(34.0,-10.0)@xmath437 ( n6)(52.0,-10.0)@xmath437 ( n7)(61.0,-10.0)@xmath437 ( n8)(70.0,-10.0)@xmath437 ( n9)(79.0,-10.0)@xmath437 ( n10)(20.0,-48.0)@xmath806 ( n11)(66.0,-48.0)@xmath838 ( n12)(7.0,-42.0)@xmath877 ( n13)(16.0,-42.0)@xmath878 ( n14)(25.0,-42.0)@xmath879 ( n15)(34.0,-42.0)@xmath880 ( n16)(52.0,-42.0 ) ( n17)(61.0,-42.0 ) ( n18)(70.0,-42.0 ) ( n19)(79.0,-42.0 ) ( n20)(55.0,-28.0)@xmath881 ( n21)(73.0,-28.0)@xmath882 ( n20,n16 ) ( n20,n17 ) ( n20,n18 ) ( n20,n19 ) ( n16,n21 ) ( n17,n21 ) ( n18,n21 ) ( n19,n21 ) ( n22)(52.0,-74.0)@xmath877 ( n23)(61.0,-74.0)@xmath878 ( n24)(70.0,-74.0)@xmath879 ( n25)(79.0,-74.0)@xmath880 ( n26)(55.0,-60.0)@xmath881 ( n27)(73.0,-60.0)@xmath882 ( n26,n22 ) ( n26,n23 ) ( n26,n24 ) ( n26,n25 ) ( n22,n27 ) ( n23,n27 ) ( n24,n27 ) ( n25,n27 ) ( 38,-67 ) this concludes the proof that @xmath831 implies @xmath883 . [ [ the - operation - otimes_j ] ] the operation @xmath647 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we now consider the operation @xmath647 where @xmath884 , @xmath199 and @xmath443 are disjoint . let @xmath825 in @xmath202 , @xmath885 in @xmath457 , @xmath886 and @xmath887 . we want to prove that @xmath864 . we first consider the very special case where @xmath888 , and the operation @xmath647 is simply the disjoint union . then @xmath694 contains @xmath793 if and only if @xmath652 or @xmath788 does , if and only if @xmath639 or @xmath889 does , if and only if @xmath835 does . asuming that @xmath694 does not contain @xmath793 , an application of theorem [ ffvv ] ensures , as for the operations of port forgetting or renaming that @xmath694 and @xmath835 satisfy the same small first - order formulas . we now consider an expansion @xmath838 of @xmath694 . it is necessarily of the form @xmath890 where @xmath805 and @xmath891 are expansions of @xmath652 and @xmath788 respectively . then there exist expansions @xmath826 and @xmath892 of @xmath639 and @xmath889 respectively , which are equivalent to @xmath805 and @xmath891 . one then verifies that @xmath893 is an expansion of @xmath835 , which is equivalent to @xmath839 . next we assume that @xmath894 is a singleton , @xmath895 , that is , @xmath896 with @xmath429 and @xmath408 . since @xmath652 and @xmath639 on one hand , and @xmath788 and @xmath889 on the other satisfy the same small first - order formulas , theorem [ ffvv ] shows that @xmath897 contains @xmath898 if and only if @xmath899 does . assume now this is not the case and consider an expansion @xmath839 of @xmath694 . again there are several cases . note that @xmath428 and @xmath437 can not both be large in @xmath652 and @xmath788 respectively . we claim that @xmath839 can defined as @xmath900 where @xmath47 is an @xmath707-term , @xmath805 is an expansion of @xmath652 and @xmath891 is an expansion of @xmath788 . as for the other operations , we consider expansions @xmath828 and @xmath901 of @xmath639 and @xmath889 , equivalent to @xmath806 and @xmath902 . although it is a bit tedious , we verify formally that @xmath903 is an expansion of @xmath835 . it follows that @xmath848 is equivalent to @xmath838 , and hence @xmath864 . case 1. @xmath428 is large in @xmath652 and @xmath437 is small in @xmath788 . then @xmath694 has edges from all @xmath428-ports of @xmath652 to all @xmath437-ports of @xmath788 , which are actually @xmath872-sources in @xmath839 . for each of these @xmath872-sources , say @xmath593 , we create a new vertex @xmath904 , and each edge coming from @xmath652 to @xmath593 is redirected towards @xmath904 . we make @xmath904 into an @xmath809-source ( for some appropriate @xmath614 ) of the expansion @xmath805 of @xmath652 we are constructing . the desired expansion @xmath891 of @xmath788 is just the subgraph of @xmath839 induced by the set of vertices of @xmath788 . and @xmath805 consists of the subgraph of @xmath839 induced by the vertices of @xmath652 together with @xmath904 and all these redirected edges . then the @xmath707-term @xmath47 needs only to fuse in @xmath905 the above described @xmath809-sources with the corresponding @xmath906-sources . this can be done by a combination of the operation @xmath209 and those of the form @xmath907 . the only point to check is that @xmath805 does not contain @xmath793 . we can apply lemma [ lemma yyy ] because @xmath839 is obtained from @xmath908 by fusions of pairs of vertices which are not adjacent and have no incoming edges with the same source ( because @xmath652 and @xmath788 are disjoint ) and no outgoing edge at all . then there exist expansions @xmath826 and @xmath892 of @xmath639 and @xmath889 respectively , equivalent to @xmath805 and @xmath891 . by letting @xmath909 , we get the desired expansion of @xmath835 , equivalent to @xmath839 . this case is illustrated in figure [ fig : figure 3 ] , where @xmath910 and @xmath88 is the constructed expansion of @xmath911 . ( 90,80)(0,-80 ) ( n0)(12.0,-24.0)@xmath652 ( n1)(34.0,-16.0)@xmath788 ( n2)(62.0,-24.0)@xmath652 ( n3)(84.0,-16.0)@xmath788 ( 53,-35)@xmath912 ( n4)(5.0,-18.0)@xmath428 ( n5)(12.0,-18.0)@xmath428 ( n6)(19.0,-18.0)@xmath428 ( n7)(28.0,-6.0)@xmath437 ( n8)(28.0,-12.0)@xmath437 ( n9)(55.0,-18.0)@xmath428 ( n10)(62.0,-18.0)@xmath428 ( n11)(69.0,-18.0)@xmath428 ( n12)(78.0,-6.0)@xmath437 ( n13)(78.0,-12.0)@xmath437 ( n9,n12 ) ( n9,n13 ) ( n10,n12 ) ( n10,n13 ) ( n11,n12 ) ( n11,n13 ) ( n14)(12.0,-72.0)@xmath806 ( n15)(34.0,-64.0)@xmath902 ( n16)(62.0,-72.0 ) ( n17)(84.0,-64.0 ) ( 70,-83)@xmath88 ( n18)(5.0,-66.0 ) ( n19)(12.0,-66.0 ) ( n20)(19.0,-66.0 ) ( n21)(2.0,-50.0)@xmath913 ( n22)(12.0,-50.0)@xmath914 ( n23)(22.0,-50.0)@xmath915 ( n24)(28.0,-54.0)@xmath916 ( n25)(28.0,-60.0)@xmath917 ( n18,n21 ) ( n18,n22 ) ( n18,n23 ) ( n19,n21 ) ( n19,n22 ) ( n19,n23 ) ( n20,n21 ) ( n20,n22 ) ( n20,n23 ) ( n26)(55.0,-66.0 ) ( n27)(62.0,-66.0 ) ( n28)(69.0,-66.0 ) ( n29)(52.0,-50.0 ) ( n30)(78.0,-54.0 ) ( n31)(78.0,-60.0 ) ( n26,n30 ) ( n26,n31 ) ( n27,n30 ) ( n27,n31 ) ( n28,n30 ) ( n28,n31 ) ( n26,n29 ) ( n27,n29 ) ( n28,n29 ) case 2. @xmath428 is small in @xmath652 and @xmath437 is large in @xmath788 . it is fully similar to the first case , creating new @xmath871-sources instead of @xmath809-sources . we omit the details . case 3. @xmath428 is small in @xmath652 and @xmath437 is small in @xmath788 . let @xmath806 be the subgraph with sources of @xmath838 consisting of the vertices of @xmath652 , and let @xmath902 be defined similarly in terms of @xmath788 . then @xmath839 is obtained from @xmath908 by the addition of edges from each @xmath807-source of @xmath805 to each @xmath918-source of @xmath891 , which can be done by an @xmath707-term ( see case 2 of the discussion of the renaming operation ) . since @xmath805 and @xmath891 are subgraphs of @xmath839 , they can not contain @xmath793 and hence , they are in fact expansions of @xmath652 and @xmath788 as desired . the proof continues as above . case 4. @xmath428 is void in @xmath652 or @xmath437 is void in @xmath788 . then @xmath919 acts as the identity on @xmath920 , so @xmath647 acts as @xmath209 on @xmath921 and we are back to a previously studied case . recall that if @xmath428 ( resp . @xmath437 ) is void in @xmath652 ( resp . @xmath788 ) , then it is void in every @xmath98-equivalent graph with source . this concludes the study of the case where @xmath894 is a singleton in @xmath922 . the case where @xmath894 is a singleton in @xmath923 is of course similar . the proof is actually the same in the general case where @xmath894 is not a singleton . we need only do the same constructions for all elements @xmath924 in @xmath894 . the only possible difficulty could arise from the use of lemma [ lemma yyy ] to verify that the graphs @xmath805 and @xmath891 obtained from @xmath839 by the creation of vertices ( like @xmath904 in case 1 above ) and the redirection of edges do not contain @xmath793 , and hence are expansions . thus let us consider the transformation of @xmath925 into @xmath839 . it consists in a sequence of fusions of pairs of vertices . whenever we fuse an @xmath811-source of @xmath805 , say @xmath593 , with an @xmath918-source of @xmath891 , say @xmath498 , we must verify that the fusions performed previously keep the hypothesis of lemma [ lemma yyy ] valid . it is clear that @xmath593 and @xmath498 are not adjacent , since @xmath593 is adjacent with vertices of @xmath652 only . because of previous fusions , there may exist an edge from some @xmath782 in @xmath652 to @xmath498 . however , this edge comes from a previously applied operation @xmath926 with @xmath927 . it follows that there is no edge from @xmath782 to @xmath593 . an analogous argument also applies to fusions between an @xmath812-source of @xmath652 and an @xmath918-source of @xmath788 , and also when we exchange the roles of @xmath652 and @xmath788 . hence , finally , we can apply lemma [ lemma yyy ] to deduce that @xmath805 and @xmath891 do not contain @xmath793 because @xmath839 does not . hence , they are expansions of @xmath652 and @xmath788 , as we needed to check . this concludes the proof of theorem [ no kmm ] . we now consider some consequences of theorem [ no kmm ] . let @xmath3 be the undirected complete bipartite graph with @xmath4 vertices , that is , @xmath3 is the undirected graph underlying @xmath780 . we say that a ( directed ) graph is _ without @xmath3 _ if its undirected underlying graph has no subgraph isomorphic to @xmath3 . we say that a graph @xmath652 is uniformly _ @xmath92-sparse _ if @xmath928 for every finite subgraph @xmath694 of @xmath652 , where @xmath929 and @xmath930 are the sets of vertices and edges of @xmath694 . a set of graphs is _ uniformly @xmath92-sparse _ if each of its elements is . [ prop k - sparse ] let @xmath775 be a set of graphs , satisfying one of the following properties : @xmath2 is without @xmath780 for some @xmath5 * @xmath2 is without @xmath3 for some @xmath5 * @xmath2 consists only of planar graphs * @xmath2 is uniformly @xmath92-sparse for some @xmath92 * @xmath2 consists only of graphs of tree - width at most @xmath92 for some @xmath92 . then @xmath2 is @xmath0-recognizable if and only if @xmath2 is @xmath1-recognizable . by corollary [ vr2hr ] , it is always the case that a @xmath1-recognizable set of graphs is @xmath0-recognizable . if @xmath2 is without @xmath780 for some @xmath5 , the converse implication was proved in theorem [ no kmm ] . lemma [ no kpp ] below shows that @xmath2 is without @xmath931 for some @xmath428 if and only if it is without @xmath780 for some @xmath5 . it is well - known that planar graphs are without @xmath932 ( planarity is a property of the underlying undirected graph , and @xmath932 is the undirected graph underlying @xmath824 ) . it follows that planar graphs are also without @xmath824 , and the result follows from theorem [ no kmm ] . it is easily seen that @xmath933 is not @xmath92-sparse . so if @xmath2 is uniformly @xmath92-sparse , then it is without @xmath933 . finally , it is known that graphs of tree - width at most @xmath92 are uniformly @xmath934-sparse ( see for instance @xcite ) , which yields the last assertion . [ no kpp ] let @xmath428 be an integer . there exists an integer @xmath5 such that a directed graph without @xmath935 , is without @xmath3 . we use the particular case of ramsey s theorem for bipartite graphs , given as theorem 1 in @xcite . it states that for each @xmath428 , there exists an integer @xmath5 such that , if the edges of @xmath3 are partitioned into two sets @xmath77 and @xmath394 , then either @xmath77 or @xmath394 contains the edges of a subgraph isomorphic to @xmath931 . so let us assume that @xmath936 , where @xmath421 and @xmath937 are disjoint sets of @xmath5 elements and there is an edge between @xmath108 and @xmath511 ( in one or both directions ) for each @xmath938 . let @xmath77 be the set of pairs @xmath938 such that the edge is from @xmath108 to @xmath511 , and let @xmath939 . then there exist sets @xmath940 and @xmath941 , with cardinality @xmath428 , such that @xmath942 or @xmath943 . in either case , we get a subgraph of @xmath652 isomorphic to @xmath935 . note hat a quick and direct proof can be given with @xmath944 , but we do not know the minimal @xmath5 yielding the result . the statement relative to bounded tree - width sets of graphs in proposition [ prop k - sparse ] is also a consequence ( in the case of finite graphs ) of lapoire s result @xcite , which states that , in a graph of tree - width at most @xmath92 , one can construct a width-@xmath92 tree - decomposition by monadic second - order ( mso ) formulas . this can be used to show that every @xmath0-recognizable set of graphs of bounded tree - width is definable in counting monadic second - order ( cmso ) logic , using edge set quantifications . courcelle showed @xcite that , for finite graphs of bounded tree - width , edge set quantifications can be replaced by vertex set quantifications . the considered set is therefore definable in cmso logic with vertex set quantifications only , and hence is @xmath1-recognizable by another of courcelle s results @xcite . it is proved in @xcite that every set of square grids is @xmath0-recognizable . it follows from theorem [ no kmm ] that every such set is also @xmath1-recognizable . hence , there are uncountably many @xmath1-recognizable sets of graphs , so we can not hope for an automata - theoretic or a logical characterization of @xmath1-recognizability in contrast with the situation prevailing for words , trees and some special classes of graphs , see @xcite . the aim of this short section is to establish the existence of @xmath0-recognizable sets which are not @xmath1-recognizable . we first establish a lemma . [ cliques hr - rec ] every set of cliques ( of the form @xmath945 , @xmath946 ) is @xmath0-recognizable . let @xmath2 be a set of undirected cliques ( recall that an undirected graph is a graph where the edge relation is symmetric ) . we provide a locally finite @xmath768-congruence on @xmath705 which saturates @xmath2 ( see section [ sec : other variants ] ) . by proposition [ prop : other variants ] , this establishes that @xmath2 is @xmath0-recognizable . for each finite set @xmath79 of source labels , let @xmath947 be the set of graphs in @xmath197 having at least one internal vertex ( i.e. , a vertex which is not a source ) , and let @xmath948 be the set of graphs in @xmath197 , in which every vertex is a source . in particular , @xmath948 is finite . let @xmath130 be the following equivalence relation on @xmath705 . we use the operation @xmath762 , as in section [ sec : other variants ] . if @xmath949 , we let @xmath950 if and only if either @xmath951 , or @xmath952 and for every @xmath953 , @xmath954 iff @xmath955 . note that for each @xmath79 , there are only finitely many @xmath130-classes in @xmath197 , namely at most @xmath956 , where @xmath428 is the cardinality of @xmath948 . moreover , @xmath130 saturates @xmath2 . indeed , suppose that @xmath957 , @xmath958 and @xmath692 . let @xmath694 be the graph in @xmath197 consisting of distinct @xmath180-sources ( @xmath230 ) and no edges . then we have @xmath959 and @xmath960 . it follows from the definition of @xmath130 that @xmath961 . finally , we check that @xmath130 is a @xmath768-congruence . let @xmath962 , with @xmath958 and @xmath963 : we want to show that @xmath964 . we observe that if both @xmath652 and @xmath694 have internal vertices , then @xmath766 is not a clique ( by definition of operation @xmath762 ) , and hence can not be in @xmath2 . the rest of the proof is a straightforward verification . we can now prove the following . there is an @xmath0-recognizable set of graphs which is not @xmath1-recognizable . let @xmath77 be a set of integers which is not recognizable in @xmath965 , for instance the set of prime numbers , and let @xmath2 be the set of cliques @xmath945 for @xmath966 . then @xmath2 is @xmath0-recognizable by lemma [ cliques hr - rec ] . we now consider a set of @xmath1-terms describing @xmath2 and using exactly 2 port labels , @xmath428 and @xmath437 . recall that @xmath447 denotes the @xmath1-constant of type @xmath448 , that is , the graph with a single vertex that is a @xmath428-port and no edges . the constant @xmath967 is defined similarly . now let @xmath968 , and @xmath969 . it is not difficult to verify that @xmath970 denotes the clique @xmath945 where all the vertices are @xmath428-ports , @xmath945 itself is denoted by the term @xmath971 , and the set @xmath788 of all @xmath1-terms of the form @xmath970 is recognizable ( as a set of terms , or trees ) . if @xmath2 is @xmath1-recognizable , then the set of @xmath1-terms in @xmath788 that denote graphs in @xmath2 is recognizable . this set consists of all the terms of the form @xmath971 with @xmath966 , and it can be shown by standard methods that it is not recognizable . it follows that @xmath2 is not @xmath1-recognizable . since graphs are relational structures , logical formulas can be used to specify sets of graphs . monadic second - order logic is especially interesting because _ every monadic second - order definable set of finite graphs is @xmath1-recognizable ( courcelle @xcite ) . _ there is actually a version of monadic second - order logic allowing quantifications on edges and sets of edges ( one replaces the graph under consideration by its incidence graph ; we omit details ) . we say that a set is @xmath972-definable if it is definable by a monadic second - order formula with edge and edge set quantifications , and that we use the phrase @xmath973-definable to refer to the first notion . it is immediately verified ( from the definition ) that _ every @xmath973-definable set is @xmath972-definable . _ the two following statements are more difficult . _ every @xmath972-definable set of simple graphs is @xmath0-recognizable ( courcelle @xcite ) . _ _ if a set of simple graphs is uniformly @xmath92-sparse for some @xmath92 and @xmath972-definable , then it is @xmath973-definable ( courcelle @xcite ) . _ this is somewhat analogous to the situation of theorem [ no kmm ] ( see proposition [ prop k - sparse ] ) . however the combinatorial conditions are different : if a set of graphs is uniformly @xmath92-sparse for some @xmath92 , it is without @xmath974 for some @xmath47 , but the converse does not hold . it is proved in the book by bollobas @xcite that , for each @xmath975 , there is a number @xmath242 such that for each @xmath5 , there is a graph with @xmath5 vertices and @xmath976 edges that does not contain @xmath974 , where @xmath977 . for these graphs , the number of edges is not linearly bounded in terms of the number of vertices , so they are not uniformly @xmath92-sparse for any @xmath92 . it is not clear how to extend courcelle s proof in @xcite , to use the condition _ without @xmath974 _ instead of _ uniformly @xmath92-sparse_. the formal setting of relational structures is very convenient to deal with simple graphs , as we have seen already . it can also be used to formalize multi - graphs ( i.e. , graphs with multiple edges ) , if we consider two - sorted relational structures . formally , a _ multi - graph with sources _ in @xmath79 is a structure of the form @xmath978 where @xmath979 is the set of vertices , @xmath980 is the set of edges , each @xmath981 is an element of @xmath979 , and @xmath982 is a ternary relation of type @xmath983 . we interpret the relation @xmath984 to mean that @xmath985 is an edge from vertex @xmath593 to vertex @xmath498 . we denote by @xmath986 the set of multi - graphs with sources in @xmath79 . as in the study of @xmath184 or @xmath701 , we assume that the finite sets of source labels @xmath79 are taken in a fixed countable set . we let @xmath987 be the union of the @xmath986 for all finite sets @xmath79 of source labels . graphs and hypergraphs with multiple edges and hyperedges are often used , see the volume edited by rozenberg @xcite . in this context , it is in fact frequent to consider operations on multi - graphs that are very similar to the @xmath0-operations on @xmath701 . more precisely , the operations of disjoint union , source renaming , source forgetting and source fusion can be defined naturally on multigraphs with sources : thus @xmath987 can be seen naturally as an @xmath0-algebra . it is clear that each simple graph in @xmath195 can be considered as an element in @xmath986 . it is important to note however that the @xmath0-operations on @xmath988 , when applied to such simple graphs , do not necessarily yield the same result as in @xmath701 . for instance , let @xmath237 be distinct elements of @xmath79 , and let @xmath989 be a simple graph . the action of fusing the @xmath242-source and the @xmath243-source of @xmath652 may now result in multiple edges : if there were arrows in both directions between @xmath990 and @xmath991 , or if there were arrows to ( resp . from ) a vertex of @xmath652 from ( resp . to ) both @xmath990 and @xmath991 . in contrast , the same operation in @xmath195 yields @xmath992 , an element of @xmath195 by definition . to avoid confusion , we will denote by @xmath993 this operation when used in @xmath987 . fortunately , we do not have this sort of problem with the other operations : applying the operations of disjoint union , source renaming or source forgetting to simple graphs considered as elements of @xmath987 yields the same result as applying the same operations within the algebra @xmath701 . we let @xmath994 be the signature on @xmath987 consisting of the operations of the form @xmath209 , @xmath244 , @xmath995 and @xmath993 . thus , @xmath987 is an @xmath994-algebra . we observe that , as a signature ( that is , as a set of symbols denoting operations ) , @xmath994 is in natural bijection with @xmath0 . so we do nt really need to introduce the new notation @xmath994 , and we could very well say that @xmath987 is an @xmath0-algebra . we simply hope , by introducing this notation , to clarify our comparative study of recognizable subsets in the algebras @xmath701 and @xmath987 . this distinction will be useful in the proofs of theorems [ l vs i(l ) rec ] and [ l vs u(l ) rec ] . to summarize and amplify the above remarks , let us introduce the following notation . we denote by @xmath996 the natural injection . for each multi - graph @xmath652 , we denote by @xmath997 the simple graph obtained from @xmath652 by fusing multiple edges ( with identical origin and end ) : that is , @xmath108 is a mapping from @xmath987 onto @xmath701 . elementary properties of @xmath998 and @xmath108 are listed in the next proposition . [ i , u morphisms ] the mapping @xmath999 is a homomorphism of @xmath0-algebras . the mapping @xmath1000 is not a homomorphism , but it commutes with the operations of the form @xmath209 , @xmath244 and @xmath240 . @xmath998 does not commutes with the operations of the form @xmath247 , but if @xmath1001 , then @xmath1002 . finally , if @xmath1001 , then @xmath1003 and @xmath1004 . we now prove the following theorems , which describe the interaction between @xmath994-recognizability of sets of multi - graphs and @xmath0-recognizability of sets of simple graphs . [ simple graphs rec ] the set of simple graphs is @xmath994-recognizable . more precisely , for each finite set of source labels @xmath79 , @xmath1005 is @xmath994-recognizable . [ l vs i(l ) rec ] let @xmath79 be a finite set of source labels and let @xmath1006 . then @xmath2 is @xmath0-recognizable if and only if @xmath1007 is @xmath994-recognizable . [ l vs u(l ) rec ] let @xmath79 be a finite set of source labels and let @xmath1008 . if @xmath2 is @xmath994-recognizable , then @xmath1009 is @xmath0-recognizable . we first introduce the notion of the type of a multi - graph : as for the elements of @xmath184 , if @xmath1010 , we let @xmath1011 be the restriction of @xmath652 to its @xmath79-sources and to the edges between them . we also denote by @xmath160 the relation on @xmath987 induced by this type mapping : two multi - graphs @xmath1012 are @xmath160-equivalent if @xmath1013 . [ zeta for multi ] the type relation @xmath160 is an @xmath994-congruence on @xmath987 . moreover , for each finite set of source labels @xmath79 , the elements of @xmath1005 can be found in only a finite number of @xmath160-classes . the result follows from the following , easily verifiable identities , where the multi - graphs @xmath652 , @xmath694 are assumed to have the appropriate sets of sources . @xmath1014 the finiteness of the number of @xmath160-classes containing elements of @xmath1005 follows from the fact that there are only finitely many source - only simple graphs with sources in @xmath79 . we also introduce the following finite invariant for a simple graph @xmath989 . we define @xmath1015 to be the set of all pairs @xmath1016 of elements of @xmath79 such that @xmath1017 , @xmath1018 and there exists a vertex @xmath593 of @xmath652 with either edges from @xmath593 to both @xmath990 and @xmath991 , or edges to @xmath593 from both @xmath1019 and @xmath991 . the set @xmath1015 can be viewed as a symmetric anti - reflexive relation on @xmath79 . [ charact multi ] let @xmath652 be a simple graph in @xmath195 and let @xmath1017 be elements of @xmath79 . then @xmath1020 has multiple edges if and only if @xmath1021 or @xmath1022 has multiple edges . we first observe that @xmath1020 has multiple edges if and only if @xmath1018 and at least one of the following situations occurs : there are edges in both directions between @xmath990 and @xmath991 , or there is a vertex @xmath593 of @xmath652 with edges from ( resp . to ) both @xmath990 and @xmath991 ( this includes the case where there is a loop at @xmath990 or @xmath991 and an edge in either direction between @xmath990 and @xmath991 ) . that is , @xmath1020 has multiple edges if and only @xmath1023 or there are edges in both directions between @xmath990 and @xmath991 . we also observe that @xmath1022 is a subgraph of @xmath1020 , so the former is simple if the latter is . finally , the existence of edges in both directions between @xmath990 and @xmath991 is sufficient to ensure that @xmath1022 has multiple edges . these observations put together suffice to prove the lemma . we are now ready to prove theorem [ simple graphs rec ] . let @xmath1024 be the following relation , defined on each @xmath986 . we let @xmath1025 if both @xmath652 and @xmath639 have multiple edges , or both @xmath652 and @xmath639 are simple graphs , @xmath1026 and @xmath1027 . it is immediate that @xmath1024 is an equivalence relation , saturating @xmath1005 . it follows from lemma [ zeta for multi ] and from the fact that @xmath1015 is a subset of the finite set @xmath1028 , that @xmath1024 is locally finite . so we only need to show that @xmath1024 is an @xmath994-congruence . we need to describe the interaction between the mapping @xmath1029 and the @xmath994-operations . as observed in proposition [ i , u morphisms ] , all @xmath994-operations preserve simple graphs except for the operations of the form @xmath993 . assuming that @xmath1030 are simple graphs with the appropriate sets of sources , we easily verify the following : @xmath1031 moreover , if @xmath1032 and @xmath1020 is simple ( if it is nt , its @xmath1029-image is not defined ) , then @xmath1033 consists of : ( 1)all pairs in @xmath1015 , ( 2)all pairs @xmath1034 such that there are edges in @xmath1011 from @xmath242 to @xmath180 and from @xmath243 to @xmath232 , or from @xmath180 to @xmath242 and from @xmath232 to @xmath243 , ( 3)all pairs @xmath1035 ( resp . @xmath1036 ) such that @xmath1037 ( resp . @xmath1038 ) , ( 4)all pairs @xmath1035 and @xmath1036 such that there are edges in @xmath1011 between @xmath242 and @xmath243 ( in either direction ) and between @xmath242 or @xmath243 and @xmath180 ( in any direction ) . let us justify this statement : it is easy to see that all these pairs belong to @xmath1033 . in particular , @xmath1039 since , as @xmath1020 is assumed to be simple , there is no @xmath1040 such that @xmath1041 and @xmath1042 . conversely , let us consider distinct edges in @xmath1043 , from @xmath498 to @xmath593 and from @xmath782 to @xmath593 , as in figure [ fig : distinct edges ] ( note that @xmath593 and @xmath498 may be equal ) , such that @xmath1044 and @xmath1045 for @xmath1046 . ( 90,25)(0,-22 ) ( n0)(0.0,0.0)@xmath498 ( n1)(0.0,-20.0)@xmath782 ( n2)(24.0,-8.0)@xmath593 ( n3)(50.0,-8.0)@xmath782 ( n4)(78.0,-8.0)@xmath1047 ( n0,n2 ) ( n1,n2 ) ( n3,n4 ) ( n4 ) if neither @xmath593 , nor @xmath498 nor @xmath782 is the @xmath242- and @xmath243-source in @xmath639 , then we are in case ( 1 ) , _ i.e. _ , @xmath1048 . if @xmath593 is the @xmath242- and @xmath243-source in @xmath639 but neither @xmath498 nor @xmath782 is , then @xmath1049 satisfies case ( 1 ) or ( 2 ) . if @xmath498 is the @xmath242- and @xmath243-source in @xmath639 but neither @xmath593 nor @xmath782 is , then @xmath1049 satisfies case ( 3 ) . the same holds by symmetry if @xmath782 is the only one of these three vertices to be the @xmath242- and @xmath243-source in @xmath639 . finally if @xmath1050 ( resp . @xmath1051 ) and is the @xmath242- and @xmath243-source , in @xmath639 then there is an edge between the @xmath242- and the @xmath243-source in @xmath652 and @xmath1049 satisfies case ( 4 ) . the case of edges from @xmath593 to @xmath498 and to @xmath782 is symmetrical . in particular , @xmath1052 , @xmath1053 , @xmath1054 and @xmath1033 are entirely determined by @xmath1015 , @xmath1011 and @xmath1055 . let us now consider @xmath1056 in @xmath987 ( with the appropriate sets of sources ) such that @xmath1025 and @xmath1057 . if @xmath652 is not simple , then neither are @xmath639 , @xmath1058 , @xmath1059 , @xmath1060 and @xmath1020 . in particular , we have @xmath1061 , @xmath1062 , @xmath1063 and @xmath1064 . assume now that @xmath652 and @xmath694 are simple . then so are @xmath1058 , @xmath1059 and @xmath1060 , and we have seen that their @xmath1029-images are determined by @xmath1015 and @xmath1055 . since @xmath160 is an @xmath994-congruence ( lemma [ zeta for multi ] ) , it follows that @xmath1024 is preserved by the operations @xmath209 , @xmath244 and @xmath240 . by lemma [ charact multi ] , whether @xmath1020 is simple , is determined by @xmath1011 and @xmath1015 , and hence @xmath1020 and @xmath1065 are both non - simple ( and then @xmath1024-equivalent ) or both simple . in the latter case , their @xmath1029-images are equal since they are both determined by @xmath1066 and @xmath1067 . thus @xmath1024 is preserved by the operation @xmath993 . this concludes the proof of theorem [ simple graphs rec ] . recall that we want to show that for each @xmath776 , @xmath2 is @xmath0-recognizable if and only if @xmath1007 is @xmath994-recognizable . one direction is quickly established : we know from proposition [ i , u morphisms ] that @xmath1068 . if @xmath2 is @xmath0-recognizable , then @xmath1069 is @xmath994-recognizable since @xmath108 is a homomorphism . in view of theorem [ simple graphs rec ] , it follows that @xmath1007 is @xmath994-recognizable as well . conversely , let us assume that @xmath1007 is @xmath994-recognizable and let @xmath130 be a locally finite @xmath994-congruence on @xmath987 saturating @xmath1007 . we want to define a locally finite @xmath0-congruence @xmath98 on @xmath701 saturating @xmath2 . for each symmetric anti - reflexive relation @xmath77 on a finite set of source labels @xmath439 and for each graph @xmath1070 , let @xmath1071 be the graph obtained from @xmath652 by deleting the edges between the @xmath242-source and the @xmath243-source for each pair @xmath1016 in @xmath439 . let also @xmath1072 be the composition of the operations @xmath247 for all @xmath1073 , in any order . for @xmath1074 , we let @xmath825 if @xmath1075 , @xmath1026 and , for each symmetric anti - reflexive relation @xmath77 on @xmath439 , @xmath1076 the relation @xmath98 is clearly an equivalence relation , and it is locally finite since @xmath130 and @xmath160 are . moreover , it saturates @xmath2 since @xmath692 if and only if @xmath1077 , and @xmath130 saturates @xmath1007 . the rest of the proof consists in showing that @xmath98 is an @xmath0-congruence . [ [ the - source - renaming - operation ] ] the source renaming operation + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath831 in @xmath1078 . then @xmath1079 . since @xmath130 is a congruence and in view of proposition [ i , u morphisms ] , @xmath1080 . it also follows from lemma [ ffvv light ] that @xmath1081 . let us now consider a symmetric anti - reflexive relation @xmath77 on the set of source labels of @xmath1060 . it is easily verified that @xmath1082 where @xmath1083 . we also note that if @xmath1084 , then @xmath1085 and @xmath240 commute . moreover @xmath1086 and @xmath1087 . thus @xmath1088 . now , using the fact that @xmath998 commutes with @xmath995 we have @xmath1089 since @xmath130 is an @xmath994-congruence , it follows that @xmath1090 and , finally , that @xmath1091 . [ [ the - source - forgetting - operation ] ] the source forgetting operation + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the proof is the same as for the source renaming operation , with this simplifying circumstance that @xmath1092 and @xmath1093 ( since @xmath242 is not a source label of @xmath1059 , and hence does not occur in @xmath77 ) . [ [ the - source - fusion - operation ] ] the source fusion operation + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath831 in @xmath1078 . here it is not immediate that @xmath1094 . however , if we let @xmath1095 , we know that @xmath1096 we note that @xmath1097 is equal to @xmath992 if @xmath652 has no edge between its @xmath242- or @xmath243-source , or if it has a loop at either . otherwise , @xmath992 is equal to @xmath1097 with a loop added to its @xmath242-source , that is : @xmath1098 where @xmath1099 and @xmath1100 are source labels not in @xmath439 and @xmath980 is the graph in @xmath1101 with 2 vertices and a single edge from its @xmath1099-source to its @xmath1100-source . observe also that the existence of loops at , or edges between the @xmath242- and @xmath243-source of @xmath652 is a condition that depends only on @xmath1011 , so it will be satisfied by both @xmath652 and @xmath639 or by neither . in the first case , where @xmath1102 , we find immediately that @xmath1103 . in the second case , the same @xmath130-equivalence is derived from proposition [ i , u morphisms ] and equation ( @xmath1104 ) above . by lemma [ ffvv light ] , @xmath160-equivalence is preserved by the operation @xmath247 . now let @xmath77 be a symmetric anti - reflexive relation on @xmath439 : we consider the graph @xmath1105 . our first observation is that @xmath1106 where @xmath1107 next , we observe that @xmath1108 . thus we have @xmath1109 and hence @xmath1110 . it follows that @xmath1111 . [ [ the - disjoint - union - operation ] ] the disjoint union operation + + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath831 in @xmath195 and @xmath832 in @xmath1078 ( where @xmath79 and @xmath439 are disjoint ) . since @xmath998 and @xmath160 preserve @xmath209 , we have @xmath1112 and @xmath1113 . now let @xmath77 be a symmetric anti - reflexive relation on @xmath1114 . let @xmath443 ( resp . @xmath171 ) be the restriction of @xmath77 to @xmath79 ( resp . @xmath439 ) and let @xmath1115 . it is easily verified that @xmath1116 it now follows from proposition [ i , u morphisms ] that @xmath1117 thus @xmath1118 , and hence @xmath1119 . this concludes the proof of theorem [ l vs i(l ) rec ] . let @xmath1120 be @xmath994-recognizable , and let @xmath130 be a locally finite @xmath994-congruence saturating @xmath2 . we want to show that @xmath1009 ( a subset of @xmath195 ) is @xmath0-recognizable . let @xmath1121 . we let @xmath1122 if , for each @xmath1123 , there exists @xmath1124 such that @xmath1125 , and symmetrically , for each @xmath1124 , there exists @xmath1123 such that @xmath1125 . the relation @xmath98 is easily seen to be a locally finite equivalence relation on @xmath701 , saturating @xmath1009 . there remains to see that @xmath98 is an @xmath0-congruence . we first establish the following lemma . [ for 6.4 ] let @xmath1126 and let @xmath1127 . * @xmath1128 if and only if there exist multi - graphs @xmath1129 such that @xmath1130 , @xmath1131 and @xmath1132 . * @xmath1133 if and only if there exists a multi - graph @xmath835 such that @xmath1134 and @xmath1131 . * @xmath1135 if and only if there exists a multi - graph @xmath835 such that @xmath1136 and @xmath1131 . * @xmath1137 if and only if there exists a multi - graph @xmath835 such that @xmath1138 and @xmath1139 . recall that @xmath652 and @xmath997 have the same set of vertices , and each edge @xmath985 of @xmath997 arises from the identification @xmath1140 edges of @xmath652 between the same vertices . if @xmath1128 , each edge of @xmath997 is in exactly one of @xmath694 and @xmath788 . let @xmath835 ( resp . @xmath889 ) be the graph obtained from @xmath694 ( resp . @xmath788 ) by replacing each edge @xmath985 by @xmath1141 parallel edges . then @xmath1142 , @xmath1131 and @xmath1132 , as required . the proof of the statements relative to the operations @xmath244 and @xmath240 is done in the same fashion . let us finally consider the case where @xmath1137 . if @xmath1143 , that is , @xmath1144 , then @xmath1145 and we can let @xmath1146 . if @xmath1147 , we let @xmath835 be obtained from @xmath694 be obtained from @xmath694 as follows : for each vertex @xmath593 , each edge @xmath985 from @xmath593 to @xmath498 ( @xmath1148 ) is replaced by @xmath1141 parallel edges , and the edges from @xmath593 to @xmath242 and @xmath243 are duplicated to a total of @xmath1141 edges . we can now conclude the proof of theorem [ l vs u(l ) rec ] , by proving that @xmath98 is an @xmath0-congruence . let @xmath831 and @xmath1149 . let @xmath1150 . by lemma [ for 6.4 ] , @xmath1151 for some @xmath1152 and @xmath1153 . since @xmath831 and @xmath832 , there exist @xmath1154 and @xmath1155 such that @xmath1156 and @xmath1157 . . then @xmath1159 and @xmath1160 . by symmetry , this shows that @xmath1161 . the verification that @xmath98 is preserved by the other @xmath0-operations proceeds along the same lines . this concludes the proof of theorem [ l vs u(l ) rec ] . the class @xmath194 , defined in section [ sec rel structures ] , has already been discussed in terms of the signatures @xmath275 , @xmath1 and @xmath0 since it is a domain in each of the three algebras @xmath184 , @xmath426 and @xmath701 . in this section , we consider a different set of operations on @xmath194 , arising from the theory of the modular decomposition of graphs , which makes @xmath194 an algebra ( one - sorted for a change ! ) . this algebraic framework was considered by the authors , in @xcite and @xcite . we first recall the definition of the composition operation on graphs . let @xmath694 be a graph with vertex set @xmath1162 = \{1,\ldots , n\}$ ] ( @xmath1163 ) . if @xmath1164 are graphs , then the composite @xmath1165 is obtained by taking the disjoint union of the graphs @xmath1166 , and by adding , for each edge @xmath1167 of @xmath694 where @xmath523 , an edge from every vertex of @xmath1168 to every vertex of @xmath1169 . we say that a graph is _ indecomposable _ , or _ prime _ , if it can not be written non - trivially as a composition ( a composition is trivial if each of its arguments is a singleton ) . it is easily verified that if @xmath694 and @xmath835 are isomorphic graphs , then the corresponding composition operations yield isomorphic graphs . so we fix a set @xmath1170 of representatives of the isomorphism classes of indecomposable graphs . in particular , we may assume that every graph in @xmath1170 has a vertex set of the form @xmath1162 $ ] for some @xmath1163 . we also denote by @xmath1170 the resulting _ modular signature _ , consisting of the composition operations defined by these graphs . the @xmath1170-algebra of graphs is denoted by @xmath1171 . it turns out that every finite graph admits a _ modular decomposition _ , that is , it can be expressed from the single - vertex graph using only operations from @xmath1170 . this fact has been rediscovered a number of times in the context of graph theory and of other fields using graph - theoretic representations . we refer to @xcite for a historical survey , and to @xcite for a concise presentation . in other words , @xmath194 is generated by the signature @xmath1170 augmented with the constants @xmath1172 and @xmath1173 , which denote a single vertex graph , respectively with and without a single loop edge . the modular decomposition of a graph is unique up to certain simple ( equational ) rules , see for instance @xcite . moreover , the modular decomposition of a graph can be computed in linear time @xcite . our first results connect @xmath1-recognizability and @xmath1170-recognizability . [ vr - f ] every @xmath1-recognizable set of graphs is @xmath1170-recognizable . in view of proposition [ easy recog facts ] and theorem [ thm vr equivalence ] , it suffices to show that every operation in @xmath1170 is @xmath1174-derived . for each integer @xmath139 , let @xmath1175 be the unary operation on @xmath426 , of type @xmath1176 , defined as follows : given a graph without ports , it simply marks every vertex with port label @xmath139 ( leaving the set of vertices and the edge relation unchanged ) . note that @xmath1175 is a qfd unary operation , and hence a @xmath1174-operation . let @xmath694 be an @xmath5-ary operation , that is , a graph in @xmath1170 with vertex set @xmath1162 $ ] , and let @xmath1177 be its edge relation . if @xmath1166 are finite graphs , the construction of @xmath1165 can be described as follows : - construct the disjoint union , @xmath1178 , an element of @xmath1179)$ ] ; - apply ( in any order ) to this disjoint union the operations @xmath1180 for all @xmath1181 $ ] such that @xmath1167 is an edge of @xmath694 and @xmath523 ; - forget all ports , that is , apply the operation @xmath1182 . this completes the verification that the operation defined by @xmath694 can be expressed as a @xmath1174-term , and hence the proof . the following result shows that the converse of proposition [ vr - f ] does not hold . every set of prime graphs is @xmath1170-recognizable , and there is a set of prime graphs which is not @xmath1-recognizable . let @xmath2 be a set of prime graphs , and let @xmath130 be the relation on @xmath194 defined as follows . we let @xmath1183 if one of the following holds : * neither @xmath652 nor @xmath694 is prime ; * @xmath652 and @xmath694 are both ( the graph with one vertex and no edge ) ; * @xmath652 and @xmath694 are both not , prime and in @xmath2 ; * @xmath652 and @xmath694 are both not , prime and not in @xmath2 . this is clearly an equivalence relation with four classes , which saturates @xmath2 . moreover , @xmath130 is an @xmath1170-congruence . indeed , let @xmath788 be a graph with @xmath5 vertices ; for @xmath1184 , let @xmath1185 for each @xmath139 . if for some @xmath139 , @xmath1186 , then @xmath1187 , and neither @xmath1188 nor @xmath1189 is prime : therefore they are equivalent . otherwise , @xmath1190 for each @xmath139 , @xmath1191 and @xmath1192 are both equal to @xmath788 , and hence they are equivalent . this concludes the proof that every set of prime graphs is @xmath1170-recognizable . before we exhibit a set of prime graphs which is not @xmath1-recognizable , we define inductively a sequence of @xmath1-terms written with three port labels @xmath1193 . we let @xmath1194 the term @xmath1195 ( forgetting all port labels in @xmath1196 ) denotes the string graph @xmath1197 , with @xmath1198 vertices , say @xmath1199 and edges from @xmath139 to @xmath1200 for each @xmath1201 . each of these graphs is prime . now let @xmath77 be a set of positive integers that is not recognizable in @xmath965 and let @xmath2 be the set of all terms @xmath1202 with @xmath966 . from the above discussion , we know that @xmath2 is @xmath1170-recognizable . if @xmath2 was @xmath1-recognizable , standard arguments would show that the set of @xmath1-terms @xmath1196 ( @xmath966 ) would be recognizable as well , and it would follow that @xmath77 is recognizable , contradicting its choice . now let @xmath7 be a finite subsignature of the modular signature @xmath1170 . a graph which can be constructed from one - vertex graphs using only operations from @xmath7 is called an @xmath7-graph . the next result deals with sets of @xmath7-graphs . this finiteness condition ( the elements of @xmath2 are built by repeated composition of a finite number of graph - based operations ) is non - trivial . in fact , for many natural classes of graphs such as rectangular grids , it is not satisfied : since grids are indecomposable , a set of graphs containing infinitely many grids can not satisfy our finiteness condition . but that condition is satisfied by other classical classes ( e.g. cographs , series - parallel posets ) , see @xcite . using results of courcelle @xcite , we can show the following result , which yields in particular a weak converse of proposition [ vr - f ] . [ fsvr ] let @xmath7 be a finite subsignature of @xmath1170 and let @xmath2 be a set of @xmath7-graphs . the following properties are equivalent : 1 . @xmath2 is @xmath275-recognizable ; 2 . @xmath2 is @xmath1-recognizable . @xmath2 is @xmath1170-recognizable . @xmath2 is @xmath7-recognizable . the equivalence of ( 1 ) and ( 2 ) can be found in theorem [ thm vr equivalence ] . proposition [ vr - f ] shows that ( 2 ) implies ( 3 ) . and ( 3 ) implies ( 4 ) as an immediate consequence of proposition [ easy recog facts ] since @xmath7 is a subsignature of @xmath1170 . the fact that ( 4 ) implies ( 1 ) is a consequence of two results of courcelle : ( * ? * theorem 4.1 ) , which states that if a set of @xmath7-graphs is @xmath7-recognizable , then it is definable in a certain extension of @xmath1203-logic ; and ( * ? ? ? * theorem 6.11 ) , which states that all sets definable in this logical language are @xmath275-recognizable . theorem [ fsvr ] states that for sets of graphs with only finitely many prime subgraphs , all four notions of recognizability are equivalent . presented in this fashion , the statement is somewhat similar to that of theorem [ no kmm ] . in this article , we have investigated the recognizability of sets of graphs quite in detail , focusing on the robustness of the notion , which was not immediate since many signatures on graphs can be defined . although we had in mind sets of graphs , we have proved that embedding graphs in the more general class of relational structures does not alter recognizability . we have proved that the very same structural conditions that equate @xmath1-equational and @xmath0-equational sets of graphs , also equates @xmath0-recognizability and @xmath1-recognizability . summing up , we have defined a number of tools for handling recognizability . some questions remain to investigate . @xmath1204when is it true that a quantifier - free operation preserves recognizability ? results in this direction have been established in courcelle @xcite . are they applicable to quantifier - free definable operations ? in particular , is it true that the set of disjoint unions of two graphs , one from each of two @xmath1-recognizable sets is @xmath1-recognizable ? @xmath1204which quantifier - free definable operations can be added to the signature @xmath0 , in such a way that the class of @xmath0-recognizable sets is preserved ( as is the case when we extend @xmath1 to @xmath1174 ) ? the paper by blumensath and courcelle @xcite , which continues the present research , considers unary non qfd operations that can be added to @xmath577 and to @xmath184 while preserving the classes of equational and recognizable sets . @xmath1204our example of an @xmath0-recognizable , not @xmath1-recognizable set of cliques , is based on the weakness of the parallel composition of graphs with sources , i.e. , the fact that this operation is not able to split large cliques . can one find another example , based on a different argument ? if one can not , what does this mean ? we conclude with an observation concerning the finiteness of signatures . whereas all finite words on a finite alphabet can be generated by this alphabet and only one operation , dealing with finite graphs ( by means of grammars , automata and related tools ) requires infinite signatures . more precisely , one needs infinitely many operations to generate all finite unlabelled graphs ( see remark [ economical ] below ) . on the other hand , applications to testing graph properties require the consideration of algebras generated by a finite signature . here is the reason . let @xmath23 be an @xmath7-algebra of graphs . if the unique valuation homomorphism @xmath1205 ( which evaluates a term into an element of @xmath23 ) is surjective , i.e. , if @xmath7 generates @xmath23 , then a subset @xmath2 of @xmath23 is recognizable if and only if @xmath90 is a recognizable set of terms ( see proposition [ easy recog facts ] and section [ significance ] ) . and the membership of a term in a recognizable set can be verified in linear time by a finite deterministic ( tree ) automaton . hence the membership of a graph @xmath652 in @xmath2 can be checked as follows : \(1 ) one must first find some term @xmath47 such that @xmath1206 , \(2 ) then one checks whether @xmath47 belongs to @xmath90 . the latter step can be done in time proportional to the size of @xmath47 , usually no larger than the number of vertices of @xmath652 . although any term @xmath47 with value @xmath652 gives the correct answer , it may be difficult to find at least one ( graph parsing problems may be @xmath670-complete ) . because of this fact many hard problems ( in particular if they are expressed in monadic second - order logic ) can be solved in linear time on sets of graphs of bounded tree - width , and also on sets of graphs of bounded clique - width , provided the graphs are given with appropriate decompositions , see courcelle @xcite , courcelle and olariu @xcite or downey and fellows @xcite . if the decompositions are not given , one can achieve linear time for graphs of bounded tree - width and @xmath972 problems using a result by bodlaender @xcite , and polynomial time for graphs of bounded clique - width and @xmath973 problems using a result by oum and seymour @xcite . however , even if @xmath7 is infinite or is finite without generating the set @xmath23 , recognizability remains interesting as an algebraic concept , and for every restriction to a finitely generated subset of @xmath23 , we are back to the `` good '' case of a finitely generated algebra . finally , we think that infinite signatures can be used for checking graph properties defining recognizable sets . this will not be possible by _ finite _ tree - automata if the graph algebra is not finitely generated , but it can perhaps be done with automata using `` oracles '' . an oracle would be a subroutine handling some verifications for big subgraphs that can not be decomposed by the operations under consideration . this idea needs of course further elaboration . [ economical ] we asserted above that finite unlabelled graphs can not be generated with a finite signature . this is not entirely correct , and we briefly describe here a signature with 6 operations on a 2-sorted algebra which generates , somewhat artificially , all finite graphs ( undirected and without loops ) . these operations have no good behaviour with respect to automata and verification questions , and such an `` economical '' generation of graphs is useless . the 2 sorts are @xmath1207 , the set of finite graphs equipped with a linear order of their vertex set , and @xmath1208 , the set of ordinary , unordered graphs . there is one unary operation of type @xmath1209 , which forgets the order on the vertex set . all other operations are unary , of type @xmath1210 : one consists in adding one new vertex , to be the new least element ; one adds an ( undirected ) edge between the two least vertices ; one performs a circular shift of the vertices ; and one swaps the two least vertices . the three last operations leave the graph unchanged if it has less than 2 vertices . finally , one adds a 6th , nullary operation , of type @xmath1207 : the constant @xmath487 , standing for the empty graph with no vertices . in this appendix , we discuss some equivalences and transformations of logical formulas which can be used to give upper bounds for the index of congruences considered in this paper , and to complete the proof of the effectiveness of certain notions ( e.g. quantifier - free definition schemes ) . more specifically , we make precise in what sense we can state , as we do in the body of the paper , that the set of first - order ( resp . monadic second - order ) formulas over finite sets of relations , constants and free variables , and with a bounded quantification depth , can be considered as finite . moreover , explicit upper bounds on the size of these finite sets are derived , which can be used to justify the termination of some of our algorithms , and in evaluating their complexity . that these upper bounds have unbounded levels of exponentiation is not unexpected , and even unavoidable by frick and grohe @xcite . let @xmath1211 be boolean variables and let @xmath1212 be the set of boolean formulas written with these variables . it is well known that @xmath1212 is finite up to logical equivalence . for further reference , we record the following more precise statement . we let @xmath1213 be the set of boolean formulas in disjunctive normal form , where in each disjunct , variables occur at most once and in increasing order , no two disjuncts are equal , and disjuncts are ordered lexicographically . these constraints guarantee the announced cardinality of @xmath1213 ; the rest of the proof is classical . let us consider finite sets @xmath171 and @xmath79 , of relational symbols and of constants ( nullary relations , source labels ) as in section [ sec rel structures ] . recall that , if @xmath64 is a finite set , @xmath1215 denotes the set of first - order formulas in the language of @xmath174-structures , with free variables in @xmath64 . for unproved results in this section , we refer the reader to @xcite . several notions of semantic equivalence of formulas can be defined . if @xmath1216 , say that @xmath1217 if for every @xmath174-structure @xmath111 and for every assignment of values in @xmath111 to the elements of @xmath64 , @xmath1218 and @xmath1219 are both true or both false . say also that @xmath1220 if the same holds for every finite or countable @xmath174-structure @xmath111 , and @xmath1221 if @xmath111 is restricted to being finite . the equivalences @xmath130 and @xmath1222 coincide by the lwenheim - skolem theorem . indeed this theorem states that if a closed formula has an infinite model , then it has one of each infinite cardinality : to prove our claim , it suffices to apply it to the formula @xmath1223 . we note that this equivalence can not be extended to monadic second - order formulas : there exists an ms formula with a unique model , isomorphic to the set of integers @xmath1224 with its order . the equivalence @xmath130 ( or @xmath1222 since we consider only first - order formulas ) is semi - decidable : by gdel s completeness theorem , @xmath1225 if and only if the formula @xmath1226 has a proof , which is a recursively enumerable property . trakhtenbrot proved that one can not decide whether a first - order formula is true in every finite structure , thus proving that @xmath1227 is not decidable . however , the negation of @xmath1227 is semi - decidable : if @xmath1228 , a counter - example can be produced by exploring systematically all finite @xmath174-structures . this is a proof also that @xmath130 and @xmath1227 do not coincide . if @xmath1230 , and if @xmath1231 , we denote by @xmath1232 the formula in @xmath1215 obtained by replacing each occurrence of @xmath1233 in @xmath243 by @xmath1234 . it is clear that if @xmath243 and @xmath1235 are equivalent boolean formulas , then @xmath1236 . a _ boolean transformation step _ consists in replacing in a first - order formula , a sub - formula of the form @xmath1232 by the equivalent formula @xmath1237 , where @xmath1238 are equivalent . then we let @xmath1229 if @xmath1218 can be transformed into @xmath1219 by a sequence of boolean transformation steps and of renaming of bound variables . it is clear that if @xmath1229 , then @xmath1239 . we want to show that each first - order formula is effectively equivalent to an @xmath83-equivalent formula of the same quantifier height , and to give an upper bound on the number of @xmath83-equivalence classes of formulas of a given height . let @xmath1240 be the set of quantifier - free formulas in @xmath1215 . such formulas are boolean combinations of atomic formulas . let @xmath1241 be the set of these atomic formulas . note that each atomic formula is either of the form @xmath1050 , where @xmath593 and @xmath498 are in @xmath1242 , or @xmath1243 where @xmath216 is a @xmath173-ary relation in @xmath171 and the @xmath169 are in @xmath1244 . letting @xmath1245 and @xmath1246 , it is easily verified that @xmath1247 we let @xmath1248 be this function . note that if we allow for the ( effective ) syntactic simplifications of identifying the formulas of the form @xmath1249 with the constant @xmath264 , and of identifying the formulas @xmath1050 and @xmath1250 , we can lower the value of @xmath1248 to @xmath1251 . [ red qf ] there exists a subset @xmath1252 of @xmath1240 , of cardinality @xmath1253 , such that every formula in @xmath1240 can be effectively transformed to an @xmath83-equivalent formula in @xmath1252 . by definition of quantifier - free formulas , @xmath1240 is the set of all formulas of the form @xmath1232 , where @xmath243 is a boolean formula and the @xmath1234 are atomic formulas . now let @xmath1252 be the set of all formulas of the form @xmath1232 , where @xmath1254 and the @xmath1234 are pairwise distinct atomic formulas . the proof of the precise statement is now immediate , using proposition [ red bool ] . let us consider graphs with sources , so that @xmath171 consists of a single , binary edge relation . then @xmath1255 and @xmath1256 . thus the type equivalence @xmath160 ( see section [ sec elementary properties ] and lemma [ type equivalence ] ) has at most @xmath1257 classes in @xmath195 . [ remark a4 ] again , we are not claiming that the set @xmath1252 is as small as possible . on quantifier - free formulas , the equivalence @xmath130 is decidable , because @xmath1239 is false if and only if the closed formula @xmath1258 is satisfiable , and the satisfiability problem for existential formulas in prenex normal form is decidable ( see @xcite ) . thus one can modify proposition [ red qf ] by letting @xmath1252 be the set of lexicographically minimal formulas in each @xmath130-class : the same statement of proposition [ red qf ] would then hold with @xmath130 instead of @xmath83 . in particular , the transformation would still be effective , although very inefficient . it is not clear whether the cardinality of the new set of reduced quantifier - free formulas would be significantly smaller . recall that the quantifier depth of a first - order formula is the maximal number of nested quantifiers . if we let @xmath1259 be the set of formulas in @xmath1215 of quantifier depth at most @xmath92 , a formal definition is as follows : @xmath1260 and , for each @xmath1261 , @xmath1262 is the set of boolean combinations of formulas in @xmath1263 using the same recursion , let us define sets of `` reduced '' formulas of every quantifier depth . first we fix an enumeration of the countable set of variables . next , we let @xmath1264 . for each @xmath1265 , we then let @xmath1266 be the set of formulas of the form @xmath1232 where @xmath1267 and the @xmath1234 s are in @xmath1268 let @xmath1270 and @xmath1246 , let @xmath1271 be the cardinality of @xmath1269 , and let @xmath1272 be the cardinality of @xmath1273 . it is elementary to verify that these functions can be bounded as follows : @xmath1274 since there is a procedure to transform each first - order formula into an @xmath83-equivalent formula in `` reduced form '' , we can consider a new equivalence relation on first - order formulas : to yield the same reduced formula . this equivalence is decidable and it refines @xmath83 ( and hence @xmath130 ) . in proposition [ red fo ] , we can still consider replacing each formula by the lexicographically least equivalent formula , but this method is not effective , since the equivalence of first - order formulas is not decidable . a very similar analysis can be conducted for monadic second - order formulas of bounded quantifier depth . one difference is that the lwenheim - skolem theorem does not hold for these formulas , so the semantic equivalence of formulas based on coincidence on all finite or countable models does not imply coincidence on all models . moreover , since there is no complete proof systems for such formulas , the equivalences @xmath130 and @xmath1222 are not semi - decidable . for the rest , one can follow the same techniques as above , to prove the following result . we denote by @xmath1275 the set of monadic second - order formulas of quantification depth @xmath92 in the language of @xmath174-structures , with their first- and second - order free variables in @xmath937 . for every finite @xmath1276 , one can construct a finite subset @xmath1277 of @xmath1275 such that , for every formula in @xmath1275 , one can construct effectively an @xmath130-equivalent formula in @xmath1277 . , the expression of graph properties and graph transformations in monadic second order logic . in : g. rozenberg ( ed . ) _ handbook of graph grammars and computing by graph transformations _ 1 , chapter 5 . world scientific , 1997 , 313 - 400 .	once the set of finite graphs is equipped with an algebra structure ( arising from the definition of operations that generalize the concatenation of words ) , one can define the notion of a recognizable set of graphs in terms of finite congruences . applications to the construction of efficient algorithms and to the theory of context - free sets of graphs follow naturally . the class of recognizable sets depends on the signature of graph operations . we consider three signatures related respectively to hyperedge replacement ( @xmath0 ) context - free graph grammars , to vertex replacement ( @xmath1 ) context - free graph grammars , and to modular decompositions of graphs . we compare the corresponding classes of recognizable sets . we show that they are robust in the sense that many variants of each signature ( where in particular operations are defined by quantifier - free formulas , a quite flexible framework ) yield the same notions of recognizability . we prove that for graphs without large complete bipartite subgraphs , @xmath0-recognizability and @xmath1-recognizability coincide . the same combinatorial condition equates @xmath0-context - free and @xmath1-context - free sets of graphs . inasmuch as possible , results are formulated in the more general framework of relational structures . the recognizability of sets of graphs .5 cm is a robust property .5 cm bruno courcelle , pascal weil
approximately 10% of all quasars are `` radio loud '' , in the sense that they are more luminous at radio than at optical wavelengths . also for some unknown reason , about 10% of radio - quiet quasars show broad absorption lines ( bal ) in their spectra . only recently have a handful of radio - loud bal quasars been discovered ( becker et al . 1997 ; brotherton et al . 1998 ) , thus changing our view that bals only appear in radio - quiet quasars . these quasars will serve as a link between radio - loud and radio - quiet quasars , and will aid in understanding the structure and kinematics of powerful outflows in these quasars . this may eventually lead to a breakthrough in understanding the origin of the radio loudness and the radio - loud / radio - quiet dichotomy . in a spectroscopic survey of a sample of 62 radio - loud quasars , we searched for emission line variations on a time scale of @xmath110 years ( ma & wills 1998 ; ma 2000 ; ma & wills 2001 ) . a bal system was serendipitously discovered in the radio - loud quasar tex 1726@xmath0344 , which showed only narrow absorption lines just 12 years ago . this quasar has a radio flux of 23.0 mjy or a luminosity of @xmath2 w hz@xmath3 at 6 cm ( richards et al . 2001 ) , a redshift of 2.43 , and a v magnitude of 18.5 ( barthel , tytler & thomson 1990 ) . the k - corrected ratio of radio - to - optical power @xmath4 ( sramek & weedman 1980 ; stocke et al . 1992 ; brotherton et al . 1998 ) , commonly used as a measure of radio loudness , is derived to be @xmath5 and is well above the criterion @xmath6 for radio - loud quasars . we observed tex 1726@xmath0344 on 2000 june 4 , using the 2.7-m telescope at mcdonald observatory . a 2-arcsecond slit was used for a total exposure of 5000 seconds . the resolution is about 0.7 nm . the spectrum was corrected with a short exposure using an 8-arcsecond slit . the corrected spectrum is denoted in the upper panel of fig . 1 by a solid line . the historical spectrum taken by barthel et al . ( 1990 ) in 1988 , after scaled to the continuum level of the 2000 spectrum , is indicated by a dotted line . this historical spectrum was taken using the 5-m hale telescope with a 1-arcsecond slit in a 3600-second exposure , and has a resolution of about 0.4 nm . the direct division spectrum and the fitting curve are shown in the lower panel . the continuum and the emission lines match well , as in most of the 62 quasars observed in a time interval of about 10 years . however , historical spectra of tex 1726@xmath0344 in 1988 ( barthel et al . 1990 ) and also in 1990 ( lanzetta et al . 1991 , not shown here ) show only narrow absorption lines . the bals blueward the siiv @xmath7139.7 and civ @xmath7154.9 emission lines in the new spectrum are dramatic , especially in the direct division spectrum . the siiv and the civ absorption lines both have a redshift of 2.36 , as compared to a redshift of 2.43 in the emission lines . the difference in the peaks between the absorption and the emission lines suggests that the absorbing gas has an outflowing velocity of 6000 km / s . both absorption features are significant with restframe equivalent widths of 0.3 nm and 0.2 nm , respectively . the full width at half maximum ( fwhm ) of these bals are approximately 1 nm in the restframe , corresponding to a velocity dispersion over 2000 km / s . the outflowing velocity and velocity dispersion coincide with those of a stellar remnant tidally disrupted by a central massive black hole . it has been suggested that tidal disruptions may be responsible for quasar broad emission lines ( shields 1989 ; roos 1992 ) . a main sequence star , after being tidally disrupted by a @xmath8 black hole in the center of a quasar , will be ejected at a high velocity . the head of the tidal debris will reach a distance of about @xmath9 cm in about 3 years , and the residual velocity will be 6000 km / s at this distance ( roos 1992 ) . the thickness of the tidal debris is stretched at a velocity of @xmath11000 km / s , and the density drops to @xmath10 @xmath11 at this point . the column density of the dense tidal stream is on the order of @xmath12 @xmath13 and is too dense to be responsible for the apparently un - saturated broad absorption lines in tex 1726@xmath0344 . however , as suggested by ma ( 2000 ) , the dense tidal stream is accompanied by faster - moving winds blown off of it due to radiation pressure . both the density and the column density of the wind are two orders of magnitude smaller than those of the dense tidal stream . a model of this piece of gas cloud with a density of @xmath14 @xmath11 and a column density of @xmath15 @xmath13 , at a distance of @xmath9 cm to a central continuum source with a luminosity of @xmath16 w , can thus be studied . we have run a photoionization calculation using cloudy ( ferland 1996 ) , and found that the ratio of civ to total carbon is @xmath17 . assuming a solar abundance with c : h@xmath18 ( grevesse & anders 1989 ) , we find that the column density of civ of this piece of gas cloud is @xmath19 @xmath13 . this value coincides well with that derived from the equivalent width of the civ broad absorption line . column densities for the civ and siiv absorbing ions can be estimated by ( wang et al . 1999 ) @xmath20 where @xmath21 and @xmath7 are the equivalent width and the wavelength , and @xmath22 and @xmath23 are the effective statistical weight and oscillator strength of the corresponding absorption lines , respectively . following wang et al . ( 1999 ) and taking the @xmath22 and @xmath23 values from korista et al . ( 1991 ) , we get the column density of civ and siiv to be @xmath24 and @xmath25 @xmath13 , respectively . we note that absorption line variations with large magnitudes have been observed in the low - redshift radio - quiet quasar pg 1126@xmath26041 ( wang et al . numerous studies on the absorption line variability of seyfert galaxies also exist in the literature ( malizia et al . 1997 ; risaliti , elvis & nicastro 2002 ; akylas et al . 2002 ) , and have been modeled with a single cloud moving through the line of sight ( akylas et al . hamann , barlow & junkkarinen ( 1997 ) have identified variable intrinsic absorption lines ( fwhm@xmath1400 km / s ) in a radio - quiet quasar q2343@xmath0125 . bal variability at a moderate level ( @xmath2740% ) has been found in several radio - quiet quasars ( barlow 1993 ) . in contrast to these previous studies , tex 1726@xmath0344 , a high luminosity _ radio - loud _ quasar developing bals in a time interval of 12 years , represents a new phenomenon . the popular disk - wind model ( murray et al . 1995 ) for both broad emission and absorption lines has been challenged by ma & wills ( 2001 ) because some radio - loud quasars were discovered to have dramatic civ emission line variability , speculated to be due to the illumination of the jet pointing to the polar region of the quasar . the tidally disrupted stars model now seems to be the best candidate . the broad absorption lines in tex 1726@xmath0344 will disappear in just a few years , if they were indeed caused by a tidal disruption event .	we report a serendipitous discovery of broad absorption lines that were newly formed in the spectrum of the high - redshift , luminous radio - loud quasar tex 1726@xmath0344 , in a time interval of only 12 years . this is the first quasar showing a transition from narrow absorption lines to broad absorption lines . it also becomes one of the few radio - loud broad absorption line quasars . the gas cloud responsible for these broad absorption lines is derived to have parameters coinciding with those of the remnant of a tidally disrupted star .
the idea to use _ initial state radiation _ to measure hadronic cross sections from the threshold up to the maximum energy of colliders with fixed energies @xmath1 , to reveal reaction mechanisms and to search for new mesonic states consists in exploiting the process @xmath2 to reduce the centre of mass energies of the colliding electrons and positrons and consequently the energy squared @xmath3 of the final state by emitting 1 or more photons . the method is particularly well suited for the modern meson factories like @xmath4 running at the @xmath5-resonance , _ pep - ii _ , _ kekb _ at the @xmath6-resonance with their high luminosities which compensate for the @xmath7 suppression of the emission of a photon . @xmath4 , _ pep - ii _ and _ kekb _ cover the energy regions up to 1.02 gev and up to 10.6 gev , respectively ( restricted for the latter actually up to 4 ... 5 gev if hard photons are detected ) . a big advantage of the _ isr _ method are the low point - to - point systematic errors of the hadronic energy spectra because the luminosity , the energy of the electrons and positrons and many of the various contributions to the detection efficiencies are determined once for the whole spectrum . as a consequence the overall normalization error is the same for all energies of the hadronic system . the term _ radiative return _ alternately used for _ isr _ refers to the appearance of pronounced resonances ( e.g. @xmath8 ) with energies below the collider energy ( see fig . [ fig : rsigma.eps ] , taken from ref . @xcite ) . reviews and updated results can be found in the proceedings of the international workshops in pisa ( 2003 ) @xcite , novosibirsk ( 2006 ) @xcite and the present ones in frascati ( 2008 ) . calculations of _ isr _ date back to the sixties - seventies of the @xmath9 century . for example photon emission for muon pair production in electron - positron collisions has been calculated in ref . @xcite , for the @xmath10-final state in ref . @xcite , resonances ( @xmath11 ) have been implemented in ref . @xcite , the excitation of @xmath12 and @xmath13 in ref . @xcite , and the possibility to determine the pion form factor was discussed in ref . the application of _ isr _ to the new high luminosity meson factories , originally to determine the hadronic contribution to vacuum polarization , more specifically the pion form factor , has materialized in the late nineties . it has started with calculations of _ isr _ for the colliders @xmath4 , _ pep - ii _ , _ kekb _ @xcite . see also refs . @xcite for calculations of radiative corrections for pion and kaon production below energies of 2 gev . an impressive example of _ isr _ is the _ radiative return _ to the region of the _ z_-resonance at _ lep 2 _ with collider energies around 200 gev @xcite , ( see fig . [ fig : rsigma.eps ] ) . -0.5 cm as function of centre of mass energy @xmath1 , courtesy pdg ref . @xcite , title="fig:",width=264 ] -1.0 cm _ isr _ became a powerful tool for the analysis of experiments at low and intermediate energies only with the development of _ eva - phokhara _ , a monte carlo generator which is user friendly , flexible and easy to implement into the software of the existing detectors @xcite . the generator was applied to an experiment to determine the cross section @xmath14 from the reaction threshold up to the maximum energy for the first time with the detector _ kloe _ at @xmath4 @xcite . the original motivation was and is still the determination of the hadronic contribution to the vacuum polarization from hadronic cross sections . the determination of the hadronic contribution to the vacuum polarization , which arises from the coupling of virtual photons to quark - antiquark - pairs @xmath16 , is possible by measuring the cross section of electron positron annihilation into hadrons with a real photon emitted in the initial state @xmath17 . it is of great importance for the interpretation of the precision measurement of the anomalous magnetic moment of the muon @xmath18 in @xmath19 @xcite and for the determination of the value of the running fine structure constant at the @xmath20 resonance @xmath21 , contributing to precision tests of the standard model of particle physics , see for details _ jegerlehner _ @xcite , also this conference @xcite . the _ novosibirsk _ groups @xmath22 and @xmath23 @xcite pioneered the measurements of hadronic cross sections by changing the collider energy ( _ energy scan _ ) . the _ isr _ method represents an alternate , independent and complementary way to determine those cross sections with different systematic errors . the hadronic vacuum polarization below 2.5 gev can be determined only experimentally because calculations within perturbative _ qcd _ are unrealistic , calculations on the lattice are not yet available with necessary accuracy , and calculations in the framework of chiral perturbation theory are restricted to values close to the @xmath27-threshold . with the maximum energy of 1.02 gev of @xmath4 almost 70 % of the hadronic correction can be determined due to the dominance of the @xmath28resonance . this correction was previously based on the measurement of the @xmath29 reaction of the _ novosibirsk _ groups @xcite by an energy scan . the _ kloe _ collaboration has carried out two analyses : in the so called small photon angle analysis the cross section @xmath30 has been measured between 0.6 and 1.0 gev with a statistical error of less than 0.2 % and a systematical error of 1.2 % . two pions have been detected in the drift chamber in the angular interval @xmath31 with excellent momentum resolution , disregarding the photon . however , an additional constraint for the missing photon angle @xmath32 has been required . as a consequence of this restricted phase space only the @xmath33 region between 0.3 and 1.0 @xmath34 is populated @xcite . the advantage of the small angle analysis is negligible background from resonant ( @xmath5 -)decays ( @xmath35 , @xmath36 and from final state radiation . in the large photon angle analysis the cross section @xmath30 can be measured down to the @xmath27- threshold . in addition to the two pions also photons are detected in the angular interval @xmath37 with an energy of @xmath38 . this analysis is hindered by large resonant background and final state radiation which can be overcome by reducing the energy of @xmath39 below the @xmath5-resonance @xcite . the results of both analyses agree very well . the hadronic correction to the vacuum polarization for @xmath40 obtained in the small angle analysis using the data of 2001 ( @xmath41 in the interval @xmath42 is @xmath43 @xcite . the total systematic error of @xmath44 of @xmath45 includes an experimental systematic error of @xmath46 and a theoretical systematical error of @xmath46 taken in quadrature . an update ( mainly due to a revised _ bhabha _ cross section to determine the luminosity @xcite ) resulted in @xmath47 . using the data of 2002 ( @xmath48 ) the small angle analysis gave @xmath49 with a noticeably reduced statistical and systematic error . in the large angle analysis a value of @xmath50 is obtained for a restricted region @xmath51 covering mainly the @xmath25-resonance to be compared with the result of the small angle analysis of the 2002 data in the same energy interval of @xmath52 @xcite . the main problem of the large angle analysis is the large systematic ( model dependent ) error from the contribution of the @xmath53 from the radiative decay of the @xmath5 according to @xmath54 at low values of @xmath33 . the _ kloe isr _ data and the ( energy scan ) results from _ novosibirsk _ @xcite agree within the error bars ( fig . [ fig : kloecmd2 ] ) . -0.5 cm soon after the application of _ eva - phokhara _ to _ kloe _ @xcite the _ babar _ collaboration also started the investigation of hadron final states with @xmath57 @xcite . in recent years a plethora of final states has been studied , starting with the reaction @xmath58 @xcite . while detecting a hard photon the upper energy for the hadron cross sections is limited to roughly 4.5 gev . final states with 3 , 4 , 5 , 6 charged and neutral pions , 2 pions and 2 kaons , 4 kaons , 4 pions and 2 kaons , with a @xmath5 and a @xmath59 , @xmath60 and 2 pions or 2 kaons , pions and @xmath61 , kaons and @xmath61 , but also baryonic final states with protons and antiprotons , @xmath62 and @xmath63 , @xmath62 and @xmath64 , @xmath65 and @xmath64 , @xmath66 and @xmath67-mesons , etc . have been investigated @xcite . in preparation are final states with 2 pions , 2 kaons . particularly important final states are those with 4 pions which contribute significantly to the muon anomalous magnetic moment and which were poorly known before the _ isr _ measurements @xcite . the _ babar _ results for the @xmath68- final states are not only consistent with previous ones but rather represent the best measurements for @xmath69 , are competitive for @xmath70 , are the best ones for @xmath71 and the only measurement for @xmath72 . detailed analyses allow the identification of intermediate states and consequently the study of the reaction mechanisms . for instance , in the case of the final state with 2 charged and 2 neutral pions @xmath73 the intermediate states @xmath74 and @xmath75 dominate , while also @xmath76 and @xmath77 have been seen . the final state with 5 pions proceeds to about @xmath78 via @xmath79 or @xmath80 and to about @xmath81 via @xmath82 , the rest being @xmath83 . for more details see @xcite . more recently also _ belle _ joined the _ isr _ programme with emphasis on final states containing mesons with hidden and open charm : @xmath60 and @xmath84 , @xmath66 and @xmath67 @xcite . a long standing problem is the nature of the low lying scalar mesons @xmath87 and their decay into pseudoscalars . _ kloe _ has studied the production of @xmath59 in the reaction @xmath88 @xcite . to separate the @xmath59 from the large irreducible background of initial and final state radiation the latter has been calculated by _ phokhara_. a large coupling of the @xmath5 to @xmath59 has been observed and no conflict found with the @xmath89 hypothesis , see also @xcite . many new highly excited states ( preliminarily denoted as _ x , y , z _ ) have been found in recent years . the first of them was discovered by _ babar _ in the reaction @xmath90 @xcite , a ( narrow ) state around 4260 mev with a width of 90 mev , later confirmed by _ belle _ via _ isr _ @xcite and by _ cleo _ in an direct energy scan @xcite . another state was detected at 2175 mev by _ babar _ in the reaction @xmath91 @xcite . _ belle _ found new states at 4080 , 4250 , 4360 , 4660 mev in the reactions @xmath92 and @xmath93 @xcite . the structure of basically all of these new states ( if they will survive ) is unknown so far . 4 quark states , e. g. a @xmath94[\bar{c } \bar{s}]$ ] state for @xmath95 , a @xmath96[\bar{s}\bar{s}]$ ] state for @xmath97 , hybrid and molecular structures are discussed , see also @xcite . also baryonic final states with protons and antiprotons , @xmath62 and @xmath99 , @xmath62 and @xmath100 , @xmath65 and @xmath64 have been investigated in agreement with previous measurements , but with significantly improved statistics . the effective proton form factor shows some nontrivial structures at invariant @xmath101 masses of 2.25 and 3.0 gev , so far unexplained @xcite . being discussed since the sixties / seventies @xmath57 became a powerful tool for experimentalists only with the development of _ eva - phokhara _ by _ j. h. khn , h. czy , g. rodrigo _ and collaborators @xcite , a monte carlo generator developed over almost 10 years , while increasing its complexity , which is user friendly , flexible and easy to implement into the software of the existing detectors . the @xmath57 method was originally thought to be an alternative to energy scans to determine hadronic cross sections from the reaction threshold up to the maximum centre of mass energy of fixed energy meson @xmath102- factories in order to determine the hadronic contribution to the muon magnetic anomaly @xmath15 and to the running fine structure constant @xmath103 , and it was very successful in that respect . but being more than a _ poor man s _ alternative to scan energies it turned out to be an extremely useful tool to clarify reaction mechanisms and to reveal substructures ( intermediate states and their decay mechanisms ) leading to more and more hadronic final states , a plethora at @xmath104-factories . statistical errors have been dramatically reduced . systematic errors are under better and better control . finally it has opened a totally new and unexpected access to highly excited states ( with @xmath0 ) the structure of which is not yet known . h. czy , j. h. khn , eur . j. * c 18 * ( 2001 ) 497 ; g. rodrigo , a. gehrmann - de ridder , m. guilleaume , j. h. khn , eur . j. * c 22 * ( 2001 ) 81 ; g. rodrigo , h. czy , j. h. khn , m. szopa , eur . j. * c 24 * ( 2002 ) 71 ; j. h. khn , g. rodrigo , eur . j. * c 25 * ( 2002 ) 215 ; h. czy , a. grzeliska , j. h. khn , g. rodrigo , eur . j. * c 27 * ( 2003 ) 563 ; h. czy , a. grzeliska , j. h. khn , g. rodrigo , eur . j. * c 33 * ( 2004 ) 333 ; h. czy , j. h. khn , e. nowak , g. rodrigo , eur . j. * c 35 * ( 2004 ) 527 ; h. czy , a. grzeliska , j. h. khn , g. rodrigo , eur . j. * c 39 * ( 2005 ) 411 ; h. czy , a. grzeliska , j. h. khn , g. rodrigo , eur . j. * c 47 * ( 2006 ) 617 ; h. czy , a. grzeliska , j. h. khn , phys . lett . * b 611 * ( 2005 ) 116 ; h. czy , e. nowak - kubat , phys . lett . * b 634 * ( 2006 ) 493 ; h. czy , a. grzeliska , j. h. khn , phys . rev . * d 75 * ( 2007 ) 074026 ; h. czy , a. grzeliska , a. wapienik , acta phys . polon . * b 38 * ( 2007 ) 3491 ; h. czy , a. grzeliska , j. h. khn , hep - ph/0702122 ; h. czy , a. grzeliska , a. wapienik , hep - ph/0710.4227 ; h. czy , j. h. khn , a. wapienik , hep - ph/0804.0359 ; h. czy , these proceedings kloe , hep - ex/0106100 ; hep - ex/0107023 ; nucl . phys . * b 116 * proc . suppl.(2003 ) 243 ; hep - ex/0205046 ; hep - ex/0210013 ; hep - ex/0211024 ; eur . phys . j. * c 33 * ( 2004 ) 656 ; nucl . phys . * b 126 * proc . ( 2004 ) 335 ; aip conf . * 717 * ( 2004 ) 83 ; nucl . * b 131 * proc . suppl.(2004 ) 75 ; nucl * b 135 * proc . suppl.(2004 ) 357 ; int phys . * a 20 * ( 2005 ) 1935 ; acta phys . slov . * 55 * ( 2005 ) 49 ; nucl * b 144 * proc . suppl . ( 2005 ) 231 ; nucl . * b 162 * proc . suppl . ( 2006 ) 81;90;95 ; nucl . * b 169 * proc . suppl . ( 2007 ) 277 , acta phys . * b 38 * ( 2007 ) 3007 ; hep - ex/0707.4078 cmd-2 , phys . lett . * b 527 * ( 2002 ) 161 ; phys . * b 551 * ( 2003 ) 27 ; phys . * b 562 * ( 2003 ) 173 ; phys . lett . * b 578 * ( 2004 ) 285 ; jetp lett . * 82 * ( 2005 ) 743 ; phys . * b 605 * ( 2005 ) 26 ; jetp lett . * 84 * ( 2006 ) 413 snd , phys * d 68 * ( 2003 ) 052006 ; j. exp . * 101 * ( 2005 ) 1053 ; nucl . * b 162 * proc . suppl . ( 2006 ) 11 ; j. exp . * 103 * ( 2006 ) 380 ; j. exp . * 103 * ( 2006 ) 720 ; phys . rev . * d 76 * ( 2007 ) 072012 babar , phys d 70 * ( 2004 ) 072004 ; * d 71 * ( 2005 ) 052001 ; * d 73 * ( 2006 ) 012005 ; * d 73 * ( 2006 ) 052003 ; * d 74 * ( 2006 ) 091103 ; * d 76 * ( 2007 ) 012008 ; * d 76 * ( 2007 ) 092005 ; * d 76 * ( 2007 ) 092006 ; * d 76 * ( 2007 ) 111105	the investigation of events with _ initial state radiation(isr ) _ and subsequent _ radiative return _ has become an impressively successful and guiding tool in low and intermediate energy hadron physics with electron positron colliders : it allows to measure hadronic cross sections and the ratio _ r _ from threshold up to the maximum energy of the colliders running at fixed energy , to clarify reaction mechanisms and reveal substructures ( intermediate states and their decay mechanisms ) and to search for new highly excited mesonic states with @xmath0 . while being discussed since the sixties - seventies _ isr _ became a powerful tool for experimentalists only with the development of _ eva - phokhara _ @xcite , a monte carlo generator developed over almost 10 years , while increasing its complexity , which is user friendly , flexible and easy to implement into the software of existing detectors .
according to the standard theoretical paradigm , the structures observed in the universe today were formed by the gravitational amplification of small perturbations in an initially gaussian dark matter density field . small scale overdensities were the first to collapse , and the resulting objects subsequently merged under the influence of gravity to form larger structures such as groups and clusters of galaxies . galaxies formed within dense _ halos _ of dark matter , where gas was able to reach high enough overdensities to cool , condense and form stars . the quiescent cooling of gas within a dark matter halo results in the formation of a rotationally - supported disk system at the centre of the halo . when halos merge with each other , a bound group of galaxies is produced . dynamical friction will cause the orbits of the group members to erode over time , and the galaxies to spiral in towards the centre of the halo and merge . galaxy - galaxy mergers are thus inevitable in this picture . n - body simulations have demonstrated that mergers between disk galaxies of near - equal mass result in the formation of remnant systems that are structurally very similar to observed elliptical galaxies ( e.g. barnes & hernquist 1996 ) . elliptical galaxies are a very homogeneous class of objects . their stellar populations are old and in addition there is a well - defined correlation between colour and absolute magnitude that exhibits a remarkably small scatter ( bower , lucey & ellis 1992 ) . there has been some doubt as to whether the old ages of ellipticals and the small scatter of the colour - magnitude relation can be explained in a scenario in which all ellipticals form by mergers of spirals . in order to address this issue , one must follow the formation and evolution of elliptical galaxies in the framework of the theoretical scenario described above . in semi - analytic models of galaxy formation ( see for example kauffmann , white & guiderdoni 1993 ; cole et al 1994 ; somerville & primack 1999 ) , an algorithm based on the extended press - schechter theory is used to generate monte carlo realizations of the merging paths of dark matter halos from high redshift to the present . these `` merger trees '' are specified for a chosen set of cosmological initial conditions and allow the progenitors of a present - day dark matter halo to be traced back to arbitrarily early times . simple recipes are introduced to describe the cooling of gas within the halos , the formation of stars from the cold gas , feedback and heavy element enrichment from supernova explosions , and the merging rates of galaxies . the models are then coupled to a stellar population synthesis code in order to generate magnitudes and colours that can be compared directly with observational data . if two galaxies of comparable mass merge , the stars of both objects are added together to create a bulge component . when a bulge is formed by a merger , the cold gas present in the two galaxies is transformed into stars in a `` starburst '' with a timescale of @xmath1 years . further cooling of gas in the halo may lead to the formation of a new disk . the morphological classification of galaxies is made according their b - band disk - to - bulge ratios ( simien & de vaucouleurs 1986 ) . if @xmath2 mag , then the galaxy is classified as early - type ( elliptical or s0 ) . it is important to define carefully what is meant by the `` formation time '' of an elliptical galaxy . the solid line in fig . [ fig : sfh ] shows the _ average _ star formation history of a present - day cluster elliptical galaxy . results are shown for a high - density cold dark matter ( cdm ) cosmology with @xmath3 , @xmath4 km s@xmath5 mpc@xmath5 and @xmath6 . as can be seen , most of the stars form at redshifts greater than 2 . because the stars form early , the small scatter of the observed colour - magnitude relation is not a problem for the model ( kauffmann 1996 ) . the dotted line shows the distribution of the time of the last major merger . this is typically much later than the epoch at which the stars form ; in this cosmology , many elliptical galaxies have their last major mergers at redshifts less than 1 . the epoch of the last major merger also depends on the choice of cosmological parameters . in a low density model ( @xmath7 ) , the epoch at which structure forms is shifted to higher redshift . at z=1 , more than @xmath8 of present - day bright ellipticals are already present in a cdm model with @xmath9 and @xmath10 ( kauffmann et al 1999 ) , but most have not yet formed by @xmath11 . finally , the formation time of an elliptical depends on its mass . more massive objects form later . because the ratio of gas to stellar mass in galaxies decreases at late times , massive ellipticals form from gas - poor progenitors and form a smaller fraction of their stars in the merger - induced starburst ( kauffmann & haehnelt , in preparation ) . the probability of merging peaks in environments where the relative velocities between galaxies are comparable to their internal stellar velocity dispersions . in the semi - analytic models , bright ( @xmath12 ) elliptical galaxies are formed in dark matter halos with circular velocities in the range 250350 km s@xmath5 . these newly - formed ellipticals are _ isolated _ galaxies , with no bright companion within a radius of @xmath13 mpc . elliptical galaxies in groups and clusters are predicted to be significantly older than isolated ellipticals . in the standard cdm model , bright ellipticals in clusters and groups have @xmath14-luminosity weighted mean stellar ages that are 34 gyr older than isolated ellipticals ( kauffmann 1996 ) . the fact that galaxies form earlier in high density environments is a direct consequence of the statistical properties of the peaks of gaussian random fields ( bardeen et al 1986 ) and is a prediction of _ all _ models based on the standard theory . another quantity of interest in the number of progenitor galaxies that merge together to form an elliptical . kauffmann & charlot ( 1998a ) show that there is a weak trend in the number of progenitors as a function of the mass of the elliptical : e galaxies with mass @xmath15 typically form from 23 progenitors , whereas e galaxies more massive than @xmath16 form from 57 progenitors . because the number of progenitors only increases by a factor of @xmath17 over more than an order of magnitude in mass , these results also imply that massive es also form from the merging of _ more massive progenitors_. as discussed by kauffmann & charlot ( 1998a ) , this provides an explanation for why a correlation between mass and metallicity is preserved in a scenario where ellipticals form via the mergers of disk galaxies . in the models , massive disks are metal rich because the energy released from supernova explosions is not sufficient to eject enriched out of the galaxy . in low - mass disks , most of the metals escape into the surrounding halo and the galaxy is metal - poor . the massive , metal - rich disks then merge to form the most massive ellipticals . the most direct test of the merger picture is the prediction that the global abundance of bright elliptical galaxies decreases at high redshift . in all currently popular theoretical models , the abundance of ellipticals decreases by at least a factor 3 by @xmath18 . because ellipticals dominate the bright end of the @xmath19-band luminosity function , kauffmann & charlot ( 1998b ) have suggested that @xmath19-band selected redshift surveys covering a wide area of the sky to a limiting magnitude of @xmath20 would a very effective means of testing the merger scenario . testing the merger picture using ellipticals in clusters at high redshift is much more tricky because of the inherent biases predicted by the theory . in particular , the dense central regions of the most massive clusters always contain the oldest galaxies at any redshift . kauffmann & charlot ( 1998a ) demonstrate that the apparent `` passive evolution '' observed for ellipticals in clusters out to @xmath21 ( stanford , eisenhardt & dicksinson 1998 ) is quite consistent with the predictions of a high - density cdm model . the agreement comes about not because ellipticals are passively evolving , but because the observational selection procedure favours the oldest systems . it is also possible to test the merger picture by looking for trends in the ages of elliptical galaxies as a function of environment . this test has been done at low redshift using a range of different stellar age indicators . rose et al ( 1994 ) , using a variety of different stellar absorption indices , and mobasher & james ( 1996 ) , using measurements of the co absorption feature at 2.3 mm , find evidence that ellipticals in low - density environments contain a substantial intermediate age population . bernardi et al ( 1998 ) , using the mg@xmath22 index , find a smaller difference between cluster and field ellipticals . while colours and balmer line strengths of galaxies are used to infer ages of stellar populations , @xmath0/fe element ratios carry information about the timescale over which star formation took place . the reason for this lies in the delayed fe enrichment from type ia supernovae , the progenitors of which are believed to be low - mass and hence long - lived binary stars . models that incorporate this delay succeed in reproducing the abundance patterns in the solar neighbourhood ( e.g. greggio & renzini 1983 ; matteucci & greggio 1986 ) . although we can not measure abundance ratios in elliptical galaxies directly , the analysis of mg and fe line indices in ellipticals ( burstein et al 1984 ) using population synthesis models indicates that the ratio mg / fe in these objects is significantly super - solar by 0.20.3 dex ( peletier 1989 ; worthey , faber & gonzlez 1992 ) . subsequent studies confirm that this so - called @xmath0-enhancement in the stellar populations of ellipticals extends out to the effective radius ( davies , sadler & peletier 1993 ; carollo & danziger 1994 ; fisher , franx & illingworth 1995 ; mehlert et al 1999 ) . in a more quantitative study using @xmath0-enhanced stellar tracks , weiss , peletier & matteucci ( 1995 ) derive @xmath23\leq\-0.7 $ ] dex and in the bulge of our galaxy , mcwilliam & rich ( 1994 ) find metal - rich stars having super - solar o / fe and mg / fe ratios . depending on the slope of the assumed initial mass function ( imf ) , such values require star formation timescales around @xmath1@xmath24 years ( e.g. matteucci 1994 ; thomas , greggio & bender 1999 ) . in models of hierarchical galaxy formation , star formation in ellipticals typically does not truncate after 1 gyr , but continuous to lower redshift ( kauffmann 1996 ) . it is therefore questionable if hierarchical clustering would lead to significantly @xmath0-enhanced giant ellipticals ( bender 1997 ) . thomas et al ( 1999 ) discuss the abundance ratios associated with mergers of evolved galaxies , and show that a late merger is incapable in producing significantly super - solar mg / fe ratios in the stellar population . this result agrees qualitatively with a study by sansom & proctor ( 1998 ) , who show that mg / fe ratios are lower when the object forms by subsequent mergers . so far , semi - analytic models have not considered this constraint . in a preliminary study , thomas ( 1999 ) demonstrated that the _ average _ star formation history ( hereafter sfh ) of cluster ellipticals predicted by the models ( see fig . [ fig : sfh ] ) led to @xmath25}\sim 0.04 $ ] dex , a value well below the observational estimate . in the present paper , we use the _ individual _ sfhs given by the models to explore the distribution of mg / fe values in both cluster and field galaxies of different morphological types and luminosities . the prescription of the chemical evolution is explained in thomas ( 1999 , and references therein ) . the chemical evolution code , particularly the supernova rates of both types , have been calibrated on the abundance patterns in the solar neighbourhood ( thomas , greggio & bender 1998 ) . the semi - analytic sfhs are input directly into this code and include both quiescent star formation in the disk galaxy progenitors as well as the starburst triggered by the merger . note that the chemical evolution models assume that galaxies evolve as single - body , `` closed - box '' systems and this is not the case in the hierarchical models , where supernova feedback effects are thought to be important in driving chemical evolution . feedback effects could either drive the derived @xmath0/fe ratios up or down , depending on the relative efficiencies with which the products of type ii and type ia supernovae are expelled . we plan to explore this in detail in future work . we also assume a universal salpeter imf slope @xmath26 . in the following , @xmath27}\equiv\log({\rm\-mg / fe})-\log({\rm\-mg / fe})_{\odot}$ ] denotes the global element ratio in the _ stellar population _ of a individual galaxy . it should be emphasized that our plots do not show mg / fe ratios for the stars inside individual galaxies , but the variation obtained from one galaxy to another . we also only consider element ratios at redshift zero . the evolution of mg / fe with redshift is discussed in thomas ( 1999 ) . we show results for a cold dark matter power spectrum with @xmath3 , @xmath28 km s@xmath5 mpc@xmath5 , and @xmath29 . [ fig : distrib ] shows the number distribution of mg / fe ratios among galaxies of different morphologies in environments of varying density . these distributions are computed by averaging over the galaxies in 15 halos with @xmath30 km s@xmath5 ( clusters ) , 50 halos with @xmath31 km s@xmath5 ( groups ) , and 100 halos with @xmath32 km s@xmath5 ( small groups ) . the mg / fe values generally scatter between @xmath33}\leq 0.16 $ ] dex , with a mean of @xmath34@xmath35 dex , significantly below the observed estimate quoted above . the spread in mg / fe results from the scatter in star formation timescales in the models ( kauffmann 1996 ) . late type galaxies experience more extended sfhs . as a result , these objects have younger stellar populations and exhibit lower mg / fe ratios , in qualitative agreement with observations . although elliptical galaxies have the highest values of mg / fe , they still do not match the observed @xmath0-enhancement . only oldest galaxies , which form most of their stars in the first 12 gyr , reach mg / fe close to the lower limit set by observations . lenticular galaxies are expected to be slightly less @xmath0-enhanced than ellipticals . galaxies in small halos form their stars over a more extended timescale than galaxies in clusters , where the supply of infalling gas is cut off as soon as the galaxy is accreted . as a result , lower mg / fe values are found for galaxies in small groups . in clusters , the formation is boosted towards higher redshift and more galaxies form on short timescales . bright , isolated ellipticals are thus predicted to have lower mg / fe values than their counterparts in clusters ( see also thomas 1999 ) . kauffmann & charlot ( 1998a ) find that the @xmath14-light weighted age difference between bright cluster and `` field '' ellipticals is @xmath13 gyr . this is in good agreement with the estimate by bernardi et al ( 1998 ) , who find that variations in zero - point and slope of the mg-@xmath36 relation of field and cluster ellipticals translate into an age difference of @xmath37 gyr . their study lacked information from fe line indices and they could not determine whether there were trends in @xmath0-enhancement from cluster to field . more work in this area would be very valuable . faint ellipticals ( @xmath38 ) exhibit a scatter in mg / fe over a range from @xmath39 dex . both the scatter and the median value decrease for brighter ellipticals because these objects form later and have younger mean stellar ages . according to kauffmann & charlot ( 1998a ) , even though bright ellipticals are younger than faint ones , they still have redder colours because they are more metal - rich . worthey et al ( 1992 ) claim , however , that the fe index remains constant for elliptical galaxies of different mass . the tight correlation between mg index and velocity dispersion @xmath36 ( bender , burstein & faber 1993 ) then implies that mg / fe is higher in more massive ellipticals . this is in clear disagreement with the trend shown in fig [ fig : lum ] . note that if the fe index is a reasonable tracer of metallicity , the results of worthey et al imply that both the mg-@xmath36 relation and the colour - magnitude relation may simply be an effect of an increase in @xmath0-element enhancement as a function of the mass of the elliptical . as we have shown , this is not easily accommodated in hierarchical models , because it is the _ low mass _ ellipticals that form on short timescales . another interpretation is that fe indices are not viable metallicity indicators ( gonzlez 1993 ) , and the super - solar @xmath0/fe ratios in luminous ellipticals actually result from an fe deficiency ( buzzoni , mantegazza & gariboldi 1994 ) . it may be possible to achieve this in models where metal ejection is treated in detail . bulges are the first spheroids to form in major mergers at high redshift ( kauffmann 1996 ) . depending on the amount of gas accreted subsequently , these spheroids may become the bulges of spiral galaxies . as discussed previously , the classification into morphological type is made according to disk - to - bulge ratios , so even ellipticals have a minor disk component . [ fig : bulges ] shows the distribution of mg / fe among the _ bulge components _ of galaxies of different types in a variety of environments . the drop of the mg / fe ratio from high to low density environment is less pronounced in the bulge component . another striking result is that the bulges of late - type galaxies are more @xmath0-enhanced and exhibit a much smaller scatter in mg / fe . this result is expected in a scenario where the bulge forms first and the disk slowly accretes over a hubble time . if the bulge is formed from the disk , one would not expect to see such an effect . it would be interesting to study mg and fe line indices in the bulges of galaxies of different morphological type in order test some of these trends ( see trager , dalcanton & weiner 1999 ) . in fig . [ fig : bg ] we plot the global value of mg / fe versus the mg / fe of the bulge component for different types of galaxies in a rich cluster . the value of mg / fe is always higher in the bulge than it is on average and this difference increases for late type galaxies . this result is interesting because the bulge component is expected to be more spatially concentrated , which implies a radial gradient in mg / fe inside the galaxy . in elliptical galaxies , observations indicate mg / fe to be constant with radius inside the half - light radius ( worthey et al 1992 ; davies et al 1993 ; fisher et al 1995 ) . \1 ) although we have adopted the sfhs predicted by the semi - analytic models , we have not yet treated chemical evolution in consistent way . inclusion of metal ejection will allow us to study abundance ratios in the intra - cluster medium ( e.g. mushotzky et al 1996 ) \2 ) indices are measured close to the centres of the galaxies where the starburst population may dominate the light ( barnes & hernquist 1996 ) . the quantities calculated in this work are for the global stellar population . we also plan to explore two - component models where chemical evolution in the burst is treated separately . \3 ) we assumed a universal salpeter imf . it would be interesting to study the effect of flattening the imf of the stars that form in the burst , which would increase production of @xmath0-elements ( see thomas 1999 ) . \4 ) a more quantitative comparison between model predictions and observables can be obtained by transforming the calculated mg / fe element ratios to line indices on the basis of mg / fe - dependent fitting functions and stellar tracks . \5 ) the choice of cosmology has an impact on the sfhs and therefore also on the resulting abundance ratios . in particular , in low-@xmath40 universes star formation is pushed towards higher redshift yielding higher mg / fe . trager , s. c. , dalcanton , j. j. , & weiner , b. j. 1999 , in when and how do bulges form and evolve , c. m. carollo & h. ferguson , stsci workshop , astro - ph/9810376 weiss , a. , peletier , r. f. , & matteucci , f. 1995 , a&a , 296 , 73	in models of galaxy formation in a hierarchical universe , elliptical galaxies form through the merging of smaller disk systems . these models yield a number of testable predictions if reliable techniques for determining the relative ages and compositions of the stellar populations of different galaxies can be found : 1 ) ellipticals in low - density environments form later than ellipticals in clusters , 2 ) more massive ellipticals form later , 3 ) more massive ellipticals form in dissipationless mergers from disk galaxies with low gas content . while colours and the balmer line strengths of galaxies can be used to infer the average ages of the stellar populations of ellipticals , @xmath0/fe element ratios carry information about the timescale over which star formation took place . here we present preliminary results from semi - analytic models for the distribution of mg / fe ratios in galaxies as a function of morphological type , luminosity and environment .
unlike stars , galaxies show diverse shapes . the belief that the difference in appearance reflects the generic character has motivated morphological classification of galaxies . since the 1920s the simple scheme suggested by hubble ( hubble 1926 ) based on single band images of bright galaxies has been widely adopted . the essence of hubble s scheme and its elaborations ( de vaucouleurs 1991 ; kormendy 1979 ; de vaucouleurs 1959 ; sandage 1961 ) is to divide galaxies into early and late types . the early types are further divided into ellipticals and lenticulars , and the late types into spirals ( unbarred and barred ) and irregulars . any one who looks directly into galaxy images immediately realize that shapes of galaxies are much more diverse than this , but also that the simple hubble sequence catches the major features of galaxy morphology . as unprecedentedly large sets of digital images of galaxies such as the sloan digital sky survey ( sdss , full details of the sdss are available at http://www.sdss.org . york 2000 ) data are available , time is ripe for understanding the origin of morphology of galaxies . in order to find the accurate relationship of galaxy morphology with other physical properties it is required to define morphological types more objectively and to use morphology subsets as homogeneous as possible . the endeavors to develop an objective , automatic and highly successful morphology classifier applicable to large libraries of digital images have resulted in various schemes ( yamauchi 2005 ; blanton 2003@xmath4 ; abraham , van den bergh , & nair 2003 ; strateva 2001 ; shimasaku 2001 ; abraham 1996 ; doi , fukugita , & okamura 1993 ) . this work further advances this topic . the philosophy of our morphological classification is to develop a scheme applicable down to the faint magnitude limit ( @xmath5 where @xmath6 is the petrosian magnitude after galactic extinction correction ) of the spectroscopic sample of the sdss using only photometric information . the fundamental features of galaxy morphology still useful for galaxies near the faint limit are the surface brightness and color profiles . their mean levels are measured by mean surface brightness and integrated color . their radial variation scaled to a characteristic size can be measured by the sersic or concentration index , and color gradient , for example . therefore , in principle , the morphology classification of galaxies should be made at least in 4-dimensional parameter space . the surface brightness at long wavelength bands represents the stellar mass distribution , and the color tells about the recent star formation history . a difficulty in dividing galaxies into early and late types using the surface brightness profile alone lies in the fact that most spirals consist of both bulge and disk . there is an unavoidable confusion between the bulge - only and the bulge - plus - disk systems when only the surface brightness information in a single band is used for classification . contamination in the early and late type subsets separated by using the concentration index , for example , is typically about 20% ( yamauchi 2005 ; shimasaku 2001 ) . the surface texture is of limited usage because it is sensitive to seeing and is simply lost for small faint galaxies ( @xmath7 in the case of sdss ; yamauchi 2005 ) . on the other hand , the information in color reveals the additional generic difference between early and late type galaxies which have experienced different star formation histories . strateva ( 2001 ) has found that the integrated ( observer frame ) @xmath8 color ( asterisks are attached to the sdss photometry based on the photometric equations used through the early data release ) of the sdss main galaxies ( stauss 2002 ) shows a bimodal distribution . however , they have shown that , when divided at @xmath9 , the early ( e / s0/sa ) and late type ( sb / sc / irr ) subsets have significant contaminations reaching about 30% for a sample with visually identified morphological types . in this paper we extend their work by incorporating color gradient and the concentration index as additional dimensions of the classification parameter space . we use a training set of 1982 galaxies to quantify the performance of our morphology classification scheme . the training set consists of two samples . the first contains @xmath10 sdss galaxies with @xmath11 after galactic extinction correction whose morphological types are assigned by the authors based on the color images retrieved by the sdss image list tool ( http://cas.sdss.org/astro/en/tools/chart/list.asp ) . galaxies are divided into six types : e / s0 ( smooth brightness and color profile ) , s0 ( disk with features , but no spiral arm ) , ae ( e or s0 with abnormal brightness or color distribution ) , s ( spiral arm , disk with dust lane ) , as ( spirals with large distortion in shape ) , i ( irregulars ) . the first three correspond to the early type , and the rest is the late types . the second training set is fukugita s catalog ( in preparation ; see also nakamura 2003 ) listing @xmath12 galaxies with @xmath13 ( after galactic extinction correction ) and with visually identified morphological types from @xmath14 to @xmath15 at half integer steps . among them we use a subset of @xmath16 galaxies that are listed both in the main galaxy and the spectroscopic sample and are not too much contaminated by foreground stars . by comparing the morphological types of @xmath17 galaxies common in our sample and fukugita s catalog , we have found that the early and late types divide at @xmath18 ( @xmath19 and 2 correspond to hubble type s0 and sa , respectively ) in the latter catalog . we use the color versus color gradient space as the major morphology classification tool , and use the concentration index as an auxiliary parameter . in the color - color gradient space spirals tend to separate from the region clustered by ellipticals and lenticulars as majority of spirals exhibit significant color gradient . as a measure of color gradient , we adopt the difference in @xmath20 color of the region with @xmath21 from that of the annulus with @xmath22 , where @xmath23 is the petrosian radius and @xmath20 is a rest frame @xmath24 color @xmath25-corrected to the redshift of 0.1 ( hereafter , @xmath1 means @xmath20 ) , and @xmath25-correction of magnitude and color is calculated based on the study of blanton et al . ( 2003@xmath26 ) . negative color difference means bluer outside . model magnitude is used for color . it is the magnitude calculated from the best - fit model profile obtained by fitting the de vaucouleur and exponential profiles to the galaxy image . we have chosen the @xmath27- and @xmath28-bands to estimate color gradient because they are widely separated in wavelength across the 4,000-@xmath29 break , and have @xmath30 ratios higher than the @xmath31- and @xmath32-bands . choice of other bands results in nosier color gradient estimates at faint magnitudes , or less successful separation of galaxies into early and late types . other measures of color gradient like the linear slope of the radial @xmath1 color profile and the @xmath1 color difference between the region with @xmath33 and the annulus with @xmath34 , are also calculated for comparisons . @xmath35 and @xmath36 are the radii from the center of a galaxy containing 50 and 90% of the pertrosian flux . we have found that the clump of the normal early type galaxies in the color - color gradient space is less tight for the latter measures at fainter magnitudes @xmath37 . throughout this paper we use ab magnitudes converted from sdss magnitudes . the @xmath27- and @xmath28-band atlas images and basic photometric parameters of individual galaxies are retrieved from the sdss data release 3 ( dr3 ) at princeton ( http://photo.astro.princeton.edu ) . we have generated a big set of sersic model images convolved with the psf of various sizes . these images are used to find a best - fit sersic model with true sersic index and true axis ratio for an observed image of a given size of the psf . the fitting is made at radii larger than @xmath38 to avoid the central region whose profile is much affected by seeing . then the best - fit sersic models in @xmath27- and @xmath28-bands are convolved with the psf of the same size , and the effects of seeing on the concentration index and color difference are estimated . we have confirmed the seeing - corrected concentration index and color difference show no apparent dependence on seeing . we use the elliptical annuli in all our parameter calculations to take into account flattening or inclination of galaxies . the position angle and axis ratio used to define the elliptical annuli and to calculate the radial surface brightness profile are isophotal ones measured from the @xmath28-band image . the petrosian radii in our analysis are usually larger than those in the dr3 catalog which adopted circular annuli . the photometric parameters we use for morphology classification are the @xmath0 color , color difference @xmath39 , and the ( inverse ) concentration index @xmath40 in the @xmath28-band , where @xmath41 . here , the concentration index is used as a complementary parameter to discriminate red disky spirals from early type galaxies . the @xmath28-band image is used to measure the index because the image at longer wavelengths is expected to better represent the stellar mass distribution . all parameters are seeing - corrected as described above . figure [ fig1 ] shows the distribution of galaxies of this training set in the @xmath0 versus @xmath39 space and in the @xmath0 versus concentration index space . the early ( e / s0 ) and late types ( sa to sd ) determined by visual inspection are marked as circles ( red ) and crosses ( blue ) , respectively . squares ( green ) are irregulars . in our automated scheme classified as early type galaxies are those lying above the boundary lines passing through the points @xmath42 , @xmath43 and @xmath44 in the @xmath0 versus @xmath39 space . they are also required to have @xmath45 . the rest are classified as late types . the completeness of this classification scheme reaches @xmath46 for early types and @xmath47 for late types . the reliabilities are @xmath47 and @xmath48 , respectively . the completeness , @xmath49 , is the fraction of galaxies of a given type that are successfully selected from the original sample by the classification scheme . the reliability , @xmath50 , is the fraction of galaxies of the desired type from the selected subsample . the @xmath51 failure is due to a few red passive spirals , galaxies with companions or foreground stars , and to incorrect visual classification ( for edge - on objects in particular ) . the parameter @xmath40 does only an auxiliary role improving the results by a few percents . we estimate the completeness and reliability of our morphology classification is about @xmath52 near @xmath53 , which is close to the faint end of the spectroscopic sample of the sdss . the performance of our classification scheme slowly degrades at @xmath54 because it becomes harder to measure the color difference and concentration index as the size of galaxies relative to the ccd pixel and the fwhm of the psf decreases . to estimate the performance of our classification scheme at fainter magnitudes we redshift the galaxies in the training sample ( @xmath55 ) with morphological types assigned by us . we calculate the new redshift at which a galaxy appear dimmer by a desired magnitude adopting a cosmology with density parameters @xmath56 and @xmath57 . the reduction factor in angular size is calculated and images are binned down ( giavalisco et al . the resulting images are convolved with the psf in such a way that they have the seeing effects equal to those of the original ones . noises are not added anew . following this procedure we generated three mock morphology samples of @xmath10 galaxies in half magnitude bins from @xmath58 to @xmath59 . the distributions of these redshifted galaxies in the color - color gradient space and in the color - concentration index space are very similar with those of three samples of randomly - drawn @xmath60 sdss galaxies with actual magnitudes in the same bins , thus justifying our procedure . we determine the classification criteria at fainter magnitudes using these redshifted morphology samples . results are summarized in table [ table1 ] . table 1 says that our classification boundaries in the color - color gradient space hardly change even at magnitudes down to @xmath53 . the only major change is the boundary at the reddest colors ( the line connecting to the point p3 ) . at fainter magnitudes the clump of the normal e / s0 s expands vertically in the color - color gradient space because the measured color - differences have more errors . the classification boundary at the reddest colors has been chosen to include more ellipticals taking into account this fact . we think the @xmath25-correction of color in this reddest color range has been accurately estimated . our @xmath25-correction might have more errors for intermediate - type spirals for which the spectra of the central part of galaxies are less representative of the total color . but in this case our classification is hardly affected because these galaxies are blue , have large color gradients , and are located far from the classification boundaries . inspection of location of galaxies in the color - color gradient space leads us to classification beyond the dichotomous division into early and late types . normal ellipticals and lenticulars are very homogeneous systems whose light is dominated by old population ii stars . they strongly concentrate within a spot centered at @xmath61 in the @xmath0 and @xmath39 plane . dispersion is mainly due to companions or foreground stars . they show very weak but definite color gradient ( , outside is bluer ) . we have discovered a trail of early types containing about 10% of early type galaxies toward the left hand direction @xmath62 from the concentration of normal early types and then upwards . the galaxies in this trail are bluer than normal ones , often show emission lines , and occasionally show star burst activity near the center . they tend to be more centrally concentrated than spirals at the same color . we have found that most @xmath63 galaxies ( dressler & gunn 1983 ) listed in yamauchi & goto ( 2005 ) are located in this trail . on the other hand , a spiral branch is extended downward of the cluster of normal early types . sa type spirals often start to have star forming zones at the outskirt of disk . their integrated color is still dominated by the red light from their bulges , but they begin to show color gradient . as the star formation activity becomes stronger , color gradient increases first and then starts to decrease as the star forming region expands toward the center . very late type spirals and most irregulars at the upper left corner are very blue and have inverted color gradient ( , center is bluer ) since they usually show strong star burst near the center . the spiral branch meets the trail of early types there . the major advantage of morphology classification in the color - color gradients space is that the blue early type galaxies are separated from the spiral galaxies because the former tend to have nearly constant color profile or bluer cores while the latter tend to have red bulge plus blue disk structure . the spiral branch is a sequence of locality of star formation activity and cold gas and dust concentration as illustrated in fig . the position of a spiral along the branch has a strong correlation with its morphological subclass as shown in fig . [ fig3 ] . after making early - late type division , one can further divide galaxies into subclasses based on their location along the early and late type branches . for example , late type galaxies can be divided into @xmath64 , @xmath65 , and @xmath66 subclasses which group galaxies lying above the line @xmath67 , between this line and the line @xmath68 , and left of the latter line , respectively . these subclasses roughly correspond to sa - sb , sb - sc , and sc - sd / irr , respectively . dispersion of the spiral branch is partly due to internal extinction which tends to make @xmath0 and @xmath69 increase . dependences of @xmath0 and @xmath69 on axis ratio of galaxies indicate that the effects of internal extinction are strongest for intermediate type spirals . our new morphology classification scheme takes into account history and locality of star formation as well as stellar mass distribution . it is easy to implement the scheme for multi - band surveys with wide spectrum coverage . we expect that the strong concentration of early types and the general shape of the spiral branch in the color versus color gradient space are still maintained for galaxies at high redshifts . any change in their location and shape will be colorful evidence for evolution of star formation activity and cold gas infall into galaxies with different morphology . cbp thanks michael vogeley for collaboration on parts of this work and for invitation to drexel university when this work was started . we also thank yasushi suto for suggesting us to use fukugita et al.s catalog . we thank the anonymous referee for helpful comments . this work is supported by the korea science and engineering foundation ( kosef ) through the astrophysical research center for the structure and evolution of cosmos ( arcsec ) . abraham , r. g. , 1996 , , 107 , 1 abraham , r. g. , van den bergh , s. , & nair , p. a. 2003 , , 588 , 218 blanton , m. r. , 2003@xmath4 , , 594 , 186 blanton , m. r. , 2003@xmath26 , , 125 , 2348 de vaucouleurs , g. 1959 , handbuch der physik , 53 , 275 de vaucouleurs , g. , 1991 , in third reference catalogue of bright galaxies , ( new york : springer ) doi , m. , fukugita , m. , & okamura , s. 1993 , , 264 , 832 dressler , a. , & gunn , j. e. 1983 , , 270 , 7 giavalisco , m. , livio , m. , bohlin , r. c. , macchetto , f. d. , & stecher , t. p. 1996 , aj , 112 , 369 hubble , e. p. 1926 , , 64 , 321 kormendy , j. a. 1979 , , 227 , 714 nakamura , o. , 2003 , , 125 , 1682 sandage , a. 1961 , in the hubble atlas of galaxies ( washington : carnegie institution ) shimasaku , k. , 2001 , , 122 , 1238 stoughton , c. , 2002 , , 123 , 485 strateva , i. , 2001 , , 122 , 1861 strauss , m. a. 2002 , , 124 , 1810 yamauchi , c. , & goto , t. 2005 , , 359 , 1557 yamauchi , c. , 2005 , ( in press ) york , d. , 2000 , , 120 , 1579 the three points , @xmath70 , @xmath71 , and @xmath72 , define the lines dividing the sdss galaxies into the early and late types in the @xmath73 versus @xmath74 plane . the early types are also required to have @xmath75 .	we have found the @xmath0 color versus @xmath1 color gradient space can be used for highly successful morphology classification of galaxies in the sloan digital sky survey . in this space galaxies form early and late type branches well - separated from each other . the location of galaxies along the branches reflects the degree and locality of star formation activity , and monotonically corresponds to the sequence of morphological subclasses . when the concentration index is used together , the completeness and reliability of classification reaches about 91% for a training set of sdss galaxies brighter than @xmath2 . at faintest magnitudes ( @xmath3 ) of the sdss spectroscopic sample , the performance still remains at about 88% . the new classification scheme will help us find accurate relations of galaxy morphology with spatial and temporal environments , and help one understand the origin of morphology of galaxies .
quantum - mechanical many - body problems can almost never be solved exactly . in this situation variational wave functions have proved to be particularly useful . although they describe correlations among the particles only in an approximate way , they have the advantage of being explicit and physically intuitive . in particular , they allow for investigations even when standard perturbation theory is not applicable , or is untractable . variational wave functions can , for example , be obtained by applying a suitably chosen correlation operator ( e. g. , the interaction part of the hamiltonian under investigation ) to a simple one - particle wave function . for the one - band hubbard model@xcite @xmath8 which is often used as an effective model to understand electronic correlation phenomena like itinerant ferromagnetism in transition metals , high - temperature superconductivity and the mott - hubbard metal - insulator transition , the simplest projected wave function is the gutzwiller wave function ( gwf)@xcite @xmath9\|\phi_{0}\rangle , \,\label{ansatz } \end{aligned}\ ] ] where @xmath4 is a variational parameter ( usually @xmath10 , @xmath11 is the operator for double occupation at lattice site @xmath12 , and the starting wave function @xmath13 is a product state of spin - up and spin - down fermi seas @xmath14 using the gwf one may , in principle , calculate expectation values of any operator @xmath15 as @xmath16 @xmath2 @xmath17 . by the variational principle the energy expectation value @xmath18 is an upper bound for the true ground - state energy of @xmath19 . the properties and quality of the gwf have been subject of detailed investigations ( for an early review see ref . ) . a diagrammatic theory for the calculation of expectation values in terms of the gwf , valid in arbitrary dimensions @xmath1 , was formulated by metzner and vollhardt.@xcite ( ref . is hereafter referred to as mv . ) in particular , for systems without net magnetic polarization @xmath0 @xmath2 @xmath20 ( i. e. , particle densities @xmath21 , with @xmath22 in eq . ( [ startwf ] ) ) they calculated the momentum distribution @xmath23 @xmath2 @xmath24 and the double occupation @xmath25 @xmath26 analytically in @xmath27 for all values of @xmath4 and @xmath5 , where @xmath28 is the number of lattice sites . the analytic calculation of correlation functions , in particular of the spin - spin correlation function , by gebhard and vollhardt@xcite showed that in the non - magnetic case , for @xmath29 , the results obtained with the gwf are in very good agreement with exact analytic and numerical results for the antiferromagnetic heisenberg chain . in fact , haldane@xcite and shastry@xcite discovered that the gwf at @xmath4 @xmath2 @xmath30 is the _ exact _ ground state wave function of the antiferromagnetic heisenberg chain with @xmath31 exchange . results were also obtained in dimensions @xmath32 using numerical techniques@xcite and finite orders of perturbation theory.@xcite within the diagrammatic approach of mv it also became possible to derive the well - known gutzwiller approximation in the limit of infinite spatial dimensions ( @xmath33).@xcite comprehensive investigations in this limit were made possible by the approach of gebhard,@xcite which allows for explicit evaluations of expectation values for arbitrary starting wave functions @xmath13 ( including ones with broken symmetry ) and facilitates the expansion in @xmath34 around @xmath33 . this approach was also extended to multi - band hubbard models;@xcite recently that method was combined with density functional theory , and applied to ferromagnetic transition metals.@xcite the gutzwiller approximation also describes a correlation - induced transition from metal to insulator , the brinkman - rice transition.@xcite we recently investigated the effect of correlated hopping , which for the gwf can be calculated in terms of @xmath23 and @xmath25 in any dimension , on this transition.@xcite up to now the analytic calculation of expectation values in @xmath27 was limited to the unmagnetized paramagnetic phase . in view of the renewed interest in the microscopic foundations of metallic ferromagnetism ( see ref . for a review ) , it is desirable to perform such evaluations also for the gwf with _ non - zero _ magnetization ( @xmath35 ) . in this paper we show that , in spite of formidable technical complications , it is indeed possible to evaluate such expectation values even for finite magnetization . from suitable canonical transformations we obtain diagrammatic relations and reduce the expectation values for @xmath35 to those for @xmath0 @xmath2 @xmath30 . thereby it becomes possible to calculate the double occupation @xmath25 and the momentum distribution @xmath36 for arbitrary values of the correlation parameter @xmath4 , density @xmath5 , and magnetization @xmath0 . furthermore we derive an expression for @xmath36 in closed form , which was not available up to now even for zero magnetization . these quantities determine the energy expectation value and thus the optimal variational parameter and spontaneous magnetization . the paper is structured as follows : in sec . [ sec : diagrammatics ] the diagrammatic formulation is used to derive diagrammatic relations from canonical transformations , valid in arbitrary dimensions . the evaluation of expectation values is derived in sec . [ sec : evaluation ] . the resulting magnetic phase diagram for the hubbard model in @xmath27 is presented in sec . [ sec : results ] . the conclusion in sec . [ sec : conclusion ] closes the presentation . the double occupation @xmath37 and the momentum distribution @xmath38 of the gwf are required for the calculation of the variational energy , @xmath39 @xmath2 @xmath40 . another quantity of interest is the discontinuity @xmath41 of @xmath23 at the fermi surface , @xmath42 @xmath2 @xmath43 @xmath44 @xmath45.@xcite the rules for the diagrammatic expansion of these expectation values in powers of @xmath46 were developed by mv , with the result [ entwicklungen ] @xmath47 where @xmath48 @xmath2 @xmath49 . the functions @xmath50 and @xmath51 can be represented by feynman diagrams corresponding to those of the energy and the greens function , respectively , of a @xmath52 theory . for later convenience we define @xmath53 the diagrams for @xmath50 can be obtained from those for @xmath51 by connecting the two external vertices ( see mv ) : @xmath54 this equation yields sum rules@xcite for the density of particles inside , @xmath55 , or outside of the fermi surface , @xmath56 , namely @xmath57 it is sometimes useful to remove the diagrams @xmath58 from @xmath59 and thus define @xmath60 note that our definitions in eqs . ( [ f0h0c0def ] ) and ( [ fhcdef ] ) differ slightly from mv . for the relations to be discussed next the hopping amplitude @xmath61 is assumed to be non - zero only for hopping between sites @xmath12 and @xmath62 on different sublattices @xmath63 and @xmath64 . in the next chapter we will see , however , that in dimension @xmath27 this requirement can be dropped . the simplest canonical transformation is the interchange of spin indices @xmath65 and @xmath66 which implies [ negmagn ] @xmath67 furthermore , a particle - hole transformation for both spins yields ( mv ) @xmath68 [ bigdens ] here @xmath69 is a vector in the first brillouin zone with @xmath70 for a lattice vector @xmath71 , @xmath64 , respectively ; @xmath72 for a hypercubic lattice with spacing @xmath73 . from eqs . ( [ negmagnkocc ] ) and ( [ bigdenskocc ] ) it follows that the discontinuity at the fermi surface obeys@xcite @xmath74 note that in particular @xmath75 for @xmath5 @xmath2 @xmath3 . for densities @xmath30 @xmath76 @xmath5 @xmath76 @xmath77 the magnetization is in the range @xmath78 @xmath76 @xmath79 . in view of eqs . ( [ negmagn ] ) , ( [ bigdens ] ) , and ( [ qdisctrafo ] ) we limit ourselves from now on to @xmath80 . therefore @xmath81 @xmath2 @xmath65 will be referred to as `` majority spin '' and @xmath81 @xmath2 @xmath66 as `` minority spin . '' note that the case @xmath5 @xmath2 @xmath0 ( the fully polarized state without doubly occupied sites ) can be obtained from the uncorrelated case @xmath4 @xmath2 @xmath3 . performing a particle - hole transformation for @xmath65-operators only,@xcite @xmath82 one may derive the following identities for @xmath37 and @xmath38 : @xmath83 [ ginvrelations ] for the uncorrelated case ( @xmath4 @xmath2 @xmath3 ) we have in particular [ nonintrelations ] @xmath84 the relations in eq . ( [ nonintrelations ] ) express a property of the starting wave function [ eq . ( [ startwf ] ) ] and can also be derived directly from the fact that hopping occurs only between @xmath63 and @xmath64 sublattices . we now derive diagrammatic relations for @xmath85 and @xmath59 from the identities in eqs . ( [ ginvrelations ] ) and ( [ nonintrelations ] ) . the following equations are valid for all @xmath86 and @xmath87 unless noted otherwise . from eq . ( [ doppelbes ] ) we obtain @xmath88 our goal is to equate coefficients of powers of @xmath46 . to this end we make use of the expansion @xmath89 we then obtain from eq . ( [ drelation ] ) @xmath90 while from eq . ( [ doppelbes ] ) we have @xmath91 we are thus led to the relation @xmath92 now we employ the _ _ binomial inversion formula__@xcite @xmath93 which is valid for arbitrary @xmath94 and @xmath95 . when applied to eq . ( [ erstecrelation ] ) it yields @xmath96 we stress that the relations in eqs . ( [ erstecrelation ] ) and ( [ crelation ] ) are valid in arbitrary dimensions for lattices with hopping between @xmath63 and @xmath64 sublattices only . in the next section this equation will be used to calculate @xmath85 in @xmath27 . an analogous procedure is used to derive relations for @xmath59 . we define the abbreviations [ fdefs ] @xmath97 and rewrite eq . ( [ kraumbes ] ) as @xmath98 this expression appears on the left - hand sides of eqs . ( [ kuprelation ] ) and ( [ kdownrelation ] ) , while their right - hand sides take the form [ nktransformiert ] @xmath99 expanding @xmath100 in powers of @xmath46 [ eq . ( [ ginvexpansion ] ) ] , comparing coefficients , and combining both cases , we find @xmath101 for clarity we will from now on label @xmath59 , @xmath102 , and @xmath23 with the subscripts @xmath103 and @xmath104 , depending on whether the momentum lies inside or outside of the fermi surface . we first simplify the equations for @xmath105 . using eq . ( [ frelation ] ) together with ( [ crelation ] ) we obtain @xmath106 valid for all @xmath107 , relating @xmath108 and @xmath109 . for the minority spin a similar calculation yields [ fdownrelation ] @xmath110 i. e. , there is a relation between the @xmath111 for momenta inside of the fermi surface , and another relation for momenta outside of the fermi surface , each linking the cases @xmath112 and @xmath113 . note that the relations in eqs . ( [ huprelation])-([fdownrelation ] ) are valid in arbitrary dimensions for lattices with hopping between @xmath63 and @xmath64 sublattices only . in the remainder of the paper we consider the gwf for the hubbard chain with a symmetric dispersion @xmath114 that increases monotonically with @xmath115 . this implies that @xmath116 ; therefore we only consider @xmath117 . for convenience we assume that the zero of energy is chosen such that the center of the band , @xmath118 , is zero . for such dispersions the free fermi sea described by the starting wave function @xmath13 is centered around @xmath119 and is simply connected , i. e. , @xmath120 is a step function . ( we follow the convention of mv of measuring @xmath121 in units of @xmath122 , where @xmath73 is the lattice spacing ; the first brillouin zone is the interval @xmath123 $ ] , the reciprocal lattice vectors @xmath124 are integers , and the nesting vector @xmath125 . ) for this fermi surface topology the fermi momentum , @xmath126 , only depends on the particle density @xmath127 , i. e. , the particular form of the dispersion is irrelevant . in general this simplification occurs only in dimension @xmath27 ; higher - dimensional tight - binding dispersions usually do not depend only on @xmath128 , although this symmetry can be artificially imposed to allow the construction of a dispersion from a given density of states.@xcite since the values of the diagrams @xmath85 and @xmath59 are completely determined by the function @xmath129 ( which is independent of the dispersion as long as @xmath130 is increasing with @xmath115 ) the relations derived in sec . [ subsec : trafos ] , and hence the relations in eqs . ( [ erstecrelation ] ) , ( [ crelation ] ) , ( [ huprelation ] ) , ( [ fdownrelation ] ) , are valid for @xmath131 increasing and symmetric dispersions in @xmath27 . the following analytic calculation of gwf expectation values with magnetization @xmath132 is based on the corresponding calculation for @xmath0 @xmath2 @xmath30 by mv . the calculation for @xmath0 @xmath2 @xmath30 was made possible by exploiting the relations following from canonical transformations , the polynomial form of the diagrams and their continuity as functions of @xmath121 and @xmath5 , and an analysis of the contribution of umklapp processes . we will now use very similar methods to express the double occupation @xmath37 and the momentum distribution @xmath133 in terms of the known quantities for @xmath0 @xmath2 @xmath30 . in section a we review the results of mv and present new closed formulas for @xmath36 for zero magnetization . for the magnetic case the double occupation , the momentum distribution , and its discontinuity at the fermi surface are calculated in sections b , c , and d. the variational energy is evaluated in section e. in mv the coefficients of the polynomials @xmath138 and @xmath139 had to be calculated recursively to obtain the momentum distribution , and convergence was problematic for @xmath153 close to 1 . in appendix [ app : polynomials ] we show how the recursion equations for @xmath138 can in fact be solved in closed form [ eqs . ( [ r - explicit1])-([r - explicit4 ] ) ] , while @xmath139 can be expressed in terms of them and their integrals . the generating functions are calculated as @xmath154 \!,\,\label{q - function } \end{aligned}\ ] ] for @xmath155 . here @xmath2 @xmath157 is the complete elliptic integral of the first kind , @xmath158 is an auxiliary function , @xmath159 where the dot indicates a partial derivative with respect to second argument , and @xmath160 is the repeated integral of @xmath161 , defined by ( @xmath162 ) @xmath163 below we will also need the following integral , @xmath164 in appendix [ app : polynomials ] we provide an explicit expression for @xmath160 [ eqs . ( [ r - integral - alt])-([r - integral - alt - polynomial ] ) ] , as well as other relations . here we note in particular the following functional relations , which are obtained from eqs . ( [ r - function ] ) and ( [ q - function ] ) , @xmath165 these expressions analytically continue @xmath160 and @xmath166 to @xmath167 . to calculate the double occupation one needs the diagrams @xmath85 . using the methods of mv one can show that for @xmath168 umklapp processes occur in the graphs for @xmath50 only if @xmath169 . this implies that for @xmath170 @xmath50 is a homogeneous function of @xmath171 and @xmath172 of order @xmath173 , since every contributing graph contains @xmath173 free momenta . since @xmath174 it follows that @xmath50 can be written as a linear combination of terms @xmath175 with @xmath176 , i. e. , @xmath177 is a polynomial in @xmath178 of degree @xmath173 this is also the case for @xmath5 @xmath2 @xmath3 since @xmath37 is continuous at @xmath5 @xmath2 @xmath3 [ see eq . ( [ bigdensdocc ] ) ] . therefore we can write the polynomial simply as @xmath179 for @xmath134 . for @xmath5 @xmath2 @xmath3 , however , we can calculate @xmath180 from eq . ( [ crelation ] ) and eq . ( [ cnonmagn ] ) : @xmath181 the polynomial structure of @xmath50 then implies @xmath182 summation of the series in eq . ( [ doppelbes ] ) yields the simple result @xmath183 it is remarkable that the double occupation at density @xmath5 and magnetization @xmath0 is obtained as the difference between the double occupation without magnetization at density @xmath5 and _ density _ @xmath0 . with the closed form of @xmath184 taken from eq . ( [ doccnonmagn ] ) we finally obtain for the double occupation , valid for @xmath87 , @xmath185 the double occupation is shown in fig . [ fig : docc ] for various parameter values . in the limit of strong correlation ( @xmath186 ) it behaves as @xmath187 + o(g^4 ) , \\ d(g , n=1,m<1 ) & = g^2 \left [ \ln\frac{1}{g } + \frac{\ln(1-m)-(1-m)}{2 } \right ] + o(g^4\ln g ) , \end{aligned}\ ] ] i. e. , the double occupation is _ non - analytic _ in the limit @xmath188 , @xmath186 , as in the non - magnetic case ( see mv ) . to calculate the momentum distribution @xmath36 one needs the diagrams @xmath59 or @xmath102 . in the following subsections we distinguish the cases of majority and minority spin , as well as whether @xmath121 lies inside or outside of the fermi surface . for finite magnetization , @xmath168 , we make the following observations , similar in spirit to those of mv for the non - magnetic case . as a function of @xmath121 , @xmath189 and @xmath190 are discontinuous at @xmath191 , since the one - particle irreducible graphs contain a factor @xmath129 . for @xmath192 momentum conservation at the outer vertices of @xmath102 can not be fulfilled , so that in this case @xmath193 . for @xmath194 umklapp processes occur and yield an additional contribution to @xmath102 of normal processes with external momentum @xmath144 . in the absence of umklapp processes @xmath59 and @xmath102 are homogeneous functions of @xmath171 and @xmath172 of order @xmath140 , since every contributing graph contains @xmath140 free momenta . then @xmath195 and @xmath196 are polynomials in @xmath197 and @xmath178 of degree @xmath140 , and due to momentum conservation at outer vertices different polynomials occur depending on whether @xmath121 is larger or smaller than @xmath198 . furthermore , different polynomials for @xmath199 occur also depending on how @xmath121 compares to @xmath200 ( see subsection 4 below and appendix [ app : down - out - sector - structure ] ) . for @xmath202 and @xmath5 @xmath2 @xmath3 the diagrams @xmath108 can be obtained from eq . ( [ huprelation ] ) in terms of the known functions @xmath109 of the non - magnetic case [ eq . ( [ fnonmagn ] ) ] : @xmath203 } & \text{for } 0\leq k\leq \frac{1 - 3m}{4 } \end{cases } \!.\,\label{hupingleichung } \end{aligned}\ ] ] since umklapp processes do not occur for @xmath204 , @xmath205 is a polynomial in @xmath197 and @xmath178 . hence we can obtain @xmath206 by replacing @xmath121 by @xmath197 and @xmath0 by @xmath178 in eq . ( [ hupingleichung ] ) . this yields @xmath207 } & \text{for } 0\leq k\leq \frac{n-3m}{4 } \end{cases } \!.\,\label{hupinergebnis } \end{aligned}\ ] ] note that @xmath208 is continuous for all @xmath204 due to @xmath209 ( see mv ) . the normal processes contribute differently depending on how @xmath121 compares to @xmath210 , as expected . for @xmath215 we can deduce @xmath109 at once by inserting @xmath216 into eq . ( [ huprelation ] ) : @xmath217 } & \text{for } 1\!-\!\frac{3n - m}{4}\leq k\leq\frac{1}{2 } \end{cases } \ ! , \end{aligned}\ ] ] which is continuous for all @xmath218 . in this sector umklapp processes occur for @xmath219 and contribute to @xmath108 like normal processes with external momentum @xmath144 . summing the series in eq . ( [ kraumbes ] ) , or applying eq . ( [ kuprelation ] ) to eq . ( [ upinresult ] ) , yields the result , valid for @xmath211 , @xmath215 , @xmath220 \!,\,\label{upoutresult } \end{aligned}\ ] ] which alternatively can be derived directly from eqs . ( [ kuprelation ] ) , ( [ upinresult ] ) , and ( [ q - fortsetzung ] ) . note also that for @xmath0 @xmath2 @xmath30 this result reduces to eq . ( [ nonmagnoutresult ] ) . the momentum distribution @xmath221 has thus been determined in the entire parameter range . it is shown for @xmath4 @xmath2 @xmath222 and densities @xmath5 @xmath2 1 and 0.8 in fig . [ fig : kocc10]a and [ fig : kocc08]a . for @xmath224 and @xmath5 @xmath2 @xmath3 the diagrams @xmath225 are given in terms of the known functions @xmath226 of the non - magnetic case [ eq . ( [ fnonmagn ] ) ] according to eq . ( [ fdownrelin ] ) : @xmath227 due to the absence of umklapp processes for @xmath228 , @xmath229 is a polynomial in @xmath197 and @xmath178 . therefore @xmath230 is given by eq . ( [ fdowningleichung ] ) with @xmath121 replaced by @xmath197 and @xmath0 replaced by @xmath178 . we then have @xmath231 where the second equation was obtained by inserting eq . ( [ fdowninergebnis ] ) in eq . ( [ fdownrelin ] ) ; it is equivalent to the functional relation in eq . ( [ r - fortsetzung ] ) . note that for @xmath0 @xmath2 @xmath30 eq . ( [ fdowninergebnis2 ] ) reduces to ( [ fnonmagn ] ) . we use eq . ( [ fdowninergebnis2 ] ) in eq . ( [ kraumbes ] ) to find the following expression for @xmath232 , valid for @xmath211 , @xmath233 , @xmath234 \!.\label{downinresult } \end{aligned}\ ] ] this expression reduces to eq . ( [ nonmagninresult ] ) for @xmath0 @xmath2 @xmath30 . finally we consider the case @xmath236 , for which the calculation of @xmath237 is somewhat more complicated . we begin with the special case @xmath5 @xmath2 @xmath3 , for which the momentum distribution can be determined immediately from eqs . ( [ kdownrelation ] ) , ( [ negmagnkocc ] ) , and ( [ kuprelation ] ) , which combine to give @xmath238 inserting eq . ( [ upinresult ] ) we arrive at @xmath239 \!.\,\label{downoutresult - n=1 } \end{aligned}\ ] ] next we consider arbitrary density and magnetization , @xmath240 . we make use of the following relation , which follows from eqs . ( [ fdownrelout ] ) and ( [ erstecrelation ] ) , @xmath241 the diagrams @xmath242 that appear in this equation can be written in terms of the contribution of normal processes , @xmath243 , as @xmath244 } & \text{for } 1\!-\!\frac{3n+m}{4}\leq k\leq\frac{1}{2 } \end{cases } \!.\,\label{hdownout } \end{aligned}\ ] ] in appendix [ app : down - out - sector - structure ] we show that depending on @xmath121 and @xmath0 , the function @xmath245 is given piecewise by four polynomials in @xmath121 and @xmath0 of order @xmath140 . the explicit determination of these polynomials is quite involved ; it is presented in appendix [ app : down - out - sector - polynomials ] . the final result for @xmath237 , valid for @xmath240 , @xmath236 , can be written as @xmath246 \!,\,\label{downoutresult } \end{aligned}\ ] ] where @xmath247 is given by @xmath248 with [ n - functions ] @xmath249 and the dot again denotes derivative with respect to second argument . it can be checked that eq . ( [ downoutresult ] ) indeed reduces to ( [ downoutresult - n=1 ] ) for @xmath0 @xmath2 @xmath30 . thus the momentum distribution @xmath250 has been determined for all parameters . it is shown for @xmath4 @xmath2 @xmath222 and densities @xmath5 @xmath2 1 and 0.8 in fig . [ fig : kocc10]b and [ fig : kocc08]b . the calculation of the correlated momentum distributions @xmath36 is now complete . we remark that they are continuous functions of @xmath121 , except at @xmath251 , and are also continuous in @xmath5 and @xmath0 for fixed @xmath121 . it can be checked that they obey the sum rule in eq . ( [ sumrule ] ) . it suffices to calculate the discontinuity @xmath252 of @xmath36 at the fermi surface for @xmath211 [ see eq . ( [ qdisctrafo ] ) ] . from our previous results we obtain the left and right limit of the momentum distribution at the fermi vector as @xmath253 where the abbreviation @xmath254 is defined as @xmath255 [ 1-(1-g^2)m]^{-\text{sgn}(\sigma ) } } .\,\label{gdef } \end{aligned}\ ] ] for the discontinuity at the @xmath81-spin fermi surface we thus obtain @xmath256 it follows that @xmath252 vanishes only for a half - filled band without double occupation ( @xmath5 @xmath2 @xmath3 and @xmath4 @xmath2 @xmath30 ) ; in this case there is exactly one particle at each site so that @xmath257 for all @xmath121 . note also that @xmath258 if @xmath5 @xmath2 @xmath3 ( or , trivially , if @xmath0 @xmath2 @xmath30 or @xmath4 @xmath2 @xmath3 ) . we plot @xmath41 for @xmath5 @xmath2 @xmath3 and @xmath5 @xmath2 @xmath259 in fig . [ fig : qdisc ] . for any symmetric dispersion @xmath130 , monotonically increasing with @xmath115 , we can now calculate the energy expectation value per site , @xmath39 @xmath2 @xmath260 , of the one - dimensional hubbard hamiltionian ( [ hamiltonian ] ) , which is then minimized w.r.t . @xmath4 to find the optimal variational energy , @xmath261 , @xmath262 note that it follows from eq . ( [ bigdenskocc ] ) that the total kinetic energy for dispersion @xmath130 at density @xmath263 can be calculated from the dispersion @xmath264 at density @xmath265 using the formulas for @xmath36 and @xmath25 that are valid below half - filling . for the hubbard chain with nearest - neighbor hopping @xmath266 the dispersion relation in our notation is @xmath267 . we assume @xmath266 @xmath268 @xmath30 without loss of generality , so that the dispersion is increasing with @xmath115 and our results for the gutzwiller expectation values apply . the optimal variational energy for this system is shown in fig . [ fig : nn - energy - m ] for densities @xmath5 @xmath2 1 and 0.8 for various magnetizations . note that at half - filling no brinkman - rice metal - insulator transition occurs at any finite @xmath7 ; i. e. , @xmath4 @xmath2 0 is the optimal variational parameter only for @xmath7 @xmath2 @xmath269 . the variational result for the ground - state magnetization is determined in the next section . in this section we determine the instability towards ferromagnetism for the hubbard chain with nearest - neighbor hopping . currently only homogeneous paramagnetic and ferromagnetic phases can be investigated analytically with the gwf in @xmath1 @xmath2 @xmath3 ; hence we do not consider antiferromagnetism or other broken symmetries . we begin by examining the energy for the special cases of the paramagnetic state ( i. e. , zero magnetization ) and the fully polarized state . the latter contains the minimum number of doubly occupied sites and is an eigenstate of @xmath19 , with eigenvalue @xmath270 where @xmath271 is the kinetic energy of one spin species for the uncorrelated state . for the case of nearest - neighbor hopping we have @xmath272 @xmath2 @xmath273 , @xmath266 @xmath268 @xmath30 . in fig . [ fig : nn - energy-0 ] the exact ground - state energy , @xmath274 , obtained from the bethe ansatz solution,@xcite is compared to the gutzwiller energy for zero and maximal polarization , at various densities . we also show the energy of the variational hartree - fock theory , @xmath275 ; it is contained as a special case in the results for the gwf , @xmath276 as expected , the @xmath4-optimized gwf significantly improves upon hartree - fock theory but overestimates @xmath277 at large @xmath7.@xcite since all spin configurations are degenerate for @xmath278 , the exact ground - state energy coincides with @xmath279 in this case ; therefore the gutzwiller energy @xmath280 necessarily crosses the value @xmath279 at some finite value of the interaction @xmath281 ( except for @xmath282 ) . the existence of a finite critical interaction @xmath281 above which the gwf predicts a ferromagnetic ground state is in contrast to the lieb - mattis theorem,@xcite which states that @xmath283 for the exact ground state ( i. e. , @xmath284 ) . the reason for this overestimation of the instability of the paramagnetic state lies in the simple structure of the gwf , which controls only _ local _ correlations and can not describe the special correlated behavior in @xmath27 microscopically . the preceeding discussion only compared the variational energies for zero and full polarization . from our new results for the ferromagnetic gwf we can also study the stability of partially polarized ferromagnetic states . we first consider hartree - fock theory . a simple calculation shows that it predicts a fully polarized ground state for @xmath285 , where @xmath286}/(\pi n^2 ) & \text{for } 0\leq n\leq1 \\ u_c^{\text{hf}}(2-n ) & \text{for } 1\leq n\leq2 \end{cases } .\,\label{ucrit - hf } \end{aligned}\ ] ] this critical interaction @xmath287 is smaller than that derived from the stoner criterion , @xmath288 , where @xmath289 is the density of states . note that @xmath290 as a function of @xmath0 never develops a local minimum at @xmath35 . on the other hand , a maximum at finite @xmath0 occurs for @xmath7 @xmath268 @xmath291 , which leads to a global minimum at full polarization already for @xmath7 @xmath292 @xmath287 . the stoner criterion , which merely signals a negative curvature of @xmath290 at @xmath283 and does not take into account a finite magnetization , is thus irrelevant for the hubbard chain with nearest - neighbor hopping . for the gutzwiller wave function we find that @xmath280 as a function of @xmath0 at fixed @xmath7 developments local extrema and global minima in a qualitatively similar fashion to @xmath290 . as a consequence the gwf also describes a discontinuous transition from the paramagnetic state to a state with full polarization at @xmath7 @xmath2 @xmath293 . this critical interaction @xmath294 is shown in fig . [ fig : ucvsn ] . compared to hartree - fock theory we find agreement in the limit of small @xmath5 . however , at intermediate densities the gwf predicts a significantly reduced ferromagnetic region . in particular for @xmath5 @xmath295 @xmath3 we have @xmath294 @xmath295 @xmath269 , as expected from the previous discussion ( see also fig . [ fig : nn - energy-0]a ) . thus , in contrast to hartree - fock theory , the gwf does not exhibit a spurious ferromagnetic transition at half - filling , since it is able to avoid double occupation not only through a ferromagnetic polarization , but also by decreasing the variational parameter @xmath4 . away from half - filling , however , the gwf predicts ferromagnetism for sufficiently large @xmath7 , in contrast to the exact solution for the hubbard chain . in this paper we presented general diagrammatic relations for the expectation values of the @xmath1-dimensional hubbard model in terms of the gutzwiller wave function ( gwf ) at _ non - zero _ magnetization @xmath0 . in @xmath27 explicit , approximation - free evaluations of the double occupation @xmath37 and the momentum distribution @xmath133 were made possible by exploiting ( i ) relations for the feynman diagrams for @xmath25 and @xmath36 , derived from canonical spin and particle - hole transformations , ( ii ) the polynomial form of the diagrams in powers of @xmath121 , @xmath5 , and @xmath0 , and ( iii ) an analysis of the contributions of normal and umklapp processes . in this way the calculation of @xmath37 and @xmath133 was reduced to that for @xmath0 @xmath2 @xmath30 . furthermore , new closed expressions for the momentum distribution @xmath36 were derived , facilitating numerical evaluation . the functions @xmath37 and @xmath133 in @xmath27 are qualitatively similar to those for @xmath0 @xmath2 @xmath30 . the discontinuity @xmath42 of the momentum distribution at the fermi energy was also calculated explicitly . it is always finite , except for the half - filled band without double occupation ( @xmath4 @xmath2 @xmath30 , @xmath5 @xmath2 @xmath3 , @xmath0 @xmath2 @xmath30 ) in which case the electrons are trivially localized . in all other cases the gwf describes a ferromagnetic fermi liquid . analysis of the gutzwiller variational energy for the hubbard chain with nearest - neighbor hopping shows that the gwf predicts a fully polarized ferromagnetic state at large enough @xmath7 and away from half - filling , in contrast to the lieb - mattis theorem.@xcite this exemplifies once more the peculiarities of the gwf which controls correlations between the electrons only globally through the local hubbard interaction . while the gwf is an excellent wave function for the one - dimensional _ heisenberg _ model ( at least for @xmath0 @xmath2 @xmath30),@xcite since this only involves spin correlations between localized spins , it is not a very good wave function for the one - dimensional _ hubbard _ model at large @xmath7 and @xmath296 since it does not describe density correlations well in this case.@xcite as a consequence the gwf can not reproduce all characteristics of the one - dimensional system . this is also apparent from the finite discontinuity of the momentum distribution at the fermi surface , which is , in fact , continuous for luttinger liquids such as the one - dimensional hubbard model . on the other hand the ferromagnetic gwf represents a trial state for partially polarized , itinerant electrons and may thus be regarded as an effective , non - perturbative description of a ferromagnetic fermi liquid . in view of the considerable technical complications involved in the present calculations it is not clear whether it will be possible to compute correlation functions with the gwf for @xmath35 . since the calculation of the spin - spin correlation function@xcite for @xmath0 @xmath2 @xmath30 helped to gain considerable insight into the properties of heisenberg - type models,@xcite a corresponding result for @xmath35 would be helpful for a better understanding of one - dimensional heisenberg models in a magnetic field . m. k. would like to acknowledge very helpful discussions with g. s. uhrig in the early stages of this research . this work was supported in part by the sonderforschungsbereich 484 of the deutsche forschungsgemeinschaft . part of this work was carried out while m. k. was at yale university , supported by dfg grant ko 2056 and us nsf grant dmr 00 - 98226 . in this appendix we describe the derivation of closed expressions for the polynomials @xmath297 and @xmath298 that appear in sec . [ sec : evaluation ] . by eliminating @xmath298 from mv s recursion formulas we obtain @xmath299 , \;\;p\geq0.\,\label{rgl1 } \end{aligned}\ ] ] furthermore , the polynomials @xmath298 can be expressed in terms of @xmath297 as @xmath300 together with @xmath301 @xmath2 0 . we define @xmath302 , @xmath303 . a closed form for @xmath138 is obtained as follows . using eqs . ( [ fdowninergebnis])-([fdowninergebnis2 ] ) we can reduce eq . ( [ rgl1 ] ) to @xmath304 this is essentially the differential equation of the legendre polynomials except for the inhomogeneity on the right - hand side . from an expansion in legendre polynomials @xmath305 we thus obtain @xmath306 which after some calculation yields eq . ( [ r - function ] ) . from @xmath150 we then obtain the explicit expressions in terms of a ( terminating ) hypergeometric function , @xmath307 furthermore an integration by parts of eq . ( [ qgl1 ] ) leads to the expression for @xmath151 in terms of @xmath160 shown in eq . ( [ q - function ] ) . by using a hypergeometric identity to rewrite eq . ( [ r - function ] ) as @xmath308 integrating term - wise w.r.t . @xmath309 , and again using several hypergeometric identities we obtain the following explicit expression for @xmath160 , @xmath310^{p+\frac{1}{2 } } } , \,\label{r - integral - alt } \end{aligned}\ ] ] where the @xmath311 are polynomials of degree @xmath140 , @xmath312 the functions @xmath160 may be evaluated via the series ( [ r - integral - alt ] ) for not too large values of @xmath313 . alternatively , the integration in eq . ( [ r - integral ] ) can be performed numerically . finally , we note a few special values . from mv s polynomial relations we obtain @xmath314 together with eq . ( [ r - function])-([w - function ] ) we find in particular @xmath315 furthermore the sum rule for particles outside of the fermi surface [ eq . ( [ sumrule ] ) ] implies @xmath316 in this appendix we derive the polynomial structure of the diagrams @xmath242 , for which a lot of cases must be distinguished . as discussed at the beginning of sec . [ subsec : magnkocc ] , different polynomials occur in the contribution of normal processes depending on the relation of @xmath121 to @xmath317 and @xmath318 . furthermore , due to a certain class of diagrams , shown in fig . [ fig : graphs ] , different polynomials for @xmath199 may in principle occur also at @xmath200 , @xmath319 , and all higher odd multiples of @xmath172 ; however some simplification takes place , as we will show below . _ case a _ : @xmath320 . here the momenta @xmath317 , @xmath318 , @xmath321 , etc . , are ordered as follows , @xmath322 so that the contribution of normal processes to @xmath242 , as it appears in eq . ( [ hdownout ] ) , can be written in terms of three polynomials @xmath323 of order @xmath140 : @xmath324 if @xmath325 umklapp processes contribute to @xmath242 with momentum @xmath144 , and several cases must be distinguished to determine the appropriate polynomials . we find for @xmath112 and @xmath326 @xmath327 while for @xmath112 and @xmath328 @xmath329 whereas for @xmath113 @xmath330 _ case b _ : @xmath331 . here , on the other hand , we have the ordering @xmath332 and the contributions of normal processes to @xmath242 are now given by four polynomials @xmath333 of order @xmath140 : @xmath334 let us determine the umklapp process contributions for the region @xmath335 , @xmath336 now we connect the polynomials with one another via eq . ( [ hdownrelout ] ) , which performs the transformation @xmath5 @xmath295 @xmath3 @xmath44 @xmath0 and @xmath0 @xmath295 @xmath3 @xmath44 @xmath5 . this provides a link between eqs . ( [ hdown - a-1b ] ) and ( [ hdown - a-2 ] ) , the right - hand sides of which are thus related , line by line , via ( [ hdownrelout ] ) . similarly , eqs . ( [ hdown - b ] ) and ( [ hdown - a-1a ] ) are connected , which reveals that the distinction at momentum @xmath337 in ( [ hdown - b ] ) is in fact absent in eqs . ( [ hdown - a-1a ] ) . hence we find @xmath338 @xmath2 @xmath339 , and also , by comparison with the first transformation , @xmath340 @xmath2 @xmath341 , @xmath338 @xmath2 @xmath342 . by inspecting the region @xmath343 @xmath76 @xmath0 @xmath76 @xmath344 of case @xmath64 , we find a similar connection to larger @xmath0 ( i. e. , _ case c _ : @xmath345 @xmath76 m @xmath76 @xmath346 ) , and use of eq . ( [ hdownrelout ] ) shows that the new distinction at momentum @xmath347 disappears in a similar fashion , and that all its polynomials likewise reduce to the ones above . it is not difficult to see that this simplification repeats for larger magnetization , i. e. , also for all @xmath348 . we conclude that new polynomials at higher odd multiples of @xmath349 are ruled out by the symmetry of the diagrams ; only the distinction at momentum @xmath350 survives . therefore the contribution ( [ hdownout ] ) of normal processes to @xmath242 can finally be written as @xmath351 i. e. , a total of four polynomials are needed to describe @xmath237 ; they are determined in appendix [ app : down - out - sector - polynomials ] . in this appendix we determine the polynomials that appear on the right - hand side of eq . ( [ downout - polynomials ] ) . first let us examine @xmath352 . for @xmath0 @xmath2 @xmath30 we immediately obtain @xmath353 @xmath2 @xmath354 by comparison with eq . ( [ fnonmagn ] ) . furthermore we can derive its behavior for small @xmath0 from the equation @xmath355 which is a simple consequence of eq . ( [ negmagnkocc ] ) . similar to mv it can be shown that @xmath356 , when regarded as a function of @xmath5 , has two continuous derivatives at @xmath5 @xmath2 @xmath3 ( for all @xmath121 @xmath357 @xmath171 ) . then eq . ( [ hupinrelation ] ) implies that @xmath356 has the same property as a function of @xmath0 at @xmath0 @xmath2 @xmath30 . hence the expression ( [ hupingleichung ] ) may be used on the right - hand side of eq . ( [ h - down - h - up ] ) for small positive @xmath0 , up to an error of @xmath358 . inserting the appropriate polynomials for momenta in the interval @xmath359 we obtain @xmath360 which holds for all @xmath121 and small @xmath0 . the following definitions will help keep the notation compact . let @xmath361 be any of @xmath362 , @xmath363 , @xmath364 , @xmath365 . in addition to the polynomials @xmath366 , we define @xmath367 @xmath368 @xmath369 , and introduce their generating functions @xmath370 now we are ready to collect the relations between eqs . ( [ hdown - a-1b ] ) and ( [ hdown - a-2 ] ) and between ( [ hdown - b ] ) and ( [ hdown - a-1a ] ) that eq . ( [ hdownrelout ] ) provides [ see appendix [ app : down - out - sector - structure ] ] . setting @xmath371 , @xmath372 , and @xmath373 in these relation , we obtain ( for arbitrary @xmath374 ) and other relations that are in fact implied by these . by repeated combination of eqs . ( [ a - poly - relations ] ) we find the important equality @xmath376 furthermore , by setting @xmath377 in eq . ( [ a3-poly - relation ] ) we immediately obtain the explicit expression @xmath378 in terms of generating functions our results so far can be expressed as @xmath379 we proceed to determine @xmath365 . first we obtain @xmath380 from eq . ( [ a2-function - relation ] ) at @xmath381 by inserting the expansion ( [ a2-function - expansion ] ) and differentiating with respect to @xmath374 ; we then set @xmath382 , so that the error term vanishes . this yields an expression for @xmath383 , which we integrate with respect to @xmath309 , using @xmath384 @xmath2 @xmath30 . this result for @xmath380 is used in eq . ( [ a3-function - relation ] ) with @xmath382 , which yields @xmath385 for arbitrary @xmath386 , @xmath387 the dot indicates derivative with respect to second argument . next we derive @xmath362 . first we obtain @xmath388 from eq . ( [ a2-function - relation ] ) at @xmath382 and @xmath381 , using ( [ a2-function - expansion ] ) at @xmath283 . this result for @xmath388 is used in eq . ( [ a1-function - relation ] ) with @xmath382 , which thus yields @xmath389 for arbitrary @xmath386 , @xmath390 finally , @xmath391 is obtained from eq . ( [ a - poly - sum ] ) . these results can be rearranged into eqs . ( [ downoutresult])-([n - functions ] ) . for the transition from a paramagnetic to a fully polarized ferromagnetic state for the hubbard chain with nearest - neighbor hopping as obtained from the gutzwiller wave function . the hartree - fock result , @xmath287 [ eq . ( [ ucrit - hf ] ) ] , is also shown ; it has a shallow maximum at @xmath398 . note that @xmath293 @xmath2 @xmath399.[fig : ucvsn],height=377 ] . solid ( broken ) lines represent majority ( minority ) spins and vanish unless they carry momentum in the interval @xmath400}$ ] ( @xmath401}$ ] ) ; see mv for diagrammatic rules . for @xmath402 the diagram in ( a ) vanishes since @xmath403 @xmath404 @xmath401}$ ] can not be fulfilled . for @xmath405 the diagram in ( b ) vanishes for similar reasons , and so on for higher odd multiples of @xmath172.[fig : graphs],height=377 ]	we present analytic results for ground - state properties of hubbard - type models in terms of the gutzwiller variational wave function with non - zero values of the magnetization @xmath0 . in dimension @xmath1 @xmath2 @xmath3 approximation - free evaluations are made possible by appropriate canonical transformations and an analysis of umklapp processes . we calculate the double occupation and the momentum distribution , as well as its discontinuity at the fermi surface , for arbitrary values of the interaction parameter @xmath4 , density @xmath5 , and magnetization @xmath0 . these quantities determine the expectation value of the one - dimensional hubbard hamiltonian for any symmetric , monotonically increasing dispersion @xmath6 . in particular for nearest - neighbor hopping and densities away from half filling the gutzwiller wave function is found to predict ferromagnetic behavior for sufficiently large interaction @xmath7 .
the description of diffractive hadronic physics in terms of the exchange of a `` pomeron '' has proved remarkably durable , and has successfully withstood the advent of qcd . interest in the pomeron has been renewed by recent results from hera . total cross sections and diffractive processes are essentially soft in nature , and thus do not fall within the realm of perturbative qcd . thus any fixed - order perturbative calculation must be regarded purely as a model of the interaction , but one which may be representative of an all - orders qcd result . nevertheless , a successful phenomenological model of pomeron exchange is provided by two - gluon - exchange ( 2ge ) @xcite . crucial to this model is that the gluons be non - perturbative , i.e. infra - red finite , both to avoid the coulomb singularity at @xmath1 , and to provide the correct @xmath2-dependence of the differential cross section @xcite . recently there has been a series of papers @xcite investigating the phenomenology of the 2ge model in the form proposed by landshoff and nachtmann ( ln ) @xcite . in these papers a non - perturbative gluon propagator extracted from the solution of the schwinger - dyson equation @xcite is inserted in the ln model . here we will explore the phenomenology of the same model , using gluon propagators extracted from lattice calculations in the landau gauge . our method has the advantage that one is inserting in the model a genuine , nonperturbative qcd quantity , which has been computed from first principles rather than derived from an approximate equation . also , as we will explain in the following , we have only one free parameter in our approach , whereas the schwinger - dyson solution depends on two parameters . for these reasons , we think that our approach provides in principle the best chance to test the success of the ln model from the point of view of qcd . the layout of the remainder of this paper is as follows . in the next section we will summarise the lattice calculations , and present the results for the lattice gluon propagator . in section 3 we will describe how we apply the measured gluon propagator to the calculation of cross - sections , and compare our results to the experimental data . in section 4 we comment on the issue of gauge dependence of our results . finally we present our conclusions . we use the standard lattice definition of the gluon fields in terms of the link variables @xcite , @xmath3 where @xmath4 is the lattice spacing and @xmath5 the bare coupling constant . the gauge configurations are fixed to the landau gauge by imposing the gauge - fixing condition @xmath6 the accurate implementation of this step is crucial , and we use the fourier - accelerated algorithm of reference @xcite . the unrenormalised gluon propagator in momentum space is obtained from the gauge - fixed gluon fields by taking the fourier transform @xmath7 in the continuum limit , we can write the landau - gauge propagator in the form @xmath8 we performed calculations on two sets of hypercubic lattices , corresponding to different physical volumes and lattice spacings . in addition , we used data for the gluon propagator as evaluated in reference @xcite ; such data are particularly useful for an analysis of finite volume effects in our calculations . the parameters used are listed in table [ tab : params ] . full details of the method and of the simulation for the hypercubic lattices are contained in ref . @xcite . .parameters of the lattices used in the calculation . the quoted value of the lattice spacing is obtained from an analysis of the string tension . [ cols="^,^,^,^,^",options="header " , ] the general agreement between the values of @xmath9 and @xmath10 obtained on the various lattices is striking , suggesting either that both quantities are subject to only small discretisation or finite volume effects , or that such effects are absorbed in @xmath11 . they are also encouragingly close to the phenomenological values of @xmath12 and @xmath13 . the quality of the fit to the data can be seen in figure ( [ fig : isr_data ] ) , where we show isr data for the differential elastic cross section at @xmath14 together with the lattice predictions , with the energy correction of equation ( [ eq : dsig_energy ] ) . though the agreement of the lattice computation with the experimental data is good , differences between the lattice results are now clearer , and suggest the need to repeat the calculation at different volumes and at different lattice spacings . in particular , our lattice at @xmath15 corresponds to a somewhat tiny physical volume for hadronic physics , so that finite volume effects may be non - negligible in that case . this process provides a further important test of the 2ge model , and in the following we adopt the analysis procedure of ducati _ et al . _ @xcite . the amplitude for meson - nucleon scattering is @xmath17 3\left[f_n(q^2 ) - f_n(q^2 - \frac{3}{2 } \vec{q}\cdot\vec{k } + \frac{3}{4}k^2)\right].\label{eq : jpsinucleon}\end{aligned}\ ] ] we use the pole approximation for the form factors @xmath18 taking @xcite @xmath19 , @xmath20 , @xmath21 and @xmath22 . two phenomenological features are evident in the calculation . firstly , the quark - counting rule is closely satisfied in the case of the hadrons composed of light quarks : @xmath23 secondly , the pomeron couples more weakly to mesons composed of heavier quarks . this can be seen in figure ( [ fig : jpsi ] ) where we show the meson - nucleon cross section as a function of the pole radius used in equation ( [ eq : meson_ff ] ) , together with a regge fit to the energy - independent part of the @xmath24 and @xmath25 cross sections performed in reference @xcite . the differences between the results on the different lattices reinforces the need to repeat the calculation on a wider range of lattice parameters . the issue of gauge dependence in the ln model is a difficult one . indeed , the starting point of ln is the leading - order perturbative calculation of a quark - quark scattering amplitude , with the constraint that no colour is exchanged between quarks . this is of course a gauge invariant quantity , expressed by a two - gluon exchange diagram . on the other hand , ln argue that the salient features of pomeron physics can be captured by substituting in such a diagram the tree - level gluon propagator with an effective one , including non - perturbative self - energy effects , which should in particular account for the removal of the infrared pole . clearly any _ ansatz _ for the self - energy introduces in principle an uncontrollable gauge dependence in the calculation . given the phenomenological success of the model , one is led to think that maybe the gauge dependence intrinsic in the propagator is either mild by itself or gets suppressed ( cancels ) in the evaluation of physical quantities . we emphasise that this problem is inherent in the model and has nothing to do with using a lattice propagator . from the point of view of our method , one may speculate that the simplest conceivable mechanism for the suppression of gauge dependence would be if it gets factored out " in the definition of @xmath26 . in other words , a gauge dependent definition of @xmath11 may cancel most of the gauge dependence in the momentum space integrals of the propagator . we plan to investigate directly the issue of gauge dependence in a future work by using lattice propagators obtained in different gauges . we have demonstrated how the landau - gauge lattice gluon propagator employed in the two - gluon - exchange model of proton - proton scattering provides a highly successful description of the data , where the only parameter that needs to be fixed is an effective quark - gluon coupling . for the first time a genuine qcd quantity , evaluated entirely from first principles , has been inserted in the ln model , providing an important consistency check of the model itself . also , given the fact that in our approximation the effect of quark loops diagrams is absent , our analysis shows the robustness of the ln model , despite its simplicity . although the description of the data appears largely independent both of the volume of the lattice and of the lattice spacing , a more detailed analysis of lattice systematic errors is necessary . we plan to perform such an analysis , together with an investigation of the gauge dependence of our results , in a future publication . we wish to thank sandy donnachie and peter landshoff for useful conversations , and chris michael both for helpful conversations and for reading the manuscript . this work was supported by the united kingdom particle physics and astronomy research council ( pparc ) under grant gr / j21347 . cp and dgr acknowledge the support of pparc through advanced fellowships held at the university of liverpool and at the university of edinburgh respectively . we are grateful to edinburgh s computing service for maintenance of service on the thinking machines cm-200 where this work was performed . 99 f.e . low , phys . * d12 * ( 1975 ) 163 . s. nussinov , phys . * 34 * ( 1975 ) 1268 . richards , nucl . b258 ( 1985 ) 267 . landshoff and o. nachtmann , z. phys . c35 ( 1987 ) 405 . f. halzen , g. krein and a.a . natale , phys . ( 1993 ) 295 . gay ducati , f. halzen and a.a . natale , phys.rev . * d48 * ( 1993 ) 2324 . cornwall , phys . rev . * d26 * ( 1982 ) 1453 . j. mandula and m. ogilvie , phys . * b185 * ( 1987 ) c.t.h . davies _ et al . _ , phys . * d37 * ( 1987 ) 1581 . p. marenzoni , g. martinelli , and n. stella , rome preprint 94/1042 , southampton preprint 93/94 - 31 , hep - lat/9410011 . henty _ et al . _ , liverpool preprint lth-361 , orsay preprint lpthe 95 - 67 , hep - lat/9510045 , to appear in the proceedings of the international europhysics conference on high energy physics ( hep 95 ) , brussels , belgium , 27th . . - 2nd . , 1995 a. donnachie and p.v . landshoff , nucl . ( 1984 ) 322 . a. donnachie and p.v . landshoff , nucl . ( 1985 ) 690 . a. donnachie and p.v . landshoff , phys . ( 1987 ) 403 cudell and d.a . ross , nucl . * b359 * ( 1991 ) 247 . gunion and d. soper , phys . * d15 * ( 1977 ) 2617 . e.m . levin and m.g . ryskin , sov . j. nucl . ( 1981 ) 619 . a. breakstone _ et al . _ , nucl . * b248 * ( 1984 ) 253 . b. povh and j. hufner , phys . * 58 * ( 1987 ) 1612 ; phys . * b245 * ( 1990 ) 653 . a. donnachie and p.v . landshoff , phys . ( 1992 ) 227 .	we investigate the landshoff - nachtmann two - gluon - exchange model of the pomeron using gluon propagators computed in the landau gauge within quenched lattice qcd calculations . we first determine an effective gluon - quark coupling by constraining the pomeron - quark coupling to its phenomenological value @xmath0 . we then provide predictions for a variety of diffractive processes . as the propagators have been evaluated entirely from qcd first principles ( although in the quenched approximation ) , our results provide a consistency check of the landshoff - nachtmann model . we address the issue of the possible gauge - dependence of our results , which will be the object of a future study . edinburgh preprint : 95/559 + liverpool preprint : lth-363 + hep - lat/9511031 the landshoff - nachtmann pomeron on the lattice + _ ukqcd collaboration _ + * d.s . henty , c. parrinello and d.g . richards * + department of physics and astronomy , university of edinburgh , edinburgh eh9 3jz , scotland +
lagrangian formulations of general relativity abound in literature , but rather than differing in the choice of the lagrangian density they differentiate one another by the variable upon which the variation of the action has to be taken . the most famous examples are pheraps the metric and palatini variational approaches . the first one , sometimes called second order variation , considers the metric tensor as the only dynamical field variable while the second one , also known as first order variation , assumes the connection to be an independent field . though both approaches reproduce the einstein field equations in vacuum , the latter gives rise to another theory whenever the independent connection is allowed to appear inside the matter action . in this case the theoretical framework goes under the name of metric - affine variational principle and it is known not to recover general relativity in the presence of spinor fields , for example . the scope of the present paper is to summarize the basic features of a new variational approach recently proposed@xcite . together with the metric , two completely independent connections , with no _ a priori _ assumpions , are considered as dynamical variables entering the gravitational action for general relativity . the variation has thus to be taken independently with respect to the metric and both the connections . as we are going to show , applying this approach to a suitable generalization of the einstein - hilbert action leads to general relativity even when the matter action is allowed to depend on one of the connections . because of the presence of two connections and in analogy with bimetric thoeries of gravity , this approach will be called _ biconnection variational principle_. denote with @xmath0 and @xmath1 two independent connections , both containing torsion and non - metricity . in order to produce a new gravitational lagrangian within our framework , we consider a generalization of the riemann tensor obtained by symmetrizing the anticommutation of the covariant derivatives over the two connections . for any vector @xmath2 we have then @xmath3v^\alpha \ , , \label{001}\end{gathered}\ ] ] where @xmath4 , @xmath5 and @xmath6 , @xmath7 are the torsion tensors and covariant derivatives of @xmath8 and @xmath9 respectively . the biconnection riemann tensor @xmath10 has been defined as @xmath11 notice that whenever the two connections coincide this reduces to the usual ( metric - affine ) riemann tensor and ( [ 001 ] ) becomes the standard formula for the commutation of covariants derivatives , which in general includes a torsion term . the action we propose is thus @xmath12 where @xmath13 in analogy with the ricci tensor and the curvature scalar . when @xmath14 this reduces to the palatini ( or metric - affine ) action . to ( [ 002 ] ) we want to add a suitable matter action , where we have to make a choice upon the connection used to build covariant derivatives . in both metric and palatini approaches this connection is assumed to be the levi - civita connection , while in metric - affine theories this is taken to be the independent connection which prevents the dynamical equivalence with general relativity . since we will make the rather physical assumptions that all the matter fields couple to the same connection , in our case we have three choices : the levi - civita connection , @xmath8 or @xmath9 . the first choice would reduce the theory to palatini general relativity and thus will not be considered . because of the exchange symmetry @xmath15 of action ( [ 002 ] ) , it does not matter which one among this two connections enters the matter action . we choose @xmath8 . the total action of the theory is thus @xmath16 where @xmath17 denotes all matter fields collectively . variation with respect to @xmath9 eventually yields the condition has to be considered here because of a projective symmetry of ( [ 002 ] ) . see ref . for full details . ] @xmath18 which means that @xmath8 reduces to be the levi - civita connection . the role of @xmath9 is thus similar to the one played by a lagrange multiplier field . the variation with respect to @xmath8 leads , after one takes into account condition ( [ 003 ] ) , to @xmath19}_\sigma\right ) \ , , \label{004}\ ] ] where @xmath20 and @xmath21 is the variation of the matter action with respect to @xmath8 , the so - called _ hypermomentum_. finally , variation of action ( [ 005 ] ) with respect to the metric tensor produces , again after having taken into account ( [ 003 ] ) , the field equations @xmath22 where @xmath23 is the standard covariant derivative of the levi - civita connection , @xmath24 is the einstein tensor and @xmath25 is the variation of the matter action with respect to the metric . note that this differs from the usual energy - momentum tensor @xmath26 of general relativity inasmuch as it lacks of the variation of the connection with respect to the metric . using now equations ( [ 004 ] ) and some algebra , we can rewrite field equations ( [ 006 ] ) as @xmath27 the divergence in the right hand side of ( [ 007 ] ) represents exactly the missing variation of the connection with respect to the metric in the tensor @xmath28@xcite . we thus have that ( [ 007 ] ) simply reduces to @xmath29 which are nothing but the _ einstein field equations_. to conclude , the biconnection variational approach for general relativity we have presented here , is physically equivalent to the more standard metric formulation of the theory . it allows the connection employed to minimally couple matter fields to be completely independent of the metric . in this respect , the biconnection approach is similar to the metric - affine one but , in contrast with the latter one , it always permits to reduce the equations of motion to the einstein field equations and thus do not alter the phenomenology of gravitation . applications of the biconnection formulation of general relativity can be relevant in quantum gravity theories , such as loop quantum gravity , which assume a metric - affine classical framework .	a recently proposed variational approach for general relativity where , in addition to the metric tensor , two independent affine connections enter the action as dynamical variables , is revised . field equations always reduce to the einstein field equations for any dependence of the matter action upon an independent connection .
gamma - ray burst ( grb ) 130427a may well be the burst with the most comprehensive afterglow follow - up , its multiwavelength monitoring covering radio , optical , x - ray , and @xmath2-ray frequencies , and extending from seconds to tens of days after trigger . the x - ray _ prompt _ emission ( up to 100 s ) was accompanied by the second brightest optical flash , monitored by raptor ( vestrand et al 2013 ) , with the optical afterglow light - curve displaying a steepening at 300 s and a flattening at 10 ks . the swift x - ray light - curve ( x - ray light - curve repository evans et al 2009 ) is consistent with a single power - law from 500 s to 5 ms . the fermi - lat @xmath2-ray light - curve ( tam et al 2013 ) displays a peak at 1020 s , simultaneous with the optical flash peak , and a steepening at 550800 s ( zhu et al 2013 ) . the vla radio light - curves ( laskar et al 2013 ) display a slow decay at 1 - 10 day . with such a rich dataset , grb afterglow 130427a demands a theoretical interpretation , done here in the framework of the external - shock model ( & rees 1997 ) where some relativistic ejecta , produced by the black - hole resulting from the core - collapse of a massive star , drive a _ forward - shock _ into the ambient medium while the ejecta are energized by the _ reverse - shock_. the synchrotron and inverse - compton emissions from both shocks are calculated assuming that electrons and magnetic field acquire a certain fraction of the post - shock energy . the shock - accelerated electrons are assumed to have a power - law distribution with energy ( hence the synchrotron and inverse - compton spectra are also power - laws ) , with a break at the cooling energy ( where the radiative - loss timescale equals the shock age ) . analytical treatments for the forward - shock emission have been provided by & rees ( 1997 ) , sari , piran & narayan ( 1998 ) , waxman , kulkarni & frail ( 1998 ) , granot , piran & sari ( 1999 ) , wijers & galama ( 1999 ) , chevalier & li ( 2000 ) , panaitescu & kumar ( 2000 ) , and for the reverse - shock by kobayashi ( 2000 ) . both shocks have been studied with 1-dimensional hydrodynamical codes by panaitescu & ( 1998 ) and kobayashi & sari ( 2000 ) , the former focusing on the two - shock synchrotron and inverse - compton emission , the latter on the dynamics of the shocks . to model the multiwavelength emission of grb afterglow 130427a , we employ a 1-dimensional code that follows the ejecta medium interaction , with the dynamics of each shock calculated from conservation of energy and using the shock jump - conditions ( blandford & mckee 1976 ) . after the onset of deceleration , the dynamics of the forward - shock is determined by the ejecta initial energy , injected energy ( rees & 1998 ) , and ambient medium density . the dynamics of the reverse - shock is determined by that of the shocked fluid and two properties of the incoming ejecta : their energy and lorentz factor . here , we consider that the ejecta add energy to the blast - wave as a power - law in observer time and that they have a single lorentz factor . the self - absorption and cooling frequencies of the synchrotron spectrum and the inverse - compton parameter are calculated self - consistently from the electron distribution and the magnetic field strength ( panaitescu & 2000 ) . radiative losses are also calculated from the electron distribution , but they are negligible for the following best - fit models . the emissions from both shocks are integrated over their motion and over the angle at which the fluid moves relative to the direction toward the observer . more details about this numerical model and its application to the multiwavelength emission of ten grb afterglows are given in panaitescu ( 2005 ) . the choice of the model features that may accommodate the temporal decay of the broadband emission of grb afterglow 130427a starts with the x - ray light - curve because its temporal decay index ( @xmath3 ) and spectral slope ( @xmath4 ) are the best determined : @xmath5 at 20 ks 5 ms ( fig [ homog ] ) and @xmath6 at mean time 24 ks ( kennea et al 2013 ) . these lead to @xmath7 , which is compatible with the value expected ( zero ) for the synchrotron emission from the forward - shock interacting with a homogeneous medium and for the x - ray being below the cooling frequency @xmath8 of the synchrotron spectrum . as the optical flux also decays at that time , the optical must be above the peak energy @xmath9 of the synchrotron spectrum , hence optical and x - ray are in the same spectral regime : @xmath10 . consequently , the intrinsic afterglow optical flux can be calculated from the x - ray flux : @xmath11 . for instance , the observed @xmath12jy implies that @xmath13 mjy , which is a factor 2.5 larger than the measured @xmath14 mjy , requiring @xmath15 mag of dust extinction in the host galaxy . the 10 kev 100 mev spectral slope @xmath16 indicates that @xmath8 is well above 10 kev . that the lat flux decays slower than in the x - ray ( @xmath17 at 500 s 50 ks ) indicates that @xmath8 is below the lat range ( otherwise , for @xmath18 mev , the model expectation is @xmath19 ) and that the electron radiative cooling is dominated by inverse - compton scatterings ( otherwise , for synchrotron - dominated electron cooling , @xmath20 and @xmath21 , incompatible with the observed @xmath22 ) . more exactly , for a compton - dominated electron cooling , the decay index of the synchrotron flux above @xmath8 is @xmath23 , which matches well the observed @xmath24 , with @xmath25 being the exponent of the power - law distribution of electrons with energy ( @xmath26 ) that is required by the forward - shock model , given the measured flux decay index @xmath27 below the cooling frequency . in summary , the optical , x - ray , and @xmath2-ray fluxes of grb afterglow 130427a , their decay indices , and the x - ray spectral slope , require that @xmath28 , if the afterglow emission is synchrotron from the forward - shock . under the assumption that the two microphysical parameters of the forward - shock ( @xmath29 and @xmath30 ) that quantify the post - shock fractional energy in the magnetic field and in electrons are constant , the forward - shock synchrotron light - curve at any frequency below the optical can be easily calculated from the optical light - curve , using the expected evolution of the synchrotron peak flux ( @xmath31 ) and peak energy ( @xmath32 ) for a homogeneous medium . if @xmath9 crosses the optical at some time @xmath33 , yielding an optical flux @xmath34 , then the radio flux at frequency @xmath35 is @xmath36 here , @xmath37 mjy is the intrinsic optical light - curve after 10 ks ( corrected for the above - inferred host extinction of @xmath15 mag ) and @xmath38 ( the forward - shock model requires that @xmath39 ) . the largest @xmath33 required by equation ( [ radio ] ) arises from the radio measurement with the highest @xmath40 ; taking the @xmath41 mjy measurement as an upper limit for the forward - shock radio flux , implies @xmath42 ks . this means that , for the forward - shock emission ( that accommodates the observed optical flux ) not to exceed the measured radio fluxes , the synchrotron peak should cross the optical at 23 ks . conversely , if the synchrotron peak crossed the optical before 23 ks , then the synchrotron flux from the forward - shock would violate vla measurements . that may be avoided if the magnetic field parameter @xmath29 decreases ( roughly as @xmath43 ) , and if energy injection in the forward - shock is allowed ( to match the optical and x - ray flux decays at 110 day , which are faster when @xmath29 decreases ) , but this scenario requires fine - tuning and we do not pursue it . fig [ homog ] illustrates the failure of the forward - shock synchrotron model with a homogeneous medium to accommodate simultaneously the radio , optical , x - ray , and @xmath2-ray fluxes of grb afterglow 130427a : while it can explain the optical afterglow emission after 10 ks and the x - ray flux at 50 s 5 ms ( excluding the second grb pulse , which is a feature that can not be accounted for by any type of external shock ) , this model over - predicts either the lowest or the highest frequency data . the closure relation @xmath7 is also compatible with the forward - shock model expectation for a wind - like medium ( with an @xmath44 particle density distribution with radius ) and for x - ray below the cooling frequency , provided that there is an energy injection in the forward - shock that slows its deceleration and the decay of the afterglow x - ray flux . if that energy injection is parametrized as @xmath45 in observer time ( a power - law flux decay requires that the dynamics of the forward - shock is a power - law in observer time ) , then @xmath46/2 $ ] , from where @xmath47 . numerically , we find that the best - fit to the x - ray emission after 500 s ( including all the gev data and the optical after 10 ks ) has @xmath48 and that forward - shock energy should increase by a factor @xmath49 until @xmath50 ms , to account for the observed x - ray flux decay . that means that the energy added to the forward - shock mitigates its deceleration after @xmath51 ks . it is important to note that the forward - shock interacting with a wind - like medium does not produce more radio emission than measured because the synchrotron peak flux decreases as @xmath52 ( instead of being constant , as for a homogeneous medium ) . the evolution of the synchrotron peak energy is the same as for a homogeneous medium ( @xmath32 ) , hence the radio flux expected from the optical emission is @xmath53 then , @xmath54 mjy and @xmath55 mjy require that the time when the synchrotron peak crosses the optical is @xmath56 ks . that brief flattening seen in the optical light - curve at 10 ks could be due to @xmath9 crossing the optical and is compatible with @xmath56 ks . the best - fit to the optical data after 10 ks , the x - ray after 500 s , and all gev measurements , with the forward - shock emission and for a wind - like medium is shown in fig [ wind ] , with a sequence of spectra shown in fig [ spek ] . the @xmath57 for 135 dof of that best - fit makes it statistically unacceptable ; the gev fit has the largest @xmath58 for 9 points , closely followed by the optical fit s @xmath59 for 39 points , with the largest contribution to the fit s @xmath60 arising from the x - ray data , @xmath61 for 79 points . the model light - curves follow well all flux trends and relative intensities except the brightness of the prompt emission until 50 s , but can not describe well the early gev light - curve and can not capture the fluctuations in the x - ray and optical measurements ( after 10 ks , optical data are from different instruments ) . compared to the parameters inferred for other afterglows by modelling their multiwavelength emission , the wind density of the best - fit shown in fig [ wind ] is very small , but not unprecedented ( chevalier , li & fransson 2004 ) . its parameter , @xmath62 , corresponds to a stellar mass - loss rate to terminal wind - velocity ratio ( @xmath63 ) that is 300 smaller than for a typical wolf - rayet ( wr ) star ( as the progenitor of long bursts with an associated type ic supernovae ) , for which @xmath64yr and @xmath65 cm / s . the reason for that low density is the requirement that the synchrotron peak crosses the optical after 3 ks and matches the optical flux detected at that time . for @xmath66 and for the fluid moving directly toward the observer , the forward - shock synchrotron peak energy and peak flux are @xmath67 @xmath68 imposing that @xmath69 ev and @xmath70 mjy , yields @xmath71 taking the ratio of these two equations leads to @xmath72 . the @xmath30 parameter that quantifies the typical electron energy corresponds to a total electron energy that is a fraction @xmath73 of the post - shock energy . equipartition with protons sets an upper limit , @xmath74 , thus @xmath75 for @xmath76 , from where @xmath77 . this wind density is about 20 times lower than the lowest value measured ( nugis & lamers 2000 ) for galactic wr stars and indicates a low mass loss - rate combined with a high wind velocity . provided that can happen at the end of a wr s life , it has a strong consequence on the medium in which that star resides , as following . owing to low wind density and high ejecta kinetic energy , the forward - shock that fits the late time broadband emission of grb afterglow 130427a is highly relativistic , having @xmath78 , hence the shock radius is @xmath79 pc . requiring that @xmath80 at the latest observation epoch ( 50 day ) is less than the size of the bubble blown by a wr star during its @xmath81 yr lifetime , @xmath82 pc ( cf . castor , mccray & weaver 1975 ) , with @xmath83 the medium density around the star , we find that @xmath84 for a wind with @xmath85 , where @xmath86 is the observer - frame epoch when the afterglow shock encounters the wind termination shock . such a low ambient density suggests that the progenitor of grb 130427a occurred in a supper - bubble ( scalo & wheeler 2001 ) blown by many preceding supernovae . there are two interesting facts related to the lat emission produced by the forward - shock synchrotron model shown in figs [ homog ] [ spek ] . first is that the scattering of the synchrotron emission ( at the peak of the spectrum ) by the forward - shock electrons ( of typical energy ) occurs near the klein - nishina ( kn ) regime . when the electron cooling is dominated by inverse - compton scatterings ( i.e. compton parameter @xmath87 ) , the cooling frequency satisfies @xmath88 . inclusion of the kn effect reduces the compton @xmath89 parameter , thus , taking into account the kn effect , increases @xmath8 and the synchrotron flux at @xmath90 : @xmath91 . in other words , the synchrotron emission from fast - cooling electrons increases when a competing radiative process ( inverse - compton ) is reduced ( by inclusion of the kn effect ) . for the forward - shock best - fit to grb afterglow 130427a , the lat range is above @xmath8 and @xmath87 ; inclusion of the kn effect reduces @xmath89 by about 10 and increases the 100 mev flux by an order of magnitude ( see fig [ wind ] ) . furthermore , as the electrons at the peak of their distribution with energy enter and exit the kn regime , the synchrotron light - curve at 100 mev displays more structure than when the kn effect is ignored . the second is that radiative cooling during one gyration time limits the energy that electrons acquire through first - order fermi acceleration to a corresponding synchrotron characteristic energy @xmath92 mev , independent of the magnetic field @xmath93 . for the best - fit parameters given in fig [ wind ] , the forward - shock has @xmath94 and @xmath95 , so the maximal synchrotron energy is @xmath96 mev ( see synchrotron spectrum cut - off in fig [ spek ] ) . at earlier times , that cut - off is higher , but the inverse - compton emission from the forward - shock takes over above 2 gev ( as shown by the @xmath97 s spectrum ) and can account for the higher - energy lat emission until about 10 ks , after which the inverse - compton flux is too low . interestingly , a hardening of the lat spectrum above several gev was identified by tam et al ( 2013 ) , from @xmath98 at 0.15 gev to @xmath99 at 5100 gev . tam et al ( 2013 ) have proposed that the harder high - energy component is inverse - compton , although we find that the observed spectrum above 5 gev is softer than the model expectation @xmath100 , corresponding to the gev range being below the peak of the upscattered spectrum . the estimation of the expected radio emission given in equation ( [ radios2 ] ) led to the conclusion that the forward - shock can not account for the optical afterglow emission prior to @xmath101 ks . also , the flat radio light - curve arising from the forward - shock interacting with a wind can not account for the radio emission at 110 day , which is slowly decaying . both these emissions are attributed to the reverse - shock ( see also laskar et al 2013 ) , as discussed below . we note that , after 10 ks , the existence of a reverse - shock is required by the energy injection into the forward - shock required by the measured decay index of the x - ray flux . the radio data are contemporaneous with the higher energy ( optical , x - ray , and @xmath2-ray ) afterglow emission accommodated by the forward - shock , thus , for the calculation of the reverse - shock emission , the dynamical parameters @xmath102 , @xmath103 , @xmath104 , and @xmath105 are fixed at the values determined from the forward - shock best - fit . the free parameters of the reverse - shock are the lorentz factor @xmath106 of the incoming ejecta ( which sets the post - shock energy density ) and the three microphysical parameters ( @xmath29 , @xmath30 , and @xmath107 ) that determine the synchrotron spectrum . the best - fit obtained with the reverse - shock emission to the 110 day radio data is shown in figs [ wind ] and [ spek ] . unfortunately , it has a large @xmath108 for 25 dof , because it underestimates the radio flux above 50 ghz . as shown in fig [ spek ] , those radio data can not be explained by the forward - shock either , if its microphysical parameters are constant . requiring the same microphysical parameters for both the reverse and forward shocks yields a much worse radio data fit , with @xmath109 . the best - fit to the early optical emission with a reverse - shock includes also the earlier x - ray data and all gev data , to ensure that the reverse - shock emission does not exceed what was observed . again , the dynamical parameters @xmath102 and @xmath105 are fixed to the values obtained for the forward - shock , but the energy @xmath110 carried by the incoming ejecta arriving at the blast - wave prior to 10 ks is only weakly constrained by the forward - shock fit to the optical and x - ray data after 10 ks , which sets an upper limit @xmath111 . with free micro - parameters , the best - fit with the reverse - shock to the early afterglow has @xmath112 for 136 dof , as it fails to account for the gev prompt emission prior to 100 s , although it explains well the early optical data and the x - ray data at 0.53 ks . we note that the reverse - shock magnetic parameter @xmath29 prior to 10 ks ( from fitting the early optical afterglow ) is 100 times larger than after 10 ks ( from modelling for the radio emission ) . if the reverse - shock microphysical parameters were held constant across 10 ks , then the fit to the radio emission would have a @xmath113 twice larger , thus a decrease in @xmath29 at 10 ks is required . that may mean that the ejecta arriving at the blast - wave later are less magnetized . as discussed in [ sypeak ] , to reconcile the radio and optical fluxes of afterglow 130427a requires that the peak of the forward - shock synchrotron crosses the optical after 3 ks . in turn , that requires ( [ tenouswind ] ) a very weak stellar wind , about 300 less tenuous ( @xmath85 ) than for the average galactic wr star . then , the forward - shock radius is @xmath114 pc , while the wind bubble radius should be @xmath115 pc . thus , if the circumstellar medium is sufficiently dense , it possible that @xmath116 . alternatively , if the stellar wind had the average density , the wind termination shock could be encountered by the forward - shock at 3 ks , provided that the burst is embedded in a hot , highly pressurized environment ( chevalier et al 2004 ) . at frequencies below the cooling break , the afterglow light - curve should display a flattening when the forward - shock crosses the wind termination shock , transiting from the @xmath117 free wind to the quasi - homogeneous shocked wind . to be self - consistent , the interpretation of the 3 ks optical light - curve flattening as the blast - wave encountering the wind termination shock should attribute the entire afterglow emission to the same shock . then , the peak of the synchrotron spectrum must be below optical at all times when a decaying optical flux is measured , a model which overproduces radio emission , if the optical afterglow originates in the forward - shock ( as shown in [ sypeak ] ) . the subsequent steepening of the optical light - curve at 20 ks can not originate in the ambient medium stratification because , outside the termination shock , the shocked wind and circumstellar medium are still homogeneous . instead , that light - curve steepening should be attributed to the cooling frequency falling below the optical , which yields a steepening of the power - law flux decay by @xmath118 ( consistent with that measured for the optical light - curve of 130427a at 20 ks ) , and a softening of the optical spectrum by @xmath119 ( consistent with the reddening reported by perley et al 2013 , after 10 ks ) . the fortuitous temporal coincidence of the cooling frequency falling below optical just after the blast - wave arrives at the free - wind termination shock is not required if the discontinuity in the ambient medium structure that yields the 3 ks optical light - curve flattening is caused by an internal interaction within an unsteady stellar wind or by the interaction between the winds of two stars . in the former scenario , considered analytically by chevalier & imamura ( 1983 ) and in the context of grb afterglows by ramirez - ruiz et al ( 2005 ) , a stronger wind produced by the wr star prior to its core - collapse interacts with a slower wind ejected previously . in the latter scenario , proposed by mimica & giannios ( 2011 ) to be the source for more diverse afterglow light - curves , the grb progenitor is in a dense stellar cluster , where the mean distance between stars is below 1 pc , and the wr wind interacts with the weaker wind of a nearby o star or a later type . in either scenario , after the interaction with the shocked wind(s ) , which yields a light - curve flattening , the blast - wave goes into an @xmath117 wind , which is the earlier wr wind or the wind of the nearby star , producing a light - curve steepening , with the flux decay index @xmath120 returning to the value it had during the interaction with the free wr wind . only the dense cluster scenario provides a natural explanation for the very weak wind inferred here from modelling the afterglow 130427a : the wind of a b star located within 1 pc of the grb progenitor . however , this scenario can not explain why that weak wind extends over tens of pcs ( as required by the duration of the afterglow , [ tenouswind ] ) despite the more powerful winds of nearby , earlier type stars . thus , the 3 ks flattening and 20 ks steepening seen in the optical light - curve of 130427a could originate from a forward - shock interacting with the more complex ambient medium resulting from an internal wind interaction provided that microphysical parameters evolve such that this model does not exceed the 110 day radio measurements . alternatively , radio emission is not overproduced if the entire afterglow emission arises from the reverse - shock , and the optical light - curve flattening and steepening could result from the changing dynamics of the reverse - shock when the shocked - wind shell is crossed . such light - curve features could also be due to variations in the density and lorentz factor of the incoming ejecta , without any need for a non - uniform ambient medium . however , a model where the entire afterglow emission arises from the same shock ( reverse or forward ) does not provide a natural explanation for the colour evolution displayed by 130427a , which becomes bluer after 3 ks ( vestrand et al 2013 ) , when the optical light - curve flattens , and redder after 10 ks ( perley et al 2013 ) , when the optical light - curve steepens . in contrast , the hybrid reverse - forward shock model explains naturally both the 3 ks spectral hardening , as due to the harder ( below the synchrotron peak energy ) forward - shock emission emerging from under the softer reverse - shock emission , and the following spectral softening , caused by the peak energy of the forward - shock synchrotron spectrum falling below optical . the closure relations expected between the forward - shock synchrotron flux decay index and spectral slope suggest a homogeneous ambient medium for grb afterglow 130427a . although long grbs arise from massive stars that drive powerful winds , a homogeneous medium is possible if the afterglow emission is produced in the shocked wind . however , this afterglow s ( 10 ks ) optical flux to ( 110 day ) radio flux ratio and its slowly decaying radio light - curves disfavour that type of mbient medium . instead , for an unevolving constant magnetic field parameter , the synchrotron spectrum peak flux is constant and the radio emission should have been brighter and slowly rising . with some fine - tuning of the evolutions of those micro - parameters , it may be possible to reduce the forward - shock model radio flux below measurements , while still accounting for the observed optical and x - ray light - curves . a wind - like medium ( @xmath44 ) is the more natural expectation for a massive star as the grb progenitor . the forward - shock emission still can not account for the radio data because the expected radio light - curve is flat , however , a wind - like medium yields a decreasing synchrotron spectrum peak flux , making it easier to keep the forward - shock radio emission below radio measurements . to explain the optical and x - ray flux decay after 10 ks with the forward - shock synchrotron emission , a moderate energy injection into the forward - shock is required , increasing the shock energy by a factor 4 from 10 ks to 1 ms . the agent of that energy injection should be some ejecta that arrive at the forward - shock at that time , which provides a natural explanation for the afterglow radio emission : the reverse - shock that crosses the incoming ejecta . the reverse - shock must have been operating at even earlier times because the high early - optical to late - radio flux ratio precludes a forward - shock origin of the optical afterglow emission prior to 10 ks . such a reverse - to - forward shock switch for the origin of the optical emission , occurring at few ks , is supported by the optical afterglow becoming bluerat that time ( vestrand et al 2013 ) , when the forward - shock emission , with a spectrum @xmath121 in the optical , begins to dominate the softer reverse - shock emission , with a spectrum @xmath122 . as the peak of the forward - shock synchrotron spectrum falls below optical at about 10 ks , the optical afterglow should become redder after 10 ks , as was observed by perley et al ( 2013 ) . however , for the reverse - shock to explain the 100 s few ks optical afterglow and the 110 day radio afterglow emission , the properties of the reverse - shock ( microphysical parameters , kinetic energy and lorentz factor of the incoming ejecta ) must change around 10 ks . furthermore , vestrand et al ( 2013 ) have shown that the reverse - shock can also account for the optical flash ( up to 100 s ) and the gev light - curve peak , but for microphysical different than after that peak . for this hybrid reverse - forward shock model , we find that the x - ray flux of grb afterglow 130427a is accounted mostly by the forward - shock emission , from the tail of the first grb pulse ( 50100 s ) up to 5 ms , excluding the second grb pulse at 100 - 500 s. the reverse - shock may have had a significant contribution to the early x - ray emission , at 500 s 2 ks . both shocks give comparable gev emissions . as shown in fig [ wind ] , the radio emission from the forward - shock is expected to overshine that from the reverse - shock at 30 day ( or somewhat later , if energy injection continues after 1 ms ) , yielding a flat flux @xmath123 mjy until @xmath124 day , when the peak of the synchrotron spectrum falls below 10 ghz . if that flat radio flux is not seen , then the magnetic field parameter @xmath29 of the forward - shock must be decreasing , so that the peak flux of the forward - shock synchrotron spectrum falls - off faster than the @xmath52 expected for @xmath125const . the relative dimness of the radio afterglow suggests that the peak of the synchrotron spectrum has crossed the optical range at 10 ks . an immediate consequence is that the wind - like ambient medium is a factor 20 less dense than the most tenuous wind measured for galactic wr stars . we can not provide a good argument for why grb 130427a s progenitor had such a low mass - loss rate to wind - speed ratio ( @xmath126 ) , but note that , owing to the weak wind , the afterglow remains highly relativistic and travels @xmath127 pc until the last observation epoch ( 50 day ) . for such a large afterglow radius to remain inside the free wr wind ( i.e. within the wind termination shock ) , the grb progenitor must have been embedded in a very tenuous medium , suggesting a supper - bubble blown by preceding supernovae and stellar winds . owing to tenuous ambient medium , the afterglow transverse size , @xmath128 pc , is unusually large , and implies a source apparent diameter of @xmath129 mas , which may be resolved with radio interferometry . if the gev emission of grb afterglow 130427a arises from the forward - shock , then the up - scattering of the synchrotron emission occurred at the onset of the kn regime , where the reduction of the electron scattering cross - section lowers the compton parameter , increases the synchrotron cooling - break frequency , and increases the synchrotron flux above that break ( i.e. in the lat range ) . furthermore , lat must have measured the forward - shock inverse - compton emission at photon energies above a few gev . this work was supported by an award from the laboratory directed research and development programme at the los alamos national laboratory and made use of data supplied by the uk swift science data centre at the university of leicester .	+ the complex multiwavelength emission of grb afterglow 130427a ( monitored in the radio up to 10 days , in the optical and x - ray until 50 days , and at gev energies until 1 day ) can be accounted for by a hybrid reverse - forward shock synchrotron model , with inverse - compton emerging only above a few gev . the high ratio of the early optical to late radio flux requires that the ambient medium is a wind and that the forward - shock synchrotron spectrum peaks in the optical at about 10 ks . the latter has two consequences : the wind must be very tenuous and the optical emission before 10 ks must arise from the reverse - shock , as suggested also by the bright optical flash that raptor has monitored during the prompt emission phase ( @xmath0 100 s ) . the vla radio emission is from the reverse - shock , the swift x - ray emission is mostly from the forward - shock , but the both shocks give comparable contributions to the fermi gev emission . the weak wind implies a large blast - wave radius ( @xmath1 pc ) , which requires a very tenuous circumstellar medium , suggesting that the massive stellar progenitor of grb 130427a resided in a super - bubble . radiation mechanisms : non - thermal relativistic processes shock waves
conventional bottomonium is an approximately non - relativistic system . out of 34 expected @xmath11 energy levels below the @xmath4 threshold @xcite 15 have been observed @xcite . the masses and decays of these states are well described by potential models @xcite . there are five states with @xmath12 pairs above the @xmath4 threshold : three isospin - zero vector states @xmath13 , @xmath14 and @xmath15 [ or @xmath0 , @xmath1 and @xmath5 according to the potential model assignment ] , and two isospin - one axial states @xmath16 and @xmath17 . the isospin - one states are obviously exotic with minimal quark content @xmath18 . but the isospin - zero states also have properties unexpected for a pure @xmath12 pair . the mass splitting between the @xmath0 and @xmath1 is larger by @xmath19 than that between the @xmath20 and @xmath0 , while for a pure @xmath12 system it is expected to be smaller by about @xmath21 @xcite . the rates of @xmath22 and @xmath23 transitions are two orders of magnitude higher than expected for a pure bottomonium @xcite . the @xmath24 transitions , that in a pure bottomonium involve spin of heavy quark and are suppressed by three orders of magnitude relative to @xmath25 transitions , are not strongly suppressed in case of the @xmath1 @xcite and are even enhanced in case of the @xmath0 @xcite . more puzzling results on hadronic transitions are discussed below . thus , there is a strong evidence that the structure of the vector states above the @xmath4 threshold is more complicated than the pure @xmath12 , i.e. it is exotic . in this paper we briefly review the experimental status of the `` botomonium - like '' states , their interpretation and prospects for future studies . our main goal is to discuss what we can learn about them based on the increased @xmath0 and @xmath1 data samples at belle - ii , and what dedicated data taking is needed . we propose to perform a high - statistics energy scan from the @xmath4 threshold up to the highest possible energy to measure exclusive open and hidden flavor cross sections . we also propose to collect data at the peak of @xmath5 and at higher mass states if they are found in the energy scan to study their properties and to search for missing bottomonia , @xmath26-wave @xmath3 mesons or exotic states partners of the @xmath16 and @xmath17 . favored interpretation of the bottomonium - like states is related to the presence of heavy mesons in their wave functions . the @xmath16 and @xmath17 states have purely molecular structures , @xmath27 and @xmath28 , respectively , while the @xmath0 , @xmath1 and @xmath5 states are mixtures of the @xmath12 and @xmath29 pairs [ with @xmath30 we denote @xmath26-wave excitations of @xmath2 or @xmath3 mesons ] . we discuss measurements that are needed to establish this interpretation and point out strategy to search for other types of exotic hadrons compact tetraquarks @xcite or hadroquarkonia @xcite . the paper is organized as follows . we first discuss the results obtained using on - resonance data , since they constitute the bulk of our current knowledge about the bottomonium - like states , and then move to the discussion of the energy scans . we conclude with a summary of the physics program for the proposed dedicated data taking . the results on bottomonium - like states presented in this section are based on large data samples collected by the belle and babar experiments . these are @xmath31 and @xmath32 collected by belle and babar , respectively , at the @xmath0 , and @xmath33 collected by belle at the @xmath1 . in the @xmath5 region only scan data are available five points with a total luminosity of @xmath34 that correspond to @xmath35 collected in the peak . known hadronic transitions from vector bottomonium and bottomonium - like states and corresponding partial widths are presented in table [ tab : partial_widths ] final states , since they proceed entirely via the intermediate @xmath36 states @xcite . we do not subtract the @xmath36 contributions from the @xmath37 final states ; they should be relatively small in all transitions except @xmath38 . ] . . known hadronic transitions from vector bottomonium(-like ) states and corresponding partial widths . [ cols="<,^,^",options="header " , ] [ tab : thresholds ] the regions of @xmath39 and @xmath40 are basically within the current reach of the superkekb collider of 11.24gev . an increase of the energy by at least 100mev will give a possibility to explore the @xmath41 threshold region and to search for baryon - antibaryon molecular states . such states are almost certainly there , as can be judged from the charmonium sector @xcite . it is also quite important to find a high - mass state to serve as a source of missing bottomonia below the @xmath4 threshold and of molecular states partners of the @xmath16 and @xmath17 , as discussed in sections [ sec : zb_partners ] and [ sec : missing_bottomonia ] . the thresholds for various transitions are listed in tables [ tab : transitions_to_missing_bb ] and [ tab : zb_partners ] . the maximal threshold is at @xmath42 . in order to cover the molecular states at the @xmath43 thresholds the peak should be above @xmath44 . this is the region of @xmath45 , @xmath46 , @xmath47 and @xmath48 thresholds . finally , investigation of the complete region of resonances would require to increase the energy up to 12gev . in the next section we discuss the @xmath40 threshold , which is within the current reach of superkekb . based on the heavy quark symmetry one should expect existence in the @xmath51-flavored sector of analogs of the charmed strange @xmath52 and @xmath53 mesons : @xmath54 and @xmath55 . their expected masses can be estimated within the heavy quark expansion as 5715mev and 5763mev , respectively . these mesons can be produced in @xmath56 annihilation in pairs with the ground - state strange @xmath3 and @xmath57 mesons . moreover , an @xmath58-wave production is allowed for the pairs @xmath59 , @xmath60 and @xmath61 . the thresholds for the former two channels practically coincide at the c.m . energy of 11.13gev , while for the latter the threshold is expected to be at approximately 11.18gev , and one can expect , due to a presence of the @xmath58-wave , a measurable production cross section at energies just above the threshold . the final states in these processes necessarily contain @xmath3 and @xmath62 mesons , and the excited @xmath54 and @xmath55 strange bottom mesons can be sought for by studying the spectrum of the invariant mass recoiling against a reconstructed @xmath3 meson . another interesting possibility may potentially arise if there is a molecular @xmath58-wave resonance , @xmath63 , near the threshold of the @xmath59 and @xmath60 pairs , i.e. near 11.13gev . an existence of such a resonance ( or a double resonance ) can be affected by the coincidence of the thresholds for the two types of pairs ( which is not the case for the final state @xmath61 ) . one can expect that such a resonance , in addition to decay channels with open @xmath51 flavor , would have a small , but possibly measurable , branching fraction for decay into bottomonium and light mesons : e.g. @xmath64 ( with @xmath65 ) , or @xmath66 . [ the former decay is somewhat similar to the part of decays @xmath67 not associated with the @xmath36 resonances . the branching fraction for this part is in the ballpark of ( a few)@xmath68 . ] thus a scan of these channels with hidden strangeness near the c.m . energy of 11.13gev can reveal existence of molecular resonances with the excited strange bottom mesons - wave production in @xmath56 annihilation of the pairs @xmath69 and @xmath70 at and/or above their common threshold at 4.43gev . the available data of a study using initial state radiation at belle @xcite indicate a certain activity in the final state @xmath71 above 4.4gev with possible structures . however more detailed data on the yield of strange @xmath72 mesons and/or of @xmath71 in that energy range are needed for further conclusions . ] . it can be also mentioned that @xmath58-wave decays into @xmath73 are also allowed for @xmath63 , in analogy with the decay @xmath74 , although it is not clear at present , how strong the breaking of hqss , that is required for such decays , is . however , if these transitions from @xmath63 proceed at a measurable rate , they can include the decay @xmath75 and thus provide a unique gateway to studies of the @xmath76 state of bottomonium . all hidden - beauty hadrons above the @xmath4 threshold have properties inconsistent with their interpretation as pure @xmath12 bottomonia . the vector bottomonium - like states @xmath13 , @xmath14 and @xmath15 [ or the @xmath0 , @xmath1 and @xmath5 ] are likely mixtures of conventional @xmath12 bottomonia and pairs of @xmath4 or @xmath78 mesons in ground or excited states . the isospin - one axial states @xmath16 and @xmath17 likely have purely molecular structures of @xmath27 and @xmath28 , respectively . we propose a dedicated data taking program that will establish the above interpretation , check its predictions and search for new bottomonium and bottomonium - like states . we propose to perform a high - statistics energy scan from the @xmath4 threshold up to the highest possible energy , with @xmath79 per point and @xmath80 step . these data will allow to measure cross sections of @xmath81 annihilation into exclusive open - flavor and hidden - flavor final states , such as @xmath4 , @xmath27 , @xmath28 , @xmath82 , @xmath83 , @xmath84 , @xmath78 , @xmath85 , @xmath37 , @xmath86 , @xmath87 and others . combined analysis of these cross sections using coupled - channel framework will determine all basic properties of the vector bottomonium - like states , such as their pole positions and interchannel couplings . these results are crucial for establishing the nature of the vector bottomonium - like states . the hidden - flavor channels will be used to search for states that do not exhibit themselves in the total cross section or in open - flavor channels ; among such states could be compact tetraquarks and hadroquarkonia . the states with molecular admixture are naturally located near the corresponding threshold . the present energy limit of the superkekb accelerator of 11.24gev will allow to investigate the @xmath39 and @xmath40 threshold regions . increase of maximal energy by at least 100mev will allow to explore the @xmath41 threshold region and search for baryon - antibaryon molecular states , which should be there , as can be judged from the charmonium sector . the region of promising thresholds extends up to 12gev . observation of new vector states is of utmost importance for studies of molecular states partners of the @xmath16 and @xmath17 . transitions from high mass bottomonium - like states provide a unique way to produce purely molecular states . energy above 11.5gev is of special interest , as most of the relevant hadronic transitions become kinematically allowed . thus searches for the molecular states provide additional motivation for increase of the superkekb energy up to at least 11.5 11.6gev . at each new vector state we propose to collect @xmath88 to perform a detailed study of corresponding transitions , since they provide information on the structure of the state , to search for missing conventional bottomonia and to search for molecular states . a region near the @xmath40 threshold is also of interest to search for missing @xmath26-wave @xmath3 mesons . available data at @xmath5 are limited to @xmath34 that were taken at five scan points , with the effective luminosity of @xmath35 . only the most prominent transitions @xmath23 and @xmath89 are known . an observed pattern of transitions is somewhat different from that at the @xmath1 , despite that the two states are close in energy . more detailed comparison requires an increase of statistics at the @xmath5 and could provide interesting information on the structure of both states . given that many transitions were observed at the @xmath1 and there are still channels to be investigated , the list of measurements to be performed at the @xmath5 is quite long . even non - observation of some channels will be of interest for comparison of the @xmath1 and @xmath5 . the @xmath5 data will provide a rich physics output even for relatively small ammount of a few tens of @xmath90 . this makes @xmath5 an attractive option for the belle - ii initial data taking . we are grateful to simon eidelman for valuable comments . the work of m.b.v . is supported in part by u.s . department of energy grant no . de - sc0011842 and was performed in part at the aspen center for physics , which is supported by national science foundation grant phy-1066293 . r.v.m . acknowledges support from the russian science foundation ( grant no . 15 - 12 - 30014 ) . there are several difficulties with the determination of the @xmath1 and @xmath5 partial widths . first , in @xmath91 there is a sizeable continuum contribution , thus the denominator is likely overestimated . second , the on - resonance @xmath1 data were taken not in the maximum of the @xmath37 cross section , but about half way down the slope . similarly , the @xmath5 data were taken at five points distributed over the peak region . thus , in both cases measured @xmath92 is underestimated .	the belle - ii experiment is expected to collect large data samples at the @xmath0 and @xmath1 resonances to study primarily @xmath2 and @xmath3 mesons . we discuss what other data above the @xmath4 threshold are of interest . we propose to perform a high - statistics energy scan from the @xmath4 threshold up to the highest possible energy , and to collect data at the @xmath5 and at higher mass states if they are found in the scan . we emphasize the interest in increasing the maximal energy from @xmath6 to @xmath7 in the future . these data are needed for investigation of bottomonium and bottomonium - like states . ftpi - minn-16/28 + umn - th-3606/16 * bottomonium - like states : physics case for energy scan above the @xmath4 threshold at belle - ii * a. e. bondar@xmath8 , r. v. mizuk@xmath9 and m. b. voloshin@xmath10 + + + + + + +
clusters of galaxies are essential tools for developing our understanding of structure formation and for probing cosmological parameters . in particular , the redshift evolution of the cluster mass function is a sensitive diagnostic of @xmath2 , sufficiently sensitive that the existence of even one or two massive clusters at @xmath3 favors a low - density universe ( donahue _ et al . _ 1998 , ebeling _ et al . _ 2000 ) . this argument assumes gaussianity in the primordial fluctuations ; clusters are equally useful at constraining primordial non - gaussianity given an independent measure of @xmath2 ( robinson , gawiser & silk 1999 ) . precise estimates of either quantity will require large , unbiased samples of clusters at a range of redshifts . locally , unbiased samples are crucial for measuring @xmath4 and for estimates of @xmath2 that rely on the fair sample hypothesis , i.e. that the composition of clusters in terms of baryon fraction or mass - to - light ratio is representative of the rest of the universe ( evrard 1997 , carlberg _ et al . _ 1996 ) . finally , the evolution of the numbers of cluster - sized masses as a function of redshift constrains the cosmological constant and the dark energy equation of state ( tegmark 2001 ) . if clusters are to serve as measures of the clumping of mass , we must identify them directly from observations of the mass distribution . this is difficult , however , because the vast majority of their matter is dark . the traditional methods of discovering clusters rely on optical emission from galaxies ( _ e.g . _ abell , corwin & olowin 1989 ; zaritsky , _ et al . _ 1997 ) or x - ray emission from a hot intracluster plasma ( _ e.g . _ boehringer _ et al . _ 2000 ) . clusters with a lower mass - to - light ratio or more baryons may well be overrepresented in these samples . radiation from baryons is a complicated proxy for mass , and hydrodynamic baryon - cdm models have been proposed to study bias in clusters ( blanton _ et al . _ 1999 ) . the development of weak gravitational lensing techniques ( see mellier 1999 and bartelmann & schneider 2001 for reviews ) suggests a different approach : shear selection ( schneider 1996 ) . because all types of matter participate in lensing , deep , wide - field imaging will reveal in its shear pattern all mass concentrations regardless of mass - to - light ratio or baryon fraction . of course , no single technique will be completely unbiased . for example , shear selection , like optical selection and the sunyaev - zeldovich effect ( joy & carlstrom 2001 ) , will tend to be more sensitive to line - of - sight superpositions of unrelated structures , to which x - ray selection , with its dependence on the square of the density , is less vulnerable . with x - ray and ( until now ) shear selection , optical followup is still necessary to provide a redshift . comparison of differently selected samples will always be necessary . still , the baryon and photon independence provided by shear selection are powerful features which may produce surprising new samples . two `` dark '' mass concentrations , found through weak lensing analyses of apparently unrelated previously known clusters , have been reported ( erben _ et al . _ 2000 ; umetsu & futamase 2000 ) . due to the absence of associated galaxies , the redshift and therefore masses of those clumps remain unknown . we report here the first spectroscopically confirmed shear - selected cluster . we also introduce photometric redshift techniques into the selection of sources for the weak lensing analysis . the recent development of these techniques ( see connolly _ et al . _ 1995 ; hogg _ et al . _ 1998 ) greatly improves the promise of shear selection , as source galaxies may be divided into redshift bins to tune the sensitivity to lensing mass concentrations through a series of lens redshifts ( mass tomography ) . in this paper , we use photometric redshifts to demonstrate that the shear - selected mass concentration is at roughly the same redshift as the cluster galaxies . this has always been an assumption in lensing analyses of clusters , but this measurement implies that mass tomography is feasible . we took @xmath5 images of a `` blank '' field ( containing no known cluster ) centered at 23:47:46 + 00:57:42 ( j2000 ) using the big throughput camera ( btc , wittman _ et al . _ 1998 ) in photometric conditions at the blanco 4-m telescope of cerro tololo inter - american observatory in 1997 and 1998 . details of the data reduction , galaxy catalogs and seeing correction may be found in wittman _ et al . _ ( 2000 , w00 ) ; here we give only the basic parameters . the final images are roughly 40square with 0.43 sampling and 1.3 fwhm after point - spread function corrections , with object counts peaking at @xmath6 , @xmath7 , @xmath8 , and @xmath9 ( isophotal magnitudes ) . in the final catalog , each object s shape is a weighted mean of the shapes measured in the different filters , as described in w00 . in addition to the w00 magnitude cut of @xmath10 , we imposed a color cut of @xmath11 to emphasize blue field galaxies at higher redshift , a tactic often used in weak lensing analyses of known clusters ( tyson , wenk & valdes 1990 ) . using the method of fischer & tyson ( 1997 ) , we constructed a mass map from the remaining 31,000 galaxies ( figure 1 ) . a mass concentration stands out near the southwest corner ; its peak is significant at the 4.5@xmath12 level , based on mass maps of bootstrap resampled catalogs . the mass concentration disappeared , as it should , under two null tests : rotating each galaxy in the catalog by 45@xmath13 , and assigning the shape of a random different galaxy to each galaxy position . to check for any bias casued by proximity to the field edge , we simulated a field of the same size and plate scale filled with random galaxies , distorted the image as if lensed by a cluster of moderate ( @xmath14 km s@xmath0 ) velocity dispersion , convolved and added noise to match the seeing and noise of the observations , cataloged , selected sources , and made mass maps as for the observations . repeating this for a series of ten center - to - corner cluster positions , we found that the bias was small ( @xmath15 in peak density ) and in the sense of reducing , not enhancing , our sensitivity to mass concentrations near the corner . we then made a color composite image from the @xmath16 and @xmath17 images ( figure 1.5 , separate jpeg ) . a concentration of reddish elliptical galaxies appears near the position of the density peak ( 23:46:23.85 + 00:45:00.8 for the brightest cluster galaxy versus 23:46:24.0 + 00:43:58.8 for the density peak , a displacement of just over 1 ) . nothing appears at that position in the rosat database , and the nasa / ipac extragalactic database contains only one object in the area , a large spiral which appears to be a foreground field galaxy . we concluded that this candidate cluster was worthy of spectroscopic followup . we designed a single slitmask containing 26 objects believed to be cluster members , selected by avoiding blue objects and choosing suitably bright yellow / red galaxies in the color composite image in the area around the putative cluster . this slitmask was used with the low resolution imaging spectrograph ( oke _ et al . _ 1995 ) on the keck telescope on the night of nov . 23 , 2000 . a single 1200 sec exposure with the 300 l / mm grating was obtained ; the spectral resolution was 10 with a 1 arc - sec wide slit , and the region from 4000 to 8000 was covered for each object . two of the objects proved to be galactic m dwarfs , while 17 are members or near outliers of a cluster at z=0.276 . one is a foreground galaxy , and six are background galaxies . we assume an instrumental contribution to the velocity dispersion of 100 km / sec in the rest frame . while there are two outliers among the 17 possible cluster members , both sigma - clipping and the biweight estimator of beers , flynn & gebhardt ( 1990 ) yield a velocity dispersion estimate ( in the rest frame ) of @xmath18 km s@xmath0 . with the cluster redshift in hand , we need only the redshift distribution of the lensed sources to derive the cluster mass from the shear . in the past it has been difficult enough to estimate the mean of this distribution , but photometric redshifts can provide a detailed distribution , in principle even an appropriate weight for each source galaxy . in this paper , we use tangential shear as a function of source photometric redshift to estimate the lens redshift @xmath19 in a way independent of the spectroscopy . we demonstrate that the mass causing the shear signal is at roughly the same redshift as the cluster . for each galaxy observed in all four filters , we used the hyperz package ( bolzonella , miralles & pell 2000 ) to compute a redshift probability density function ( pdf ) . we multiplied this by another pdf computed from the galaxy s apparent magnitude , assuming that a schechter ( 1976 ) luminosity function with @xmath20 and @xmath21 ( folkes _ et al . _ 1999 ) holds at each redshift ( we assume @xmath22 km s@xmath0 mpc@xmath0 , @xmath23 , and @xmath24 throughout ) . the latter pdf is quite broad , but serves a valuable purpose by suppressing high - redshift peaks which often appear in the former pdf . the first and second moments of the final pdf product provide the estimated redshift and its statistical error . henceforth , we use a catalog of 26,000 galaxies which is defined not by the magnitude and color cuts that went into figure 1 , but by the requirement of detection in each of four filters leading to a photometric redshift . of these 26,000 galaxies , 13,000 were also in the initial catalog . the median @xmath25 in this catalog is 0.58 . we verified the accuracy of the photometric redshifts by comparison with spectroscopic redshifts of 31 galaxies in the range @xmath26 . this sample was composed of 19 galaxies with redshifts described in section [ sec - spec ] ( the ones detected in all four filters ) , plus another 12 galaxies ( @xmath27 ) in a different region of the 40field , kindly provided by r. guhathakurta . we find little bias , with @xmath28 on average , and an rms of only 0.059 in the same quantity . a detailed analysis of this photometric redshift method and its application to larger datasets is in preparation ( margoniner _ et al . _ 2001 ) . before examining the dependence of the shear on source redshift , we must first account for the equally large dependence on projected position relative to the lens . we are not interested in a detailed reconstruction of the lens ; rather , for a given @xmath29 we would like to collapse information from all sources at a wide range of projected radii into a single number characterizing the lensing signal at that source redshift . hence at a given @xmath29 we fit a very simple model , a singular isothermal sphere ( sis ) , to the radial dependence of the tangential shear ( centered on the peak in the mass map ) , and use the fitted amplitude and its uncertainty . the assumption of a particular profile should not introduce any bias as a function of @xmath29 . to test this , we also considered nfw ( navarro , frenk & white 1997 ) profiles . although our data can not constrain the scale radius @xmath30 , for fixed @xmath30 ( 225 kpc , bartelmann , king & schneider 2001 ) , the results do not change significantly from the sis case . we therefore choose the sis for simplicity . for an sis , the tangential shear @xmath31 , where @xmath32 is the projected mass density , @xmath33 is the critical density , and @xmath34 , @xmath35 , and @xmath36 are the angular diameter distances from observer to source , lens to source , and observer to lens , respectively . since only @xmath34 and @xmath35 are changing with source redshift , the amplitude of an @xmath37 fit to the tangential shear should grow with the distance ratio @xmath38 . figure 2 shows the fitted tangential shear ( at a fiducial radius of 1 mpc ) as a function of @xmath29 . it is consistent with zero for @xmath39 and increases monotonically for @xmath40 ( the upper limit is to avoid extrapolating too far beyond the range of spectroscopic verification of @xmath29 ) . the dotted line illustrates the shear expected from a lens at @xmath41 , with @xmath42 fit to the points . this is a good fit ( @xmath43 ) . we explore the range of @xmath19 allowed by the shear data by stepping @xmath19 through the range @xmath44 in steps of 0.025 , and refitting at each step . the probabilities corresponding to @xmath45 at each step are plotted in figure 3 . the median and mode of this distribution are at @xmath46 and 0.30 respectively , with a 68% confidence interval @xmath47 ( these numbers change by less than 0.01 when an nfw profile is used ) . the fit for @xmath48 is also shown in figure 2 ( dashed line ) . thus the lens roughly coincides with the cluster of galaxies in redshift as well as in right ascension and declination . this method can be used to estimate the redshift of any newly discovered lensing mass from the lensing data alone . such a procedure may well become a standard part of shear - selected cluster surveys . finally , we estimate the mass and mass - to - light ( m / l ) ratio using the best - fit projected mass . still assuming an isothermal sphere , the projected mass inside of 250 kpc ( where it is convenient to measure the light ) is @xmath49 , assuming @xmath41 . for the range of @xmath19 allowed by the @xmath50 curve , @xmath51 ranges from @xmath52 . the velocity dispersion implies @xmath53 under the sis assumption . the discrepancy may be due to the sis assumption : unlike the redshift estimate , the mass estimate is sensitive to the profile assumed . converting from observed @xmath54 band to rest - frame @xmath17 using the approach of fischer & tyson ( 1997 ) , we find @xmath55 in this region . compared to other clusters ( mellier 1999 ) , this is high but not exceptional . baryon - unbiased samples of mass concentrations over a wide range of redshift will be of critical importance in constraining cosmological parameters . to realize the potential of this technique , weak lensing observations must have the sensitivity to discover clusters over a broad part of the cluster mass function . for example , constraints on @xmath2 from the mass function at high redshift currently involve only a few very massive clusters , and such extreme clusters lie far out on the tail of the mass distribution , which may not be gaussian . weak lensing surveys can probe clusters an order of magnitude less massive . unlike x - ray and optical selection , a shear signal does not diminish as the square of the luminosity distance , so that low - mass clusters should be detectable even at high redshift as long as the photometric redshifts are accurate in eliminating foreground sources . we have demonstrated the serendipitous discovery , with high signal - to - noise , of a rather modest cluster via a weak lensing analysis of a single 40 field . the cluster was spectroscopically confirmed at @xmath56 , with a velocity dispersion of 615 km sec@xmath0 . the tangential shear follows source ( photometric ) redshift in a manner which requires the lens to lie at or near the cluster redshift . thus , all the ingredients are in place for a truly shear - selected sample of clusters , in which any putative mass clump can be confirmed , and its redshift estimated , from the multi - color imaging data alone . this technique is also capable of answering the question of the existence of any truly dark clumps . the redshift ( and therefore mass , and mass - to - light ratio lower limit ) of any such clumps can only be derived from the shear versus source redshift curve . this further suggests that the promise of three - dimensional mass tomography ( tyson 1995 , 2000 ) over cosmologically significant volumes can be realized in wide - field deep imaging surveys . note that such a cluster is not unexpected in the volume probed by this data ( rahman & shandarin 2001 ) . ongoing cosmic shear surveys covering tens of square degrees ( _ e.g. _ the deep lens survey , descart ) should discover significant samples of shear - selected clusters ( kruse & schneider 1999 ) and begin to constrain @xmath2 and dark energy through the redshift evolution of the cluster mass function . we thank p. guhathakurta for providing spectroscopic redshifts outside the cluster , and the ctio staff for their help with the btc project and for their upgrading and maintenance of the delivered image quality of the blanco telescope . cerro tololo inter - american observatory is a division of national optical astronomy observatory ( noao ) , which is operated by the association of universities for research in astronomy , inc . , under cooperative agreement with the national science foundation . btc construction was partially funded by the nsf .	we report the discovery of a cluster of galaxies via its weak gravitational lensing effect on background galaxies , the first spectroscopically confirmed cluster to be discovered through its gravitational effects rather than by its electromagnetic radiation . this fundamentally different selection mechanism promises to yield mass - selected , rather than baryon or photon - selected , samples of these important cosmological probes . we have confirmed this cluster with spectroscopic redshifts of fifteen members at z=0.276 , with a velocity dispersion of 615 km s@xmath0 . we use the tangential shear as a function of source photometric redshift to estimate the lens redshift independently and find @xmath1 . the good agreement with the spectroscopy indicates that the redshift evolution of the mass function may be measurable from the imaging data alone in shear - selected surveys .
the perturbative qcd description of many observables measured at colliders is plagued by large corrections arising from soft and collinear parton emission , even for fairly generic kinematical conditions . for example , near threshold , large logarithmic corrections remain @xcite after cancellation of singular virtual and real gluon contributions , their large size being a result of the nearby threshold restricting the real gluons to be soft . in terms of a ( mellin ) variable @xmath5 , in terms of which threshold is approached by @xmath6 , such large threshold corrections take the form ( @xmath7 ) , @xmath8 where the @xmath9 depend in general on the process . another example @xcite is when an identified part @xmath10 of a final state has acquired small transverse momentum by soft recoil ( @xmath11 ) against the remaining , unmeasured part of the final state . then the perturbative expression for the differential cross section with respect to @xmath12 of @xmath10 takes again the form of eq . , but with different coefficients @xmath9 and with @xmath13 , @xmath14 being the impact parameter fourier conjugate to @xmath11 . such large logarithmic corrections can be brought under control by all - order resummation , and there exists a large literature demonstrating the viability , where applicable , of threshold , recoil as well as their joint resummation , for a wide variety of observables . it is interesting to try to extend all - order control to classes of large terms beyond the logarithmic corrections . one such new set consists of numerically large constants ( `` @xmath15 terms '' ) originating from the same infrared - sensitive regions of those feynman diagrams that also produce the logarithms @xcite . another important class of potentially large terms , of soft - collinear origin , can be represented as @xmath16 their phenomenological importance was first demonstrated in ref . @xcite in which the leading terms @xmath17 were also summed to all orders for higgs production and drell - yan . the assessment of these terms was made more meaningful in the context of a complete next - to - next - leading order ( nnlo ) @xcite calculation , and a consistent next - to - next - leading logarithmic ( nnll ) threshold - resummed result @xcite . it is not yet clear how to sum next - to - leading terms in . in this paper we examine the impact of the leading terms in for a single particle inclusive observable , the @xmath12 spectrum of prompt photons produced in hadronic collisions . we do this in the context of both a threshold @xcite and joint @xcite resummed calculation for this spectrum . the paper is organized as follows . in section 2 we review briefly the threshold and joint resummed prompt photon @xmath12 distribution . in section 3 we describe and motivate our extension to include the leading @xmath18 terms . in section 4 we assess the numerical impact of these corrections , and we conclude in section 5 . we consider the inclusive @xmath12 distribution of prompt photons produced in hadron - hadron collisions at center of mass ( cm ) energy @xmath19 @xmath20 where @xmath21 refer to the two incoming hadrons and @xmath22 to the unobserved part of the final state . the lowest order qcd processes producing the prompt photon at partonic cm energy @xmath23 are @xmath24 the distance to threshold is customarily measured by the variable @xmath25 , where @xmath26 . at the parton level this distance is given by @xmath27 . below we review the result for the joint resummed prompt photon @xmath12 distribution . at the end of this section we recall how the threshold resummed result may be derived from it . the joint resummation formalism for prompt photon production @xcite implements the notion that , in the presence of soft qcd radiation with summed transverse momentum @xmath28 of soft recoiling partons , the actual transverse momentum produced by the hard collision is not @xmath29 but rather @xmath30 . stated more precisely : in the context of a refactorization analysis @xcite one can identify a short - distance process at cm energy @xmath31 that produces a prompt photon with momentum @xmath32 in a recoiling frame . one defines accordingly @xmath33 . the extreme situation @xmath34 in which all transverse momentum is produced through soft recoil leads to a singularity in the short - distance process , which we avoid by imposing an upper limit @xmath35 on @xmath11 @xcite . a recently proposed extension @xcite of joint resummation avoids this singularity . the joint resummed @xmath12 distribution of prompt photons in hadronic collisions is written as @xmath36\ , .\end{aligned}\ ] ] let us explain each of the terms on the right hand side of eq . . the top line displays the moments of standard parton distribution functions , as well as the sum over initial state parton flavors . the next line contains the mellin transform over the partonic scaling variable @xmath37 in the recoiling frame , the born amplitudes , and the @xmath5- and @xmath14-independent hard virtual corrections summarized in @xmath38 . the second to last line contains the integral over the recoil momentum of the soft partons , as well as a kinematic factor linking recoil and threshold effects . the last line contains the sudakov exponentials from initial and final state partons , as well as soft wide - angle radiation in combined mellin - impact parameter space . as indicated in the last line of eq . , large threshold and recoil logarithms , expressed through @xmath39 and @xmath40 , can be resummed into an exponential form . the perturbative exponential moment dependence at next - to - leading logarithmic ( nll ) accuracy is given by @xmath41 let us discuss each of these terms in turn . the initial state perturbative exponent reads , in integral form @xmath42 where @xmath43 are the renormalization and factorization scale , respectively . the function @xmath44 defines the @xmath5- and @xmath14-dependent scale of soft gluons to be included in the resummation , and is chosen as @xcite @xmath45 where @xmath46 is a suitably chosen constant and @xmath47 with @xmath48 the euler constant . an older form used in @xcite @xmath49 generates spurious subleading logarithms in @xmath11 @xcite . we postpone elaborating on the integral in eq . to the next section . the final state jet exponent reads to nll accuracy @xmath50 where @xmath51 the exponent associated with wide angle soft radiation is @xmath52 the functions @xmath53 and @xmath54 as well as the functions @xmath38 @xcite are listed in the appendix . a nonperturbative term must be added to the perturbative exponent in eq . in order to regularize the limit in which @xmath11 is very small . . @xcite we take @xmath55 the threshold - resummed result can now be derived by simply neglecting @xmath56 in the factor @xmath57)^{n+1}$ ] in eq . . then the @xmath56 integral sets @xmath58 to zero everywhere , yielding the threshold - resummed result . the leading terms in eq . originate from both initial and final state radiation , and to resum them we will use two different methods upon which we elaborate in this section . there are moreover two classes of functions in momentum space at order @xmath60 that generate the leading @xmath61 terms upon mellin transformation . in terms of the variable @xmath62 which in the present case can represent either @xmath63 or @xmath37 , one of the two classes is formed by the singular plus distributions @xmath64_+$ ] , the other by the singular but integrable @xmath65 . the @xmath59 contributions from the former can be computed using the methods of @xcite and can be found e.g. in ref . the @xmath59 contributions from the latter can be generated at any order in perturbation theory by a simple replacement in the resummed expression ( see below in eqs . ) , expanding the resulting expression to the desired order , and keeping the leading term in eq . . roughly speaking , the replacement is equivalent to exchanging at order @xmath66 one soft - collinear gluon ( corresponding to one factor @xmath67 ) for a hard - collinear one ( corresponding to a factor @xmath68 ) @xmath69 this replacement is in fact easily included in the existing threshold resummation formulae . a preliminary study for prompt photon production was carried out in ref . we employ this replacement method in fact for the final state related @xmath70 terms . it was pointed out in refs . @xcite that the initial state related @xmath70 terms could be generated in the context of joint resummation by extending evolution of parton densities to a soft scale . we will use this method as well , for the first time for a one - particle inclusive observable . we now discuss the initial and final state @xmath59 contributions in turn . our procedure for the initial state follows refs . @xcite , where the joint resummation for electroweak or higgs boson production at mass @xmath31 and transverse momentum @xmath11 was given . we recall the key points here . the integral form of the initial state nll exponent can be written as @xmath71 the first term in this expression leads to @xmath72 where @xmath73 we recall that the @xmath74 depends on @xmath5 and @xmath14 through eq . . the functions @xmath75 are listed in the appendix . the second term represents flavor - conserving evolution to nll accuracy ( the integrand consists of the @xmath39 and constant terms for the anomalous dimension matrix @xmath76 for @xmath77 ) from the hard scale @xmath78 to the soft scale @xmath79 . one now performs the replacement @xcite @xmath80 that includes the leading , flavor - diagonal @xmath59 effects generated by the @xmath81 integral ( the @xmath82 part of @xmath83 combines with the @xmath39 terms ) . in fact , one may go further and include the off - diagonal contributions via the replacement @xmath84 \ , f_{g / h}(n , \mu_f)\ ; \longrightarrow \ ; { \cal e}_{ik } \left(n , q/\chi,\mu_f\right ) \ , f_{k / h}(n , \mu_f)\ , .\ ] ] where @xmath85 plus nll corrections . as a result , we can replace in eq . the combination @xmath86\ ] ] by @xmath87\ ] ] where @xmath88 the matrix @xmath89 implements evolution from scale @xmath78 to scale @xmath79 , and is normalized to be the unit matrix if these two scales are equal . note that the dependence on @xmath78 cancels among the factors in eq . . leading @xmath59 effects arising from final state radiation can be derived from the jet functions @xcite that enter threshold or joint resummed expressions for observables having final state partons at lowest order . the integral form of the final state exponent @xmath90 in eq . reads @xmath91 to include leading @xmath59 dependence in @xmath92 we make the replacement @xcite @xmath93\ , a^{(1)}_q + { \cal o}(\frac{1}{n^2})\,,\ ] ] where @xmath94 . the extra terms can be cast in a more convenient form . using @xmath95 and the replacement ( accurate to nll ) @xmath96 one finds @xmath97 where the extra terms @xmath98 that include the leading @xmath59 terms due to final state radiation read @xmath99\,,\ ] ] @xmath100\,.\ ] ] there is no leading @xmath59 contribution arising from wide angle soft radiation . as a result , we finally arrive at the following equation for the joint resummed prompt photon hadroproduction @xmath12 spectrum in which leading soft - collinear effects are included : @xmath101\nonumber \\ & \ & \hspace{-15 mm } \times \exp\left[\frac{1}{\alpha_s(\mu ) } f^{(0)}_k(\lambda ) + f^{(1)}_k(\lambda , q,\mu ) + f'_k(\lambda,\alpha_s ) + g^{(1)}_{ijk } ( \lambda)\right]\,.\end{aligned}\ ] ] as before , the corresponding threshold result may be obtained by neglecting @xmath102 in the last factor on the second line . here we study numerically the inclusion of the @xmath59 terms for the case of prompt photon production for two kinematic conditions : those of @xmath103 collisions at the tevatron at @xmath104 tev @xcite , and those of the @xmath105 collisions in the e706 @xcite fixed target experiment with @xmath106 gev . our main aim is to assess the effect of such terms in relevant kinematic conditions , rather than provide optimized and realistic theoretical calculations for comparison with data ( see ref . @xcite for recent study ) . for instance , we do not include contributions from fragmentation processes , which have recently been addressed in ref . @xcite and shown to be significant . our assessments mainly consist of comparing the same calculation with and without @xmath59 terms . our default choices for various input parameters are as follows . we use the grv parton density set @xcite , corresponding to @xmath107 , with the evolution code of ref . @xcite , changing flavor number at @xmath108 and @xmath109 . we choose the factorization and renormalization scale equal to @xmath12 , and the non - perturbative parameter @xmath110 in eq . equal to @xmath111 . for the parameter @xmath74 we use the expression in eq . , following @xcite , with @xmath112 . does not substantially modify results , but choosing the form in eq . , which generates spurious subleading recoil logs @xcite , does lead to significant changes at larger @xmath12 . ] for our joint - resummed results , we chose for tevatron ( e706 ) kinematics the cut - off @xmath35 in eq . equal to @xmath113 gev . regarding logarithmic accuracy , and unless otherwise stated we refer to ll when using only @xmath114 in eq . , @xmath115 in eq . , and @xmath116 for the processes in ; we refer to nll when also including @xmath117 and @xmath118 and the virtual corrections in . for the evolution from scale @xmath78 to @xmath79 in eq . we use the full nlo anomalous dimension in all cases . starting with tevatron kinematics we compare in figs . [ fig1]-[fig3 ] results at ll and nll accuracy , with and without the leading @xmath119 contribution for joint resummation . contributions for tevatron kinematics . left pane : ll without @xmath59 ( @xmath120 , solid ) , nll without @xmath59 ( @xmath14 , dashed ) , nll with @xmath59 ( @xmath121 , short - dashed ) . right pane : ratio of nll to ll ( solid ) , ratio of nll with @xmath59 to nll without ( dashed).,title="fig:",scaledwidth=49.0% ] contributions for tevatron kinematics . left pane : ll without @xmath59 ( @xmath120 , solid ) , nll without @xmath59 ( @xmath14 , dashed ) , nll with @xmath59 ( @xmath121 , short - dashed ) . right pane : ratio of nll to ll ( solid ) , ratio of nll with @xmath59 to nll without ( dashed).,title="fig:",scaledwidth=47.0% ] for clarity we have here included the constant corrections in also for the ll case . [ fig1 ] shows that the effect of the leading @xmath59 is appreciable when compared to the effect of passing from ll to nll , the latter difference being almost negligible . inclusion of @xmath59 effects leads to noticeable suppression for most of the @xmath12 range , and to enhancement at very small and very large @xmath12 . to better understand the origin of these @xmath59 suppressed contributions , we examine in figs . [ fig2 ] and [ fig3 ] for each channel in the contributions from the initial and final state . effects for @xmath122 channel at ll , tevatron kinematics . ratio to ll without @xmath59 of initial state ( solid ) and final state ( dashed ) effects , and both ( short - dashed ) . , scaledwidth=50.0% ] effects for @xmath123 channel at ll , tevatron kinematics . labels as in fig . [ fig2 ] . , scaledwidth=50.0% ] we plot these contributions for the ll cross sections only to facilitate interpretation . to help understand the results , we can expand the perturbative exponent in eq . to lowest order in @xmath124 , keeping only the @xmath125 and @xmath59 terms @xmath126 the expressions suggest that the initial state @xmath59 terms enhance the cross section for the @xmath127 and in particular the @xmath123 channels , while the final state @xmath59 terms suppress it , again by an amount that depends on the channel . the net result turns out to be suppression in the former channel and enhancement in the latter . these qualitative aspects are indeed borne out if we use the same method to compute initial state @xmath59 effects as we did for the final state in section [ sec : final - state ] . is actually still enhancement , because the contribution of the @xmath128 functions in , is very small . ] in the present case however , the net @xmath59 effect in both channels is suppression , indicating that the non - diagonal terms in the evolution matrix give a sizeable negative contribution . note that for tevatron kinematics , when combining channels , the @xmath123 channel dominates at low @xmath12 , because the required momentum fractions are not too large . at large @xmath12 , where parton momentum fractions are larger , the valence - quark dominated @xmath122 channel takes over . turning to e706 kinematics we perform the same studies as we did for the tevatron . the results are shown in figs . [ fig4]-[fig6 ] . contributions for e706 kinematics . labels as in fig . [ fig1].,title="fig:",scaledwidth=49.0% ] contributions for e706 kinematics . labels as in fig . [ fig1].,title="fig:",scaledwidth=47.0% ] we observe an overall enhancement due to the @xmath59 effects , somewhat smaller than the change from ll to nll . both effects are more pronounced than for the tevatron . this is due both to a larger value of @xmath124 as well as being closer to threshold in this fixed target kinematical regime . examining the effects per channel in figs . [ fig5 ] and [ fig6 ] , we now see a noticeable enhancement from the initial state @xmath59 effects in the @xmath127 channel , but still suppression in the @xmath123 channel . clearly the non - diagonal terms in the evolution matrix play a significant role for the e706 case as well . effects for @xmath127 channel at ll , e706 kinematics . labels as in fig . , scaledwidth=50.0% ] effects for @xmath123 channel at ll , e706 kinematics . labels as in fig . , scaledwidth=50.0% ] next , we examine the differences between threshold and joint resummation . in fig . [ fig8 ] we compare resummed results directly by showing the ratios with respect to the joint - resummed @xmath12 distribution without @xmath59 terms . we see for tevatron kinematics that the threshold resummed dominates the joint resummed at large @xmath12 , while at low @xmath12 the converse is true . for the e706 case the threshold resummed results are entirely below the joint - resummed ones . for tevatron ( left pane ) and e706 ( right pane).,title="fig:",scaledwidth=46.0% ] for tevatron ( left pane ) and e706 ( right pane).,title="fig:",scaledwidth=46.0% ] the threshold resummed curves are shown separately in fig . [ fig7 ] , which is analogous to the rightmost panels in figs . [ fig1 ] and [ fig4 ] . effects in threshold resummation , for tevatron ( left pane ) and e706 ( right pane ) . labels as in fig . [ fig1 ] right pane.,title="fig:",scaledwidth=46.0% ] effects in threshold resummation , for tevatron ( left pane ) and e706 ( right pane ) . labels as in fig . [ fig1 ] right pane.,title="fig:",scaledwidth=46.0% ] for tevatron kinematics the inclusion of @xmath59 terms in threshold resummation leads , as for joint resummation , from suppression at small @xmath12 to enhancement at larger @xmath12 , but more noticeably . for e706 kinematics , different from the joint resummation case , the enhancement at small @xmath12 turns to suppression just below @xmath129 gev . the cross section even becomes negative beyond @xmath130 gev , which is due to the fact that the nearness of the threshold drives the scale @xmath79 in eq . effectively below the starting scale of the pdf evolution we have examined the effects of including terms of the form @xmath131 in joint - resummed and threshold - resummed prompt photon @xmath12 distributions at both collider and fixed target kinematics , at leading accuracy ( @xmath77 ) . the complete structure of subleading terms of the form is still unknown . note that we have not considered the fragmentation component of the prompt photon production cross section in our analysis . to the extent that terms of the form arise from initial state radiation effects , we used the method of refs . @xcite to include them , now in a single - particle inclusive cross section . those arising from final state emission we included by extending the jet function to leading @xmath59 accuracy . numerically we found the combined @xmath59 terms to be comparable to nll corrections , and dependent on kinematics either enhancing or suppressing . the final state @xmath59 contributions were particularly small , while in the initial state the effects of non - leading @xmath82 effects are appreciable , depending again on channel and kinematics . the flavour non - diagonal terms in the evolution matrix were found to be numerically significant , and the main source of discrepancy with expectations based on simple approximations . we conclude that , because the effects , though small , are non - negligible , understanding the structure of @xmath59 terms better is a worthwhile pursuit . this work was supported by the foundation for fundamental research of matter ( fom ) and the national organization for scientific research ( nwo ) . rb and am would like to thank nikhef and el and am the imsc in chennai for local hospitality . here we list the exponents used in section [ sec : including - leading - ln ] . the initial state exponents involve @xmath132\ , , \\ h_a^{(1 ) } ( \beta , q,\mu ) & = & \frac{a_a^{(1 ) } b_1}{2\pi b_0 ^ 3 } \left [ \frac{1}{2 } \ln^2 ( 1 - 2 \beta ) + \frac{2 \beta + \ln(1 - 2 \beta)}{1 - 2\beta } \right ] + \frac{b_a^{(1)}}{\pi b_0 } \ln(1 - 2 \beta ) \nonumber \\ & + & \frac{1}{2\pi b_0 } \left [ a_a^{(1)}\ln \left ( \frac{q^2}{\mu^2 } \right ) -\frac{a_a^{(2)}}{\pi b_0}\right ] \ ; \left [ \frac{2 \beta}{1 - 2\beta}+ \ln(1 - 2 \beta ) \right ] \ ; .\end{aligned}\ ] ] here @xmath133\ ] ] where @xmath134 and @xmath135 . also @xmath136 the final state exponents involve the functions @xmath137\ ] ] @xmath138\\ + \frac{b^{(1)}_a } { \pi b_0 } \ln(1-\lambda ) -\frac{a^{(1)}_a\gamma_e}{\pi b_0 } [ \ln(1-\lambda)-\ln(1 - 2\lambda ) ] -\frac{a^{(2)}_a}{2\pi^2 b_0 ^ 2}[2 \ln(1-\lambda ) - \ln(1 - 2\lambda)]\\ + \frac{a^{(1)}_a}{2\pi b_0}[2\ln(1- \lambda)- \ln(1 - 2\lambda)]\ln\frac{q^2}{\mu^2}\end{gathered}\ ] ] the wide - angle soft radiation exponents are @xmath139 in these equations @xmath140 where @xmath141 . these expressions are obtained by expanding the perturbative functions @xmath142 , @xmath143 and @xmath144 in powers of @xmath124 , @xmath145 and so on . finally , the explicit forms of @xmath38 @xcite are @xmath146 \\ & c^{qg\rightarrow \gamma q } = 1 + \frac{\alpha_s}{\pi}\bigg [ -\frac{1}{10}(c_f - 2c_a)\ln 2 - \frac{1}{2}k_q + \frac{\zeta(2)}{10 } ( 2c_f + 19c_a ) \\ & \hspace{0.5 cm } + \frac{1}{2}c_f\ln^2 2 + \frac{3}{4}(c_f+\pi b_0 ) ( -\ln 2 ) - \pi b_0\ln\frac{2p_t^2}{\mu^2 } \bigg ] \end{aligned}\ ] ] where @xmath147 we note that there is no factorization scale dependence in @xmath148 and the coefficient functions in eq . because of complete evolution from scale @xmath78 to @xmath79 in eqs . , and . g. sterman , nucl . b281 * , 310 ( 1987 ) . s. catani and l. trentadue , nucl . phys . * b327 * , 323 ( 1989 ) . y. l. dokshitzer , d. diakonov and s. i. troian , phys . b79 * , 269 ( 1978 ) . g. parisi and r. petronzio , nucl . b154 * , 427 ( 1979 ) . g. altarelli , r. k. ellis , m. greco and g. martinelli , nucl . phys . * b246 * , 12 ( 1984 ) . j. c. collins and d. e. soper , nucl . b193 * , 381 ( 1981 ) . j. c. collins and d. e. soper , nucl . b197 * , 446 ( 1982 ) . j. c. collins , d. e. soper and g. sterman , nucl . b250 * , 199 ( 1985 ) . g. parisi , phys . b90 * , 295 ( 1980 ) . l. magnea and g. sterman , phys . rev . * d42 * , 4222 ( 1990 ) . t. o. eynck , e. laenen and l. magnea , jhep * 06 * , 057 ( 2003 ) , [ hep - ph/0305179 ] . m. kramer , e. laenen and m. spira , nucl . * b511 * , 523 ( 1998 ) , [ hep - ph/9611272 ] . r. v. harlander and w. b. kilgore , phys . lett . * 88 * , 201801 ( 2002 ) , [ hep - ph/0201206 ] . r. v. harlander and w. b. kilgore , phys . rev . * d64 * , 013015 ( 2001 ) , [ hep - ph/0102241 ] . c. anastasiou and k. melnikov , nucl . * b646 * , 220 ( 2002 ) , [ hep - ph/0207004 ] . c. anastasiou , k. melnikov and f. petriello , phys . . lett . * 93 * , 262002 ( 2004 ) , [ hep - ph/0409088 ] . v. ravindran , j. smith and w. l. van neerven , nucl . phys . * b665 * , 325 ( 2003 ) , [ hep - ph/0302135 ] . v. ravindran , j. smith and w. l. van neerven , mod . a18 * , 1721 ( 2003 ) , [ hep - ph/0307005 ] . s. catani , d. de florian and m. grazzini , jhep * 0105 * , 025 ( 2001 ) , [ hep - ph/0102227 ] . s. catani , d. de florian , m. grazzini and p. nason , jhep * 0307 * , 028 ( 2003 ) , [ hep - ph/0306211 ] . e. laenen , g. oderda and g. sterman , phys . lett . * b438 * , 173 ( 1998 ) , [ hep - ph/9806467 ] . s. catani , m. l. mangano and p. nason , jhep * 9807 * , 024 ( 1998 ) , [ hep - ph/9806484 ] . s. catani , m. l. mangano , p. nason , c. oleari and w. vogelsang , jhep * 9903 * , 025 ( 1999 ) , [ hep - ph/9903436 ] . n. kidonakis and j. f. owens , phys . d61 * , 094004 ( 2000 ) , [ hep - ph/9912388 ] . p. bolzoni , s. forte and g. ridolfi , nucl . b * 731 * , 85 ( 2005 ) , [ hep - ph/0504115 ] . e. laenen , g. sterman and w. vogelsang , phys . lett . * 84 * , 4296 ( 2000 ) , [ hep - ph/0002078 ] . e. laenen , g. sterman and w. vogelsang , phys . rev . * d63 * , 114018 ( 2001 ) , [ hep - ph/0010080 ] . g. sterman and w. vogelsang , phys . * d71 * , 014013 ( 2005 ) , [ hep - ph/0409234 ] . li , phys . lett . * b454 * , 328 ( 1999 ) , [ hep - ph/9812363 ] . a. kulesza , g. sterman and w. vogelsang , phys . rev . * 014011 ( 2002 ) , [ hep - ph/0202251 ] . s. catani , m. l. mangano , p. nason and l. trentadue , nucl . phys . * b478 * , 273 ( 1996 ) , [ hep - ph/9604351 ] . p. mathews _ et al . _ , pramana * 63 * , 1367 ( 2004 ) , proceedings of the eighth workshop on high energy physics phenomenology , mumbai , india , jan . 2004 . a. kulesza , g. sterman and w. vogelsang , phys . rev . * d69 * , 014012 ( 2004 ) , [ hep - ph/0309264 ] . n. kidonakis , g. oderda and g. sterman , nucl . b525 * , 299 ( 1998 ) , [ hep - ph/9801268 ] . d0 , v. m. abazov _ et al . _ , phys . lett . * b639 * , 151 ( 2006 ) , [ hep - ex/0511054 ] . cdf , d. acosta _ et al . _ , phys . rev . * d70 * , 074008 ( 2004 ) , [ hep - ex/0404022 ] . fermilab e706 , l. apanasevich _ et al . _ , phys . rev . * d70 * , 092009 ( 2004 ) , [ hep - ex/0407011 ] . p. aurenche , m. fontannaz , j .- guillet , e. pilon and m. werlen , phys . rev . * d73 * , 094007 ( 2006 ) , [ hep - ph/0602133 ] . d. de florian and w. vogelsang , phys . rev . * d72 * , 014014 ( 2005 ) , [ hep - ph/0506150 ] . m. gluck , e. reya and a. vogt , eur . j. * c5 * , 461 ( 1998 ) , [ hep - ph/9806404 ] . a. vogt , comput . . commun . * 170 * , 65 ( 2005 ) , [ hep - ph/0408244 ] .	we extend next - to - leading logarithmic threshold and joint resummation for prompt photon production to include leading collinear effects . the impact of these effects is assessed for both fixed - target and collider kinematics . we find them in general to be small , but noticeable . nikhef/2007 - 008 + itp - uu-07/23 + imsc-2007/04/5 * soft - collinear effects in prompt photon production * + rahul basu@xmath0 , eric laenen@xmath1 , anuradha misra@xmath2 , patrick motylinski@xmath3 + @xmath0 _ the institute of mathematical sciences , cit campus , taramani , chennai 600 113 , india _ + @xmath3_nikhef theory group , kruislaan 409 , 1098 sj amsterdam , the netherlands _ + @xmath4_institute for theoretical physics , utrecht university + leuvenlaan 4 , 3584 ce utrecht , the netherlands _ + @xmath2 _ department of physics , university of mumbai , santacruz(e ) , mumbai , 400098 , india _ +
the textbook notion of a perfect crystal at @xmath1 , is that of a periodic array of unit cells , all comprising the same _ integer _ number of particles , @xmath2 , where @xmath3 is the unit cell volume and @xmath4 is the average particle density profile . for simplicity , let us consider a single component crystal similar to @xmath0he . we assume that the elementary constituents ( atoms or molecules ) can be regarded as structureless particles ; furthermore , we assume that they obey bose statistics . the supersolid phase ( sfs ) can be generally defined as one that combines crystalline properties , such as shear modulus and broken translation symmetry , with frictionless mass transport through the solid bulk . the striking , simultaneous presence of solid and superfluid properties in the same condensed matter system , will result in a number of phenomena that defy our everyday experience . one such phenomenon is schematically depicted in fig . [ fig1 ] , showing a sfs sample ( region with periodic arrays of dots ) placed inside a u - shaped vessel , in coexistence with the superfluid liquid . at @xmath1 undamped oscillations of the solid levels in the two sections of the u - shape vessel will be observed , as long as the maximum flow velocity remains below the critical value and the liquid - solid interface is rough . [ angle=-90 , scale=0.5]fig1 another landmark of sfs behavior ( see , for instance , ref . @xcite ) , is the reduction of the moment of inertia of a solid sample with respect to its classical value ( commonly referred to as non - classical rotational inertia , or ncri ) @xmath5 here @xmath6 is the superfluid density , and @xmath7 is the particle density in the system . the ncri effect can be observed by enclosing a known amount of solid in a vessel , which is then set in rotation about its axis . the moment of inertia of the system is related to the ( measurable ) resonant period of oscillation of the rotating system . a drop of the moment of inertia at low temperature , is interpreted as _ decoupling _ of part of the solid ( the superfluid fraction ) from the rotation . the history of ideas on how superfluidity can occur in a crystalline solid is quite old . on very general grounds , one expects that the `` textbook '' crystal ought to be insulating , i.e. , with vanishing superfluid density @xmath8 , and off - diagonal correlations decaying exponentially in space . penrose and onsager ( po ) were the first to argue in favor of this point of view @xcite , starting from the picture of atoms localized around their equilibrium lattice positions . however , their argument has never been regarded as the final word in the discussion of whether the superfluid solid phase of matter is possible . n. yang made a comment that long - range phase correlations in the solid state could occur in systems characterized by a high degree of atomic delocalization @xcite . helium is unique in this regard ; quantum properties of light helium atoms in combination with the relatively weak interatomic potential ( the well depth is only @xmath9 k ) prevent them from making a crystal at zero pressure , see fig . an objection to the po treatment was that regarding atoms as being localized around lattice points underestimates the role of exchange processes , which may allow for the ( perhaps remote ) possibility of superflow , even with an integer @xmath10 @xcite . another proposal for the supersolid phase has its origin in the observation that there is no fundamental reason why crystals should necessarily be _ commensurate _ , i.e. , feature an integer number of particles per unit cell , on average . a dilute gas of vacancies , or interstitials , which occur in all solids at finite temperature and ought to be highly mobile in a helium crystal , could be also present in the ground state giving rise to bose - einstein condensation and superfluidity at low temperature @xcite . though certainly plausible , most theoretical scenarios leading to the sfs phase are phenomenological and qualitative when it comes to making predictions for a particular hamiltonian ; they have never been supported by robust microscopic calculations , including all quantum - mechanical effects and based on realistic models of crystals . still , their appeal is so strong that they inspired decades of active research and continue to do so . remarkable manifestations of quantum behavior on a macroscopic scale in supersolids are expected to render the observation of the sfs state rather unambiguous , as long as it occurs under conditions ( e.g. , temperature ) accessible to the experiment . such an observation has eluded experiments for some thirty years ( for a review of the experimental evidence prior to 2004 , see , for instance , ref . @xcite ) ; it was only in 2004 that the first convincing evidence of a sfs phase of 4 was published . the phenomenon was first reported for a solid embedded in a porous vycor matrix by @xcite kim and chan ( kc ) ; shortly thereafter , it was extended to the system embedded in a matrix of porous gold ( with a characteristic pore size two orders of magnitude greater than in vycor ) , and eventually to the bulk crystal @xcite . in all cases , the observed ncri at low temperature was identified as the onset of supersolidity in 4 . it should be emphasized that the supersolid interpretation of the data is supported by the velocity / amplitude dependence of the ncri signal and the crucial test that the ncri effect goes away ( is suppressed by two orders of magnitude ) when the flow channel around the rotation axis is blocked . these results constitute a major problem for the non - superfluid kinetic interpretation . the kim and chan discovery sparked a renewed effort in the investigation of the sfs phase of matter . at the time of this writing , it seems fair to state that things have turned out to be considerably more complicated ( and thus more interesting ) than expected . as the initial experiment by kim and chan is being repeated by other groups , and concurrently different experiments and reliable microscopic ( first - principles ) calculations , based on state - of - the - art computational techniques are being carried out , consensus is building that the early microscopic scenarios of supersolidity , described above , are not viable to explain the body of experimental data . for example , there is now strong theoretical and experimental evidence that 4 _ is _ a commensurate crystal , and such crystals made of a single species of particles obeying bose statistics ( e.g. , 4 atoms ) are always insulating ( more precisely , the commensurate sfs phase has zero probability of being observed ) . measurements and calculations of vacancy and interstitial properties indicate that the andreev - lifshits - chester scenario of vacancy - induced supersolidity does not occur in solid helium , even out of equilibrium . therefore , attention is now shifting to a variety of inhomogeneous scenarios , that may account for many puzzling ( and often apparently conflicting ) experimental results . this paper is organized as follows : we shall first review the theoretical framework for supersolidity , including the early suggestions as well as some more recent proposals of microscopic mechanisms . they will be discussed in the light of both early and recent analytical and numerical results . we shall then examine the core of experimental data , which at this time includes observations of ncri as well as flow experiments @xcite and the most recent experiment on grain boundary superfluidity @xcite , and propose that , while a definitive theoretical explanation of the observed phenomena is not yet available , there are good reasons to look at disordered and/or inhomogeneous scenarios . specifically , we shall illustrate how the effect of disorder can hold the key to interpreting an important part of the phenomenology . finally , we shall outline some possible directions of experimental and theoretical investigation . this perspective is a quick response to the fast developing topic , not a comprehensive review . i apologize if some important contributions to the field were overlooked . given a broad range of opinions and opposing points of view on the subject , there is certainly a degree of personal bias towards which arguments and results are more convincing and should be given more `` weight '' in the discussion . the hope is that the story is provocative enough to stimulate additional research efforts to understand the remarkable phenomenon of supersolidity . often the discussion starts with the proposal for the supersolid ground state wave function @xmath11 though it is certainly allowed by the theory that in a particular material sfs exists only as a finite - temperature phase . in the penrose - onsager picture of atoms localized around equilibrium points @xmath12 , the ground state wave function is simply a product of localized wannier orbitals ^(po)_g= _ p _ j=1^n ( r_j - r_pj ) [ po ] where the sum @xmath13 is over all permutations @xmath14 of particle labels . using the definition of the single - particle density matrix at zero - temperature n(r , r)= dr_2 dr_n _ g(r , r_2 , , r_n ) _ g(r,r_2 , , r_n ) , [ density_matrix ] and simple properties of the variational state @xmath15 one readily obtains ( by substituting @xmath15 for @xmath11 above ) n(r , r)= _ i ( r_i - r ) ( r_i - r ) . [ podm ] since in the @xmath16 limit either @xmath17 or @xmath18 has to be far from the equilibrium lattice point @xmath19 , the density matrix decays to zero exponentially fast for localized wannier orbitals . this immediately implies no off - diagonal long - range order ( odlro ) and thus no superfluidity in the system . here , we adopt the following definition of odlro : we say that it is present in the system , if the integral @xmath20 diverges in the thermodynamic limit . this definition is _ different _ and more general than that based on finite condensate density ( @xmath21 ) , which requires that the integral be proportional to the system volume , in the thermodynamic limit @xcite . it makes perfect sense to define odlro in such a way that superfluidity automatically leads to odlro and vice versa . in particular , eq . ( [ odlro ] ) allows one to talk about odlro in two dimensional ( 2d ) systems at finite temperature ( often referred to as topological order @xcite ) for which the condensate density is zero . in this article , we restrict ourselves to the single - particle superfluidity . in general , one has to look at the many - body density matrix to see if there is odlro in some finite - order channel . for example , it is certainly possible that m atoms form m - molecules and superfluidity happens than at the molecular level in this case one has to study the m - particle density matrix . or , molecular hydrogen would be a good example here , by neglecting the internal structure of molecules and treating them as point particles ( probably at the expense of introducing complex effective many - body interactions ) odlro can again be discussed in the single - particle molecular " density matrix . there is no fundamental difference between the two pictures . the superfluid density ( formally a tensor ) is a linear response coefficient which controls the free - energy increase in a system with periodic boundary conditions ( bc ) when these bc are twisted , i.e. if @xmath22 is the linear system size in direction @xmath23 we require that @xmath24 ( there are other definitions of @xmath8 , e.g. through the ncri , eq . ( [ ncri ] ) , which can be shown to be identical to the one given below ) f()-f(0)= _ _ _ . [ nsdef ] here @xmath25 is the system volume . twisted bc can be avoided by introducing instead a gauge phase with the gradient @xmath26 and writing the free - energy density ( on the largest scales ) as @xmath27 using connectivity " arguments @xcite one can show that the system is sensitive to twisted bc only if it has odlro defined in eq . ( [ odlro ] ) . the wave function ( [ po ] ) takes into account zero - point motion and local exchange processes between identical particles ( these effects are included in the shape of the wannier orbitals ) ; nevertheless , it describes an insulating state , as long as @xmath28 are localized . by moving away from two major assumptions made in the construction of the insulating ground state ( [ po ] ) , one arrives at two , apparently quite different possibilities for the supersolid wave function . if the role of exchange is underestimated in localizing particles around different lattice points , then a better variational state might be of the bec form ^(bec)_g=_i=1^n ( _ j=1^n ( r_j - r_i ) ) , [ bec ] with macroscopic condensate in the single - particle state @xmath29 . the bec state ( as well as @xmath15 ) can be further improved by multiplying it by a many - body correlation factor of the jastrow form , namely @xmath30 , whose purpose is to account for the strong repulsive core of the interatomic potential , and suppress the condensate fraction ( @xmath21 ) to a value much smaller than unity ( @xmath21 is already below 10% in the superfluid 4 at zero pressure @xcite ) . by construction , the jastrow factor ought not alter , nor eliminate the underlying lattice structure . equation ( [ bec ] ) offers an alternative view of the supersolid state , namely that of a superfluid phase with a density wave modulation . however , it suffers from a fundamental shortcoming , in that the number of lattice points @xmath31 is arbitrarily assumed to be the same as the number of particles . in a superfluid the density - wave parameters of any variational _ ansatz _ are independent thermodynamic variables , to be selected through energy minimization . at this point , we recall that superfluidity and odlro imply that part of the system matter is characterized by a classical field component , for which the notion of particle number can be safely ignored without loss of generality . [ this component represents low - momentum states with large occupation numbers . ] thus , the condition @xmath32 can be satisfied only by accident , which is to say that a commensurate supersolid in continuous space has a probability of zero measure to be found in nature . as a result , ( [ bec ] ) can not describe the generic ground state of a realistic system @xcite . in order to illustrate more vividly how odlro relates to the presence of gapless vacancies and/or interstitials and incommensurability of a single component solid in continuous space , consider the simple example of a large - amplitude ( classical ) standing wave of the electro - magnetic field in the typical optical table experiment . in this case , the question would never be asked whether the number of photons per wave period is an integer . another possible modification of eq . ( [ po ] ) , one that retains the picture of localized orbitals , consists of allowing the number of particles @xmath33 to be less than the number of lattice sites @xmath31 , i.e. , assuming that the ground state has no energy gap for the creation of vacancies ^(al)_g=_\ { k_1 k_n } _ j=1^n ( r_k_j - r_j ) , [ al ] where the the sum is over all possible sets of @xmath33 lattice points out of @xmath31 available . now , the outcome of the calculation for the density matrix is quite different because in the sum over @xmath34 n(r , r)= _ i , j=1^n_l ( r_i - r ) ( r_j - r ) . [ aldm ] one can always find terms with @xmath19 and @xmath35 close , respectively , to @xmath17 and @xmath18 , no matter how large their separation . if , in the spirit of a tight - binding approach , we move from a description in terms of continuous space variable @xmath17 and @xmath18 , to that of lattice site positions for the particle , by introducing site creation ( and , analogously , annihilation ) operators @xmath36 , then eq . ( [ aldm ] ) can be recast in the following form ( taking lattice translation invariance into account ) n(l)=_k n(k , k+l ) = _ k b_k^ b_k+l^ = n_v . [ aldmv ] the final result is consistent with the picture of bec of a non - interacting vacancy gas , with dimensionless lattice concentration @xmath37 , introduced by andreev and lifshitz @xcite . one has to assume next that the effective interaction between gapless vacancies is repulsive , in order to ensure system stability ( i.e. , to prevent the vacancy gas from collapsing ) . as was noted by reatto @xcite and chester @xcite , it is not at all necessary to break translational invariance explicitly in the sfs state with vacancies , by specifying the set of equilibrium particle positions @xmath12 . an analogy between the jastrow wave function ^(j)_g \{-_i < j=1^n u(r_i - r_j ) } , [ rc ] and the partition function of a classical system of interacting particles with pairwise potential @xmath38 at finite temperature , suggests that among ( [ rc ] ) there are variational states which feature spontaneous crystalline order , as classical systems are known to crystallize at sufficiently low temperature . states ( [ rc ] ) are also superfluid since they feature odlro , and incommensurate @xcite . regarding the last property , one may observe that , due to lack of vacancy - interstitial symmetry in continuous solids , the probability of having @xmath39 is of zero measure @xcite ( an obvious statement for a classical solid at finite @xmath40 ) . based on all of the above considerations , one might be led to thinking that the jastrow wave function may hold the key to the microscopic understanding of supersolidity . however , there are fundamental problems associated with eq . ( [ rc ] ) , chiefly that it can not possibly describe realistic solids of systems with short - range interactions ( such as helium ) , for which the condensate fraction is tiny ( less than 10@xmath41 ) . the reasoning supporting this criticism is offered in the next section . the most important conclusion of this section is that only zero - point vacancies or interstitials can make an ideal crystal supersolid . there is no other ideal - crystal scenario for the sfs state though microscopic mechanisms leading to the incommensurate crystalline groundstate might be very involved . however , the story does not end here , since we did not question yet what happens when quantum crystals are not perfect , i.e. contain defects such as dislocations and grain boundaries , or even loose their crystalline order completely and form an amorphous solid . it appears , that topological defects and glasses offer an alternative approach to supersolid phenomenon . in what follows we examine which of the two possibilities is most likely to operate in helium . existing theories can not answer the question at what pressures and temperatures one has to look for the sfs phase , or even whether the sfs phase exists at all in a given hamiltonian , e.g. in helium . several aspects have emerged as necessary conditions that the sfs state has to satisfy , in order to exist ; consequently , any of them can be used to probe _ hcp _ crystals of helium theoretically . [ at the time of this writing , there is no experimental evidence of structural transitions in solid helium at low temperatures @xmath42 and moderate pressures , say @xmath43 bar . ] by definition , the sfs state has a finite superfluid density @xmath8 . since superfluidity and odlro ( as defined above ) must be present simultaneously , one can also look at asymptotic properties of the single - particle density matrix , which is nothing by the zero time limit @xmath44 of the matsubara green function @xmath45 here @xmath46 stands for the thermal expectation value , @xmath47 is the time - ordering operator , @xmath48 , and @xmath49 is the bose particle creation operator in the matsubara representation . for @xmath50 ( @xmath51 ) one is computing the green function for a vacancy ( interstitial atom ) . if , over large distances and/or long times , @xmath52 and @xmath53 decay exponentially , then the state is non - superfluid , i.e. , insulating . moreover , if @xmath52 and @xmath53 are temperature - independent up to a certain length / time scale , then one can claim that ground state properties on the corresponding scales are being explored ( this is most readily seen in the path - integral framework ) . finally , by performing simulations in the grand canonical ensemble one can investigate whether the solid is commensurate or has gapless vacancies / interstitials in the ground state . all of the above criteria have been used in recent first - principles simulations of _ hcp _ helium crystals . they are based on the standard pairwise interatomic potential for helium @xcite which essentially did not change in the last decades and is known to capture all important properties of condensed and gaseous helium with the relative accuracy of the order of one percent . the study of exchange cycles@xcite reveals that they are extremely rare in the solid phase and indicative of the insulating behavior @xcite . one may recall that macroscopic exchange cycles are necessary for superfluidity , since @xmath54 where @xmath55 is the helium mass , @xmath56 is the dimensionality , @xmath57 is the linear system size ( periodic boundary conditions are assumed ) , and @xmath58 are winding numbers , counting how many times exchange cycles involving many particles wind around the system boundaries@xcite . to simplify , we ignore here the tensor structure of @xmath8 in the crystal . initial concerns about ergodicity problems of simulations probing large exchange cycles @xcite were eliminated in subsequent path - integral monte carlo ( pimc ) simulations based on the worm algorithm @xcite , which established that the superfluid fraction @xmath59 is unobservably small ( below @xmath60 ) , even in finite - size crystals comprising only 800 atoms . more detailed evidence for the insulating behavior of helium crystals is provided by computations of the single - particle density matrix and the zero - momentum matsubara green function@xcite . in fig . [ fig2 ] , we clearly see an exponential decay of the one - body density matrix @xmath52 at large distances , both in the vicinity of the melting curve ( pressure @xmath61 bar ) and at the pressure of @xmath62 bar . the observed behavior is independent of temperature for @xmath63 k , and there is good agreement between independent calculations of @xmath52 performed by different groups at the melting pressure @xcite . from the exponential decay of the zero - momentum @xmath64 ( left panel of fig . [ fig2 ] ) , one can obtain an activation energy @xmath65 k for a vacancy ( @xmath66 k for an interstitial ) by studying @xmath67 for @xmath50 ( @xmath51 ) . we note that these energies are quite large compared to the characteristic temperature of @xmath68 k of the kc experiment , all but ruling out the possibility that thermally excited vacancies may be behind the supersolid phenomenon ( it should also be noted that @xmath69 and @xmath70 increase with pressure ) @xcite . this conclusion is consistent with the absence of thermally activated vacancies at low @xmath40 , as inferred from impurity mobility experiments ( summarized in ref . @xcite ) , as well as high - precision studies of the liquid - solid phase diagram @xcite . one may note that simulations of gaps in a system comprising 800 atoms are essentially probing the thermodynamic limit , as finite - size effects are negligible ( see inset of left panel of fig . [ fig2 ] ) . the analysis of the x - ray data at low temperature in terms of activated vacancies @xcite was questioned in ref . specifically , the formula @xmath71 with reported activation energies , in combination with the calculated @xcite vacancy effective mass , can not account for the observed change of the lattice constant ( the effect would be too small ) . however , since energies of optical and acoustic phonons at the brillouin zone boundary are also @xmath72 10 k , one can not exclude a conventional explanation in terms of crystal anharmonicity . it seems , that all one can state with relative confidence is that the vacancy concentration is below 1% . in the most conservative approach , the above - mentioned theoretical results show that commensurate _ hcp _ crystals of 4 are insulating and locally stable since single vacancies have a finite energy cost . this does not exclude , strictly speaking , a remote possibility that the true ground state may have a finite , albeit small concentration of vacancies@xcite . such scenarios , however , are ruled out by strong attractive interaction between vacancies ; indeed , when vacancies are forcibly introduced in the crystal at any finite concentration ( in a computer experiment " ) , they immediately phase separate from the crystal bulk @xcite . under realistic experimental conditions , they will quickly anneal at grain boundaries and dislocations ; if the latter are absent , then vacancies will form liquid droplets ( at the melting curve ) or dislocation loops . in any case , the net result is that the commensurate insulating _ hcp _ crystal is the true ground state . in addition to the original picture of the dilute zero - point vacancy gas @xcite , over the past few years there have been several theoretical proposals for the ideal supersolid phase of helium @xcite , differing in the microscopic mechanisms of supersolidity . some of the early works suggesting commensurate sfs phases , were subsequently amended , to respect the theorem that such phases should necessarily feature gapless vacancies ( not necessarily in the dilute gas phase ) . in this respect , @xcite are still in error . an interesting idea was proposed in ref . @xcite , that _ hcp _ crystals may become sfs only under anisotropic stresses , which are likely to occur in the experiment . however , simulation results for _ hcp _ crystals at the melting curve , show that vacancy and interstitial gaps hardly change , even under very large anisotropic stress @xcite . it appears at this point that phenomenological scenarios and mean - field theories of supersolidity do not work for 4 , and one has to seek the explanation of recent experimental results outside the paradigm of the homogeneous _ hcp _ crystal . before reviewing crucial experimental facts , let us discuss the most important shortcomings ( including some unphysical properties ) of jastrow , as well as of the related shadow"@xcite variational wave functions , in the context of helium . calculations based on these wave functions predict both ( i ) a very small , but finite , condensate fraction for solid 4 near the melting curve ( below @xmath73 ) and ( ii ) large activation energies for vacancies in solid helium samples comprising several hundred atoms@xcite . in the presence of short - range interactions among atoms , the physics of vacancy formation is local in nature . all of the energy contributions quickly approach their thermodynamic limit value , as the system size is increased . indeed , the kinetic energy of localization in a volume @xmath74 is proportional to @xmath75 ( this contribution is not present in a system with periodic boundary conditions ) . the direct coupling of the removed atom with the rest of the system converges as @xmath76 for the van der waals interaction ; the same law describes convergence of the deformation energy . the fast disappearance of finite - size effects as @xmath57 is increased , is observed in essentially all numerical simulations of condensed helium ( see , e.g. inset in fig . [ fig2 ] ) , including those based on the variational approach . on the other hand , vacancies are gapless in a state described by a jastrow wave function by construction ! it means that if variational calculations were taken to larger system sizes they would have predicted negative corrections to the vacancy energy which _ increase _ , not decrease , with @xmath57 , making @xmath77 negative for @xmath78 . this unphysical behavior is only possible if irreducible long - range multi - particle interactions , which can not be reduced to the effective chemical potential shift , are present in the system something that is not observed in nature . the other problem with the jastrow variational _ ansatz _ ( and , of other such wave functions as well ) is that it does not easily describe the physics of phase separation . because , as shown above , a dilute bose gas of vacancies in helium has attractive interactions and thus is an unstable , phase separating system , any discussion of bec and superfluidity of the uniform vacancy gas under such circumstances is meaningless . in this section , we shall argue that many experimental facts suggest that the observed supersolid phenomenon in 4 is induced by defects , or disorder , in the torsion oscillator samples . this point of view is supported by several numerical simulations , as well as by the recent direct observation of the grain boundary superflow @xcite . let us assume that the observed drop in the resonant period of the torsion oscillator can be attributed to a non - zero superfluid response at a given frequency . this seems plausible , as the period drop or , equivalently , mass decoupling increases as the oscillator amplitude @xmath79 and velocity @xmath80 decrease . [ to understand why @xmath8 decreases with @xmath80 , one has to consider the non - linear response of vortex loops and pinned vortex lines to the flow . ] in normal viscous media , an opposite correlation between the mass decoupling and @xmath80 is expected . one may also attempt to understand the ncri effect in terms of kinetic relaxation or mass redistribution in the cell @xcite . however , in helium samples , these explanations face the crucial blocked - channel test @xcite only superflow can be sensitive to the sample topology on macroscopic length - scales . it is also hard to believe that mass redistribution kinetics in solid helium is reversible at low temperature , i.e. the signal would not be reproducible upon heating and cooling ( assuming that @xmath81 in the cycle to avoid possible annealing effects ) . with the superfluid response in mind , we observe that many features in the data simply do not fit into the homogeneous supersolid picture . in particular : + it is a well established fact that in a continuous normal - superfluid ( n - sf ) transition , the superfluid density dependence on temperature near the transition point at @xmath82 is given by @xmath83 , where @xmath84 is the correlation length exponent for the 3d xy - universality class @xcite . so far , no exception to this law , which predicts an _ infinite _ derivative @xmath85 at @xmath82 has been found , either theoretically or experimentally , and the same is expected for the transition to the sfs state @xcite . on the other hand , in all torsional oscillator experiments displaying mass decoupling at low @xmath40 , the rise of @xmath6 at low @xmath40 starts off with _ zero _ derivative near @xmath82 , in turn rendering a precise determination of @xmath82 rather ambiguous . the most straightforward explanation for such a gradual increase of the superfluid density would be a broad distribution of local transition temperatures @xmath86 within a strongly inhomogeneous sample . the shape of the @xmath6 curve is then controlled by the probability distribution of transition temperatures . diffraction experiments might provide information on the crystal quality in experimental samples but these were not reported yet . + the other puzzling feature , which does not conform to the established picture of the continuous n - sf transition , is the absence of the specific heat maximum . given experimental uncertainties @xcite , the possible amplitude of the specific heat anomaly is orders of magnitude too small for the observed amount of the superfluid density @xcite . moreover , it was found that the dependence of the specific heat @xmath87 on @xmath40 is linear in the supersolid regime . a linear @xmath88 may originate from one - dimensional luttinger - liquid type structures , e.g. , superfluid dislocation cores @xcite and ridges between grain boundaries @xcite , or just from a collection of two - level systems in the amorphous sample @xcite . + if the supersolid phenomenon observed in recent experiments is indeed a homogeneous bulk effect , it is difficult to explain the observed dependence of the microscopic parameter @xmath8 on the geometry of the experimental setup . the scatter of the reported values of @xmath8 ( spanning almost a decade ) , depending on annulus or open geometry in the kc experiment , can be in principle explained if the sample quality , and thus the distribution of @xmath82 , is strongly dependent on the cell geometry ( certainly a plausible argument ) . + rittner and reppy @xcite have reported the `` elimination of supersolid by annealing '' , in an experimental setup very close to that of kim and chan . the other important observation , is that the torsional oscillator quality factor , or inverse dissipation rate , is higher in normal samples . this result should be regarded as direct evidence in favor of disorder - induced supersolidity . however , such annealing effects have not been confirmed by kim and chan @xcite , and the explanation for this discrepancy is lacking , at the time of this writing . + bosonic superfluidity is a very robust phenomenon , and one does not expect any significant changes in the superfluid properties when 3 impurities are added , at a concentration @xmath89 . it is well established @xcite that 3 substitution atom in the 4 solid matrix is described by a tight - binding model , with a tiny hopping amplitude @xmath90 k. the small value of @xmath91 is consistent with the characteristic energy for nuclear magnetism in solid 3 , which is believed to be due to exchange of helium atoms@xcite . this result is important in several ways . first , it proves that 3 atoms do _ not _ induce vacancy formation in their vicinity since the 3-vacancy complex would be extremely mobile ( the vacancy hopping amplitude is nearly four orders of magnitude higher @xcite ) . second , a small value of @xmath91 makes 3 atoms extremely sensitive to virtually any type of crystalline disorder , since defects will certainly create tight bound states for 3 . even a weak deformation potential between 3 impurities is sufficient to bind them @xcite , or localize dilute 3 solutions @xcite . it is therefore very surprising , that both the superfluid fraction and the onset of the superfluid response observed in kc experiments , are _ significantly _ altered when 3 impurities are added , even if the concentration @xmath92 is as small as few @xmath93 , ( or even few @xmath94 ! ) . the mismatch between @xmath8 and @xmath92 is so large that it casts serious doubts on the homogeneous supersolid scenario @xcite , even on phenomenological grounds@xcite . in a disordered sample , the lighter 3 atom will try to minimize its kinetic energy by binding to static defects which have lower local particle number density , such as dislocation cores , grain boundaries and ridges between them ( vacancies are mobile and it is not clear whether they can form a bound state with 3 atoms ) . one thus expects an accumulation of 3 on defects at low temperature ; it is plausible that the local concentration of isotopic impurities may be fairly high in certain regions of space . the other , somewhat speculative , effect of 3 substitution might be that the quality of samples , i.e. the amount of disorder , is itself a function of @xmath92 . for example , by concentrating at the perimeter of the microcrystal , 3 atoms may inhibit fast crystal growth , thereby helping to make the `` ice cream '' . if we now couple these considerations with the theoretical conclusion that only disorder can be responsible for the supersolid phenomenon in helium , we find a very subtle interplay between the superfluid response and 3 content . imagine for a moment , that superfluidity occurs along grain boundaries , while 3 is concentrated at the boundary ridges . assuming that 3 atoms suppress superfluidity , we arrive at the picture where 3-rich ridges form a josephson junction network completely surrounding grain boundaries . in this scenario it is conceivable that even a tiny amount of 3 ( per volume ) might be relevant for supersolidity @xcite . + the obvious expectation is that supersolid properties ought to be suppressed , as pressure is applied to the sample . applying pressure indeed drives down the superfluid transition temperature @xmath82 . practically all known properties of helium crystals are consistent with the overall tendency of helium atoms to behave more classically at higher densities ( see , e.g. fig . [ fig2 ] ) . surprisingly , the kc data @xcite show hardly _ any _ pressure dependence for @xmath8 and @xmath95 , up to @xmath96 bar ; at higher pressures , the supersolid signal weakens , and apparently disappears at @xmath97 bar . at this point one may wonder whether experimental data on pressure dependence should be attributed to properties of the _ same _ sample , as if the pressure was changed by squeezing the volume , or different point represents properties of _ different _ samples . if supersolidity is directly related to the sample quality , which , in turn , depends on initial conditions and the solidification protocol , then data become very ambiguous , and their interpretation intricate . while the above mentioned experimental facts do not fit into the homogeneous supersolid crystal picture , there are others which are hard to reconcile with _ any _ existing theoretical framework . for example : + superfluid decoupling in the torsion bob has the most unusual dependence on the oscillator amplitude / velocity . though in the current setup velocity and amplitude are strictly related , it is assumed that forces equilibrate in the sample fast enough to attribute all changes in the superflow to the velocity @xmath98 . it is found that @xmath8 saturates at macroscopically small velocities , @xmath99 ( related to just few circulation quanta ) . these values clearly have nothing to do with the naive sound velocity estimate @xmath100 based on typical helium parameters . the discrepancy is so large , that there is virtually no room left for explaining @xmath101 using microscopic mechanisms . on the other hand , if extended defects in the superfluid phase order are involved , e.g. vortex lines of length @xmath57 , then one faces the problem that the vortex motion timescale @xmath102 is @xmath103 to @xmath104 orders of magnitude too long , when compared to the oscillation period . + superfluidity is associated with an anomalous mass current response to a gauge phase gradient or chemical potential difference . there have been numerous attempts in the past to detect superflow in the solid phase with negative results ( see ref . the same conclusion was reached in more recent studies @xcite , both in vycor and in bulk samples . as they stand now , flow experiments do _ not _ support , and indeed are incompatible with the supersolid phase of helium . we can only note here that in disordered solids large pressure gradients are common , and it is not obvious what chemical potential difference is applied to the superfluid component in the disordered sample . for example , air pressure in the cave deep under the mountain is still at a modest 1 bar value . aside looking at defects that are typical for the polycrystalline sample , one may wonder if helium can exist in an `` ultimate '' disordered solid phase , namely , a glass . all known structural glasses are in a normal , non - superfluid state . in this regard , 4 offers an intriguing possibility , unique for a quantum solid , of being in a superfluid glassy " phase , or superglass ( sg ) . by definition , in the metastable sg phase the translation invariance is broken , but the pair correlation function @xmath105 features no diagonal long range order . at the same time , a superglass has odlro and non - zero superfluid response . strange as it is , the sg is reminiscent of a sponge soaked in a superfluid liquid made of the same atoms . superglass was observed in numerical simulations of high - pressure ( about 150 bar ) samples prepared by fast temperature quench from the normal liquid state @xcite . clearly , the dynamics of real helium under cooling is vastly different from the thermalization dynamics in the monte carlo simulation done ( i ) in imaginary time , and ( ii ) in the absence of energy conservation . thus , numerical simulations rather answer the question of existence of the metastable phase and its quasi - equilibrium properties , but not how easily it can be prepared experimentally . one may also wonder if helium sg is merely an artifact of the simulation algorithm ( though we do not see any obvious reason to suspect that ) , or a long - lived , metastable , physical phase responsible for the unexpected outcome of the acoustic - pulse experiment @xcite which aimed at nucleating the crystal phase in the middle of the cold , @xmath106 , superfluid liquid but failed . the local pressure in the pulse reached 160 bar , well above the theoretically predicted threshold of 60 bar for quantum nucleation @xcite . it is not known what the normal - superfluid transition temperature in the glassy phase is , at what temperature the metastable liquid freezes into a glass , and what role ( if any ) the sg might play in the torsion oscillator experiments , e.g. , by being trapped in some amount between micro - crystals . even if the bulk phase of the material is an insulating solid , there is no physical reason why grain boundaries ( gb ) in the same material ought to be insulating as well , especially in the vicinity of the ( weakly first - order ) liquid - solid transition . it means that sf in lower dimensionality can be obtained by simply placing two solid pieces right next to each other . conceptually , this possibility was demonstrated in model simulations of domain walls in the checkerboard solid formed by hard core bosons with the nearest - neighbor repulsion on a simple cubic lattice @xcite . subsequent path - integral monte carlo simulations of 4 polycrystals @xcite did reveal that some ( not all ! ) grain boundaries are likely to support superfluidity . superfluidity in the layer at the disordered vycor substrate was also reported in ref . @xcite . simulations also find that nearly all ridges ( lines of contact between different gb ) show robust phase coherence properties . it is thus possible that superfluidity across grain boundaries of small ( 216-atom ) crystals is nothing but the proximity effect . this issue was addressed in simulations of much larger gb @xcite in a system of about 2000 atoms consisting of two randomly oriented crystallites at the melting density @xmath107 3 . the main conclusions did not change ; while grain boundaries at special angles and high symmetry directions are insulating , generic grain boundaries appear to be superfluid with typical transition temperatures ( orientation dependent ) of about half a kelvin and the maximum possible @xmath95 at about @xmath108 k. the width of the superfluid gb region is about @xmath109 where @xmath110 is the nearest - neighbor distance in the hcp crystal , see fig . [ fig3 ] . in a remarkable recent experiment @xcite , superflow along the grain boundary was detected in an experiment whose setup is similar to that shown in fig . the difference between the solid levels was found to decay according to the linear law @xmath111 characteristic of the superfluid flow at the critical velocity ( estimated to be of the order of meters per second ) . the flow was detected only in the presence of grain boundaries ; otherwise the sample did not relax . on one occasion , the dynamics abruptly came to a halt , as the boundary had suddenly disappeared . a grain boundary is a topological crystalline defect , and its quick ( in less than @xmath112 min time ) disappearance is by itself an amazing observation , since it has no obvious explanation other than suggesting that boundaries are in a rough superfluid phase , supporting coherent recrystallization waves . it is currently our strong belief , that all known experimental facts , together with results of first - principles numerical simulations , rule out any explanation of the observed supersolid phenomenon in 4 within a homogeneous crystal framework . the only reasonable alternative , consistent with many observations , is that of supersolidity induced by crystalline disorder . at the moment , very little is known for sure about structure properties of solid samples in the torsion oscillator experiments . hopefully , in future investigations it will be possible to control sample quality _ in situ _ , e.g. by forming solids under constant pressure , rather than under constant volume conditions . helium crystals of high quality are usually made out of the superfluid liquid at constant pressure . solids grown from normal liquid under constant volume conditions are not transparent ( similar in appearance to milk ) @xcite . it is also desirable to have direct information on the solid order in the same supersolid sample , e.g. , by optical means . even more urgently , several experimental observations of different groups have to be reconciled , especially regarding pressure dependence and annealing effects . it seems also that additional measurements of the specific heat and other thermodynamic properties can help in quantifying the amount of `` disorder '' in experimental samples @xcite . analytic theories of the sfs state in continuous space are phenomenological in nature when it comes to predicting , for a particular hamiltonian , whether this state exists and what are crucial parameters for the sfs state in terms of bare couplings . supersolids arise from competition between the kinetic and potential energy terms invalidating perturbative and mean - field treatments . at present , we do not know a single example of a system characterized by a realistic interatomic potential , which has a stable sfs phase in continuous space . this work is worth pursuing , since there are exciting new possibilities in engineering interparticle interactions in cold atomic systems . it would be also interesting to see , once the ideal sfs phase is found , which of the recent theoretical predictions capture the microscopic picture of the supersolid phase more adequately . so far , the only observation of the superfluid glass phase was numerical ( at @xmath113 and @xmath1143 ) . most properties of the helium glass phase , and even its very existence , are a mystery . theoretically , more work has to be done to quantify the sg state , and to see whether the heuristic phase diagram shown in fig . [ fig4 ] takes place . experimentally , one may think of preparing glassy solid samples by fast density increase at low temperature rather then temperature quench at high density . experimentally , it seems that in helium pressure increase can be achieved on a much shorter time scale than cooling down to @xmath113 . [ one objection to this proposal is that crystal growth from the cell wall at low temperature is extremely fast @xcite . ] another way of looking for non - ergodic properties of helium samples , i.e. system dependence on the intial conditions and preparation protocol , is to compare samples cooled while rotating with samples rotated after cooling @xcite . at the moment , the grain boundary flow experiments give us the most unambiguous clue to the solution of the supersolid puzzle . on the other hand , a whole plethora of new questions arise in this connection . the problem is very broad , not only because crystalline interfaces can be prepared at different angles with respect to the crystal axes , but also because one can get different results , depending on how well the initial state was `` annealed '' ( with respect to minimizing the classical deformation energy ) , and whether defects are `` doped '' with out - of - equilibrium vacancies and/or interstitials ( one may wonder if vacancies cluster on defects the same way they do it in the bulk and whether they can make , say a dislocation core , superfluid ? ) . in the experimental setup of ref . @xcite , the superflow in the boundary layer was observed only at the coexistence curve . it is important to extend this type of studies to higher pressures , though the experimental design has to be modified . one possibility @xcite , is to look at the same topology as in fig . [ fig1 ] with the bulk solid being sandwiched between two vycor plates . since helium in vycor remains superfluid at pressures about @xmath115 higher than the bulk melting value @xmath116 one may think of measuring the flow through the solid in response to the pressure difference between the two arms of the u - tube . another fascinating subject is recrystallization waves in the boundary layer . though possible in principle @xcite and being a logical explanation for the fast disappearance of grain boundaries in the experiment @xcite , recrystallization waves still remain to be seen and detected . formally , there is no one - to - one connection between recrystallization waves and superfluidity in the boundary , e.g. smooth domain walls in the 3d checkerboard solid remain superfluid at zero - temperature , but the theory of the interplay between the two phenomena is basically non - existent . previous nmr studies of 3 solid solutions @xcite provided detailed information on the isotopic impurity hopping dynamics , as well as on interparticle interactions . these studies were performed at relatively high temperatures @xmath117 k and in high quality crystals . it would be interesting to see if nmr techniques can be used to understand what happens to 3 atoms in disordered samples grown from the normal high - density liquid and cooled down to the @xmath118 mk range . if 3 atoms form a dense solution at the defects , one should be able to see it in the increased nmr relaxation rates since coupling between the 3 magnetic moments is the only available mechanism . another exciting possibility is to use nmr techniques to see if 3 solids can be in the metastable glass phase . the anomalously low critical velocity in the torsional oscillator studies is yet to be understood , even at the expense of admitting that the observed phase is , in fact , not superfluid @xcite . supersolids may happen in two ways : either as ideal incommensurate crystals with zero - point vacancies or commensurate crystals full of topological defects . though the original observation of supersolidity in 4 can not be interpreted in terms of zero - point vacancy mechanism in a perfect crystal , things turn out to be far more interesting and exciting than expected . instead of one well - characterized crystal , we now face a whole variety of disordered solid samples . this immediately brings into focus grain boundaries , dislocations , and glassy phases which are , on their own right , new superfluid physical systems . for example , dislocations and grain boundaries should be considered as special low - dimensional objects inseparable from their three - dimensional environment . studies of their fascinating properties make the experimental and theoretical landscape multi - dimensional , since in addition to pressure and temperature , the set of important parameters also includes various orientation angles . there is little doubt that this research field is going to last . we also believe that superfluid amorphous helium is a unique and remarkable state which will add a new angle to studies of glassy systems . apart from understanding the role of crystalline defects in the supersolid phenomenon , it is extremely important to formulate what type of realistic interaction potential between particles can lead to the ideal supersolid phase this long - term goal which started more than 50 years ago is yet to be achieved . existing experimental and theoretical tools are more than capable of making significant progress in achieving these goals , both at the phenomenological and microscopic levels . i am indebted to b. svistunov , m. boninsegni and m. troyer for numerous discussions and valuable suggestions which lead to this work . i acknowledge financial support by the national science foundation under grant no . phy-0426881 . s. wessel and m. troyer , phys . lett . * 95 * , 127205 ( 2005 ) ; d. heidarian and k. damle , phys . lett . * 95 * , 127206 ( 2005 ) ; r. g. melko , a. paramekanti , a. a. burkov , a. vishwanath , d. n. sheng , and l. balents , phys . . lett . * 95 * , 127207 ( 2005 ) ; m. boninsegni and n. prokofev , phys . lett . * 95 * , 237204 ( 2005 ) . l. pollet , m. boninsegni , a. kuklov , n. prokofev , b. svistunov , and m. troyer , in preparation . preliminary results on superfluidity of grain boundaries in 4 were reported at the kitp miniprogram : _ the supersolid state of matter _ , february 6 - 17 , ( 2006 ) ; [ @xmath119 . r. hallock and m. ray , in preparation . h. p. bchler , e. demler , m. lukin , a. micheli , n. prokofiev , g. pupillo , and p. zoller , _ strongly correlated 2d quantum phases with cold polar molecules : controlling the shape of the interaction potential _ , cond - mat/0607294 .	for nearly half a century the supersolid phase of matter has remained mysterious , not only eluding experimental observation , but also generating a great deal of controversy among theorists . recent discovery of what is interpreted as a non - classical moment of inertia at low temperature in solid @xmath0he @xcite has elicited much excitement as a possible first observation of a supersolid phase . in the two years following the discovery , however , more puzzles than answers have been provided to the fundamental issue of whether the supersolid phase exists , in helium or any other naturally occurring condensed matter system . presently , there is no established theoretical framework to understand the body of experimental data on 4 . different microscopic mechanisms that have been suggested to underlie superfluidity in a perfect quantum crystal do not seem viable for 4 , for which a wealth of experimental and theoretical evidence points to an insulating crystalline ground state . this perspective addresses some of the outstanding problems with the interpretation of recent experimental observations of the apparent superfluid response in 4 ( seen now by several groups @xcite ) and discusses various scenarios alternative to the homogeneous supersolid phase , such as superfluidity induced by extended defects of the crystalline structure which include grain boundaries , dislocations , anisotropic stresses , etc . can a metastable superfluid glassy " phase exist , and can it be relevant to some of the experimental observations ? one of the most interesting and unsolved fundamental questions is what interatomic potentials , given the freedom to design one , can support an ideal supersolid phase in continuous space , and can they be found in nature .
the helffer - sjstrand formula was established in @xcite in the following proposition [ hsformula ] let @xmath0 be a self - adjoint operator ( not necessarily bounded ) on a hilbert space @xmath1 . suppose @xmath2 is in @xmath3 and @xmath4 in @xmath5 is an extension of @xmath2 such that @xmath6 on @xmath7 . then we have @xmath8 where @xmath9 is the lebesgue measure on @xmath10 . the existence of the functional calculus was assumed by the authors . davies@xcite showed that the formula ( equation [ hsformula ] ) yielded a new approach to the functional calculus for linear operators on banach spaces under the following hypothesis [ e : hypothesis ] @xmath0 is a closed densely defined operator on a banach space @xmath11 with spectrum @xmath12 . the resolvent operators @xmath13 are defined and bounded for all @xmath14 and @xmath15 for some @xmath16 and all @xmath14 , where @xmath17 . his functional calculus , for operators on banach spaces , was defined for an algebra of slow decreasing smooth functions . davies@xcite pointed out that a functional calculus based upon almost analytic extensions was also constructed by dynkin@xcite . however , the two approaches were quite different and that davies approach was more appropriate for differential operators . + the spectral mapping theorem for the helffer - sjstrand functional calculus was also independently proved by btkai and faanga @xcite . they applied methods from abstract functional analysis and their primary tool was an existing abstract spectral mapping theorem from the theory of banach algebras : let @xmath18 be a commutative , semisimple , regular banach algebra , @xmath19 be a banach algebra with unit , @xmath20 be a continuous algebra homomorphism and @xmath21 . then @xmath22 and @xmath23 denotes the gelfand transform . our exposition , part of the ph.d thesis referred to in the introduction of @xcite , takes a very non - abstract and direct approach to the proof . in particular an existing spectral mapping is not assumed . our sole ingredients , supplementing the tools provided in davies@xcite , are the very elementary observations : * it is possible to join two non - zero points in @xmath24 smoothly without passing through the origin . * ( @xcite problem 8.1.11 ) if @xmath0 is a closed operator and @xmath25 lies in the topolgical boundary of the spectrum of @xmath0 then for every @xmath26 there is a vector @xmath27 with length @xmath28 such that @xmath29 * stokes formula has similarities to the cauchy integral forumla . a compelling argument for a direct proof that does not rely on spectral mapping results from the theory of banach algebras follows from the claim by davies@xcite that all the calculations in the construction of his functional calculus can all be carried out at a banach algebra level rather than at an operator level , provided one has a resolvent family in the algebra satifying the obvious analogue of hypothesis [ e : hypothesis ] . + + in the last part of our exposition we derive a functional calculus for operators with spectra bounded on one side . our main tool here is an extension operator of seeley : @xmath30 we summarize some of the main aspects of the helffer - sjstrand functional calculus presented in davies@xcite and some properties of the algebra . let @xmath31 be a smooth function such that @xmath32 then given an interval @xmath33 $ ] we define the approximate characteristic function @xmath34 , { \epsilon } } } } $ ] @xmath35 , { \epsilon } } } } { \left ( x \right ) } = \psi_{{}_{a,\epsilon}}{\left ( x \right)}- \psi_{{}_{b+\epsilon,\epsilon}}{\left ( x \right)}\ ] ] which has support @xmath36 $ ] and is equal to @xmath28 in @xmath33 $ ] and is smooth . for @xmath37 let @xmath38 to be the set of all complex - valued smooth functions defined on @xmath7 where for every @xmath39 there is a positive constant @xmath40 such that where @xmath41 we then define the algebra @xmath42 @xmath43 is an algebra under point - wise multiplication . for any @xmath2 in @xmath43 the expression @xmath44 defines a norm on @xmath43 for each @xmath39 . moreover @xmath45 is dense in @xmath43 with this norm . the function @xmath46 is in @xmath43 for each @xmath47 the statement follows from the observations that if @xmath48 and @xmath49 then @xmath50 and @xmath51 [ e : second ] let @xmath52 . if @xmath2 is in @xmath53 then the function @xmath54 is also in @xmath53 when @xmath55 is large then @xmath56 for some @xmath57 . moreover @xmath58 and @xmath59 [ lem : stable ] if @xmath60 for @xmath61 and @xmath62 then @xmath63 @xmath64 the following concept of almost analytic extensions is due to hrmander ( * ? ? ? * p63 ) . [ almostanalytic ] let @xmath65 be a smooth function such that @xmath66 then given @xmath67 we define an almost analytic extension @xmath4 as @xmath68 moreover we define @xmath69 the following lemma establishes the construction of the new functional calculus [ e : important ] let @xmath70 then define @xmath71 where @xmath4 is an almost - analytic version of @xmath2 as defined in definition [ almostanalytic ] . then 1 . if @xmath72 then subject to hypothesis [ e : hypothesis ] the integral is norm convergent for all @xmath2 in @xmath43 and @xmath73 2 . the operator @xmath74 is independent of @xmath39 and the cut - off function @xmath75 , subject to @xmath72 3 . if @xmath2 is a smooth function of compact support disjoint from the spectrum of @xmath0 then @xmath76 4 . if @xmath2 and @xmath77 are in @xmath43 then @xmath78 5 . if @xmath79 and @xmath80 for all @xmath81 then @xmath82 and @xmath83 given @xmath84 , @xmath85 in @xmath24 we define the curve @xmath86 in the complex plane @xmath87 where @xmath88 $ ] and @xmath89 the important property of @xmath86 is that it is able to connect two non - zero points in the complex plane without intersection with the origin . [ deeep ] let @xmath90 . if @xmath2 is a smooth complex valued function in the interval @xmath91 $ ] where @xmath92 and @xmath93 then there is a smooth function @xmath94 in @xmath95 \right)}$ ] such that @xmath96 : h{\left ( x \right)}=\lambda\ } \text { is empty } \ ] ] moreover @xmath97 and all derivatives of @xmath97 vanish at @xmath98 and @xmath99 . let @xmath100 since @xmath2 is continuous we know there is an @xmath101 such that @xmath96/{\left ( a+\epsilon , b-\epsilon \right ) } : f{\left ( x \right)}=\lambda\ } = \emptyset\ ] ] then we can define @xmath102 , { \epsilon } } } } \right)}f + { \psi _ { { } _ { \small[{a+\epsilon},{b-\epsilon}\small ] , { \epsilon } } } } g\ ] ] [ gotcha ] given @xmath70 , let @xmath25 be a non - zero point in @xmath24 and let @xmath103 + if @xmath104 is empty then there is a function @xmath105 such that @xmath106 for all @xmath107 and @xmath108 if @xmath109 is empty then we put @xmath110 . + if @xmath109 is not empty then @xmath109 is a compact subset of @xmath111 . moreover @xmath109 can be covered by a finite set of closed disjoint intervals @xmath112 $ ] which are also subsets of @xmath111 . by applying theorem [ deeep ] to each interval we can find a function @xmath94 in @xmath113 such @xmath114 and @xmath115 . moreover since @xmath116 has compact support in @xmath111 then it follows from lemma [ e : important](iii ) that @xmath117 . we let @xmath118 be a bounded operator satisfying hypothesis ( [ e : hypothesis ] ) . moreover let @xmath119 [ compactifyf ] for any @xmath70 and @xmath26 @xmath120 , { \epsilon } } } } { \left ( b \right ) } = f{\left ( b \right)}\ ] ] where @xmath121 and @xmath122 suppose @xmath2 has compact support then @xmath123 , { \epsilon } } } } $ ] has compact support disjoint from the spectrum of @xmath118 hence by lemma [ e : important](iii ) the statement of the lemma is true for functions in @xmath3 . the statement for all @xmath70 follows from density of @xmath3 in @xmath113 . [ calcforbo ] let @xmath124 . if @xmath26 and @xmath125 then @xmath126 by lemma [ compactifyf ] we can assume that @xmath2 has compact support in @xmath127 $ ] + if @xmath128 and @xmath129 is the annulus @xmath130 then @xmath131 applying stokes theorem @xmath132 and letting @xmath133 be large enough for @xmath4 to vanish on @xmath134 completes the proof . [ oneforbo ] let @xmath26 . if @xmath135 and @xmath136 then @xmath137 , { \epsilon } } } } { \left ( b \right ) } = 1\ ] ] let @xmath138 and define @xmath139 as the open rectangle @xmath140 using a similar argument to that given in the proof of lemma [ calcforbo ] we see that @xmath137 , { \epsilon } } } } { \left ( b \right ) } = \frac{1}{2\pi i}\int_{\partial \omega } { \widetilde{\psi } _ { { } _ { \small[{l'},{u'}\small ] , { \epsilon } } } } { \small ( z,\overline{z } \small)}{\left ( z - b \right)}^{-1 } dz - \frac{1}{\pi}\int_{\omega } \frac{\partial { \widetilde{\psi } _ { { } _ { \small[{l'},{u'}\small ] , { \epsilon } } } } } { \partial \overline{z}}{\left ( z - b \right)}^{-1}dxdy\ ] ] when @xmath141 then @xmath142 , { \epsilon } } } } { \small ( x \small ) } = 1 $ ] . moreover when @xmath141 then @xmath143 , { \epsilon } } } } { \small ( x \small ) } = 0 $ ] for all @xmath144 . recalling definition ( [ e : ae ] ) we can see that @xmath145 , { \epsilon } } } } { \small ( z,\overline{z } \small)}=1 \quad \text { for all } z\in \overline{\omega_0}\ ] ] hence @xmath137 , { \epsilon } } } } { \left ( b \right ) } = \frac{1}{2\pi i}\int_{\partial \omega } { \left ( z - b \right)}^{-1 } dz\ ] ] and we conclude with an application of cauchy s integral formula . ] we extend the algebra @xmath43 of slow decaying functions in a trivial but necessary way . let @xmath146 where for each @xmath147 we define @xmath148 moreover we define point - wise addition and multiplication : @xmath149 it is clear that @xmath150 the multiplicative identity and @xmath151 the additive identity are in @xmath152 and the algebra is closed under these operations . + for any @xmath153 we will denote @xmath154 simply by @xmath84 . + given @xmath155 , let @xmath156 and let @xmath157 we have the extended functional calculus . for @xmath158 let @xmath159 along with the implied norm @xmath160 for @xmath158 let : @xmath161 [ lem : thisone ] if @xmath158 and @xmath162 is not in @xmath163 then @xmath164 is in @xmath43 and @xmath165 by re - writing @xmath166 then it is routine exercise in differentiation to show that @xmath167 is in @xmath168 . then since @xmath169 is in @xmath43 , lemma [ lem : stable ] implies the statement . [ inversecor ] given @xmath170 and @xmath171 such that @xmath172 for all @xmath107 then @xmath173 [ auto ] if @xmath174 is in @xmath152 then @xmath175 given @xmath176 which is not in @xmath177 we have by by corollary [ inversecor ] @xmath178 hence @xmath179 exists and is bounded and therefore @xmath180 . [ unbounded ] if @xmath174 is in @xmath152 then @xmath181 let @xmath171 be such that @xmath182 and let @xmath183 if @xmath184 then by lemma [ gotcha ] we have that there is function @xmath94 in @xmath113 such that @xmath185 and @xmath186 moreover @xmath187 if @xmath188 it follows from lemma [ e : important ] we have @xmath189 since @xmath190 , the statement of the lemma follows from lemma [ auto ] . [ bbounded ] let @xmath191 . if @xmath0 is bounded and @xmath192 then @xmath193 let @xmath194 and @xmath195 . + let @xmath196 such that @xmath197 is not zero on @xmath198 $ ] and on @xmath199 $ ] . + then let @xmath200 and @xmath201 . + the set @xmath202 : \pi_{{}_{{{\cal a}},\phi}}{\small ( x \small)}=0\}\ ] ] can be covered by a finite number of disjoint intervals @xmath203 $ ] which are all disjoint from @xmath204 and are all in @xmath205 $ ] . applying lemma [ deeep ] to each @xmath203 $ ] we can find a function @xmath124 such that @xmath202 : f{\small ( x \small)}=0\}=\emptyset\ ] ] and @xmath206 for all x in @xmath207 $ ] . + let @xmath77 be any function in @xmath113 such that @xmath208 for all @xmath209 $ ] + by lemma [ e : important](iii ) we have @xmath210 and by lemma [ compactifyf ] we have @xmath211 , { \epsilon } } } } \small)}{\small ( h \small ) } = { \psi _ { { } _ { \small[{l'},{u'}\small ] , { \epsilon } } } } { \small ( h \small)}\ ] ] hence by lemma [ oneforbo ] we have @xmath212 and consequently @xmath213 if @xmath174 in @xmath152 then @xmath214 if @xmath0 is unbounded then @xmath215 and the theorem follows from lemma [ unbounded ] . if @xmath0 is bounded and there is an @xmath216 such that @xmath217 then @xmath215 and again the theorem follows from [ unbounded ] . if @xmath0 is bounded and @xmath218 for all @xmath216 then @xmath219 by lemmas [ unbounded ] and [ bbounded ] . [ e : third ] given @xmath220 and a function @xmath124 , let @xmath221 and let the function @xmath222 be defined as in lemma [ e : second ] then @xmath223 this statement follows directly from the functional calculus and the observation @xmath224 let @xmath2 be a function in @xmath113 then @xmath225 we observe the identity @xmath226 for some @xmath107 . + let @xmath220 . suppose there is a sequence of unit length vectors @xmath227 such that @xmath228 . using identity ( [ loid ] ) we have @xmath229 . by applying lemma [ e : third ] we can conclude that @xmath230 . + the accumulation points of @xmath231 are in @xmath232 since the latter is closed . let @xmath174 be a function @xmath152 then @xmath233 we now assume that @xmath0 is self - adjoint and @xmath11 is a hilbert space . + the following theorem of davies extends the helffer - sjstrand functional calculus to @xmath234 for self - adjoint operators . [ safc ] the functional calculus may be extended to a map from @xmath235 to @xmath236 with the follwoing properties : 1 . @xmath237 is an algebra homomorphism . 2 . @xmath238 3 . @xmath239 4 . if @xmath79 and @xmath80 for all @xmath81 then @xmath83 moreover the functional calculus is unique subject to these conditions . if @xmath235 then @xmath240 this is a consequence of the density of @xmath113 in @xmath234 . by the stone - weierstrass theorem the linear subspace @xmath241 is dense in @xmath234 . if @xmath242 is close to @xmath2 and if @xmath243 is of length 1 then @xmath244 the statement then follows from lemma [ safc](iii ) if @xmath235 then @xmath245 let @xmath246 suppose @xmath90 is not in the closure of @xmath231 . then there is @xmath247 such that @xmath248 also for all large enough @xmath39 we have @xmath249 . then from @xmath250 we can deduce that @xmath251 hence @xmath252 and @xmath253 . + from the identity @xmath254 we can deduce that @xmath255 . we modify our main hypothesis ( [ e : hypothesis ] ) by assuming that the spectrum of @xmath0 is bounded below and without loss of generality @xmath256 . + we introduce a new ring of functions @xmath257 @xmath258 is the set of smooth functions on @xmath259 with the same decaying property as @xmath260 that is for every @xmath39 there is positive constant @xmath40 such that @xmath261 then @xmath257 is defined appropriately and similarly we define the banach space @xmath262 with norm @xmath263 if @xmath2 and @xmath77 are elements of @xmath43 such that @xmath291}=g\|_{[0,\infty]}$ ] and the spectrum of @xmath0 is @xmath292 then it is not necessary that @xmath293 is empty , since @xmath294 is possible . let @xmath309 , then by seeley s extension theorem there exists an extension @xmath310 . we define @xmath311 . this definition is independent of the particular extension by corollary [ last ] . the functional analytic properties are inherited from the extension . this research was funded by an epsrc ph.d grant 95 - 98 at kings college , london . i am very grateful to e. brian davies for giving me this problem , his encouragement since and for continuing to be a mentor in mathematics long after having finished supervising my ph.d . i am indebted to anita for all her support . 99 a. btkai and e. faanga _ the spectral mapping theorem for davies functional calculus_. rev . roumaine math . pures appl . 48 ( 2003 ) , no . 4 , 365372 . e.b . davies , _ the functional calculus _ , j. london math . soc vol 52 , number 1 166 - 176 e.b . davies , _ spectral theory and differential operators _ , cambridge studies in adv . math . 42 , cambridge univ . press , 1995 . davies , _ linear operators and their spectra _ , cambridge studies in adv . 106 , cambridge univ . press , 2007 . e. m. dynkin , _ an operator calculus based upon the cauchy - green formula , and the quasi - analyticity of the classes d(h ) _ sem . math . v a steklov math . leningrad 19 ( 1972 ) 128 - 131 . b. helffer and j. sjstrand , _ quation de schrdinger avec champ magntique et quation de harper , schrdinger operators _ ( snderborg , 1988 ) eds . h. holden and a. jensen , lecture notes in phys . 345 , springer - verlag , berlin , 1989 , pp . 118 - 197 l. hrmander , _ the analysis of linear partial differential operators _ vol 1 , springer , new york 1993 s.t . seeley _ extensions of @xmath318 functions defined on a half space _ , proc . soc 15 1964	we give a direct non - abstract proof of the spectral mapping theorem for the helffer - sjstrand functional calculus for linear operators on banach spaces with real spectra and consequently give a new non - abstract direct proof for the spectral mapping theorem for self - adjoint operators on hilbert spaces . our exposition is closer in spirit to the proof by explicit construction of the existence of the functional calculus given by davies . we apply an extension theorem of seeley to derive a functional calculus for semi - bounded operators . + + * ams subject classification : 47a60 + keywords : functional calculus , spectral mapping theorem , spectrum , *
the topic of open access publishing has been extensively and controversially discussed . for general information see , for instance , @xcite . there are many models of open access named after different colors such as golden or green roads to open access @xcite , but we do not want to go into more details , here . concerning scientific publishing particularly in mathematics , we refer to @xcite and @xcite as well as to the references therein . in @xcite the authors illustrate a game - theoretic approach to open access publishing in order to understand different publication patterns within different scientific disciplines . the underlying observation is that there are communities where open access publishing is widely adopted , whereas other scientific communities are far away from practicing any open access publishing . first , different classical game settings are discussed namely a zero sum game , a game similar to the prisoners dilemma ( up to sign ) , and a stag hunt game version , all describing a nash equilibrium dilemma of the non - open access communities . second , these classical settings are transferred into their quantum game extensions which allow to tackle the dilemma mentioned that can not be solved within the classical setting . the classical approach in @xcite is formulated as a two - player game where the players are authors ( scientists ) in symmetrical situations . the two authors have the same set @xmath0 of strategies at their disposal . consequently , the game looks exactly the same to both of its players . for the open access game each player has to decide between the option to publish open access as the first strategy @xmath1 or , as the second strategy @xmath2 , to conventionally publish in traditional journals where articles go through peer reviewing . depending on the strategy chosen by the co - player , the players aim to maximize their success of the game . due to the symmetric situation both players do not only have the same set of strategies but also have the same payoff matrices . in contrast to the approach of a symmetric two - scientists game , the present paper attempts to explain mathematically the open access play as a conflict of interest between scientists and publishers , as it is emphasized by the rapidly accumulating literature on open access and electronic publishing . assuming that the open access problem is not a conflict between those scientists who publish open access and those who do not , one takes the view that it is more adequate to understand the problem of open access as an asymmetric conflict between scientists and publishers . the crucial point of this view is that it actually suggests a bimatrix game describing this asymmetric setting . we do not explain why both players scientists and publishers behave as they do in the game situation described , but we will give a mathematical description of their interaction in the game . moreover , we do not argue for preferences in choosing one game strategy over another . we consider scientists and publishers as players in different positions them having different strategy sets as well as different payoff matrices . moreover , we take into consideration whole populations of players parts of which choose either one or the other strategy . the set of pure strategies for the scientists is @xmath0 , where @xmath1 is publishing open access and @xmath2 stands for conventional publishing . furthermore , let @xmath3 and @xmath4 denote the relative parts of the scientists population playing strategy @xmath1 and @xmath2 , respectively . if , for example , half of all scientists publishes open access and the other does not , one has @xmath5 . or , if the relation is @xmath6 to @xmath7 , then @xmath8 and @xmath9 . since @xmath10 and @xmath11 describe relative frequencies , we see that @xmath12 always holds true . hence , the set of all possible mixed strategies for the scientists population is @xmath13 the opponent population of publishers consists of two types playing pure strategies @xmath14 . here strategy @xmath15 means the publisher accepts or realizes open access publishing , whereas @xmath16 represents the strategy of definite declining any way of open access . for the publishers population let @xmath17 be the frequency of strategy @xmath18 , where @xmath19 . thus , all possible mixed strategies for the publishers population are given by the set @xmath20 in order to discuss the conflict of open access publishing , we proceed with establishing the payoff matrices @xmath21 for the scientists and @xmath22 for the publishers . each payoff depends on the strategies chosen by the participating players . a player in the scientists population using strategy @xmath23 against a player from the publishers population using strategy @xmath24 obtains the payoff @xmath25 , whereas the opponent obtains @xmath26 , with @xmath27 . as reasoned in @xcite , it is convenient to assume that the scientist tries to maximize his scientific reputation . let @xmath28 denote the reputation payoff that the scientist is awarded . this payoff will be slightly reduced by @xmath29 in case of publishing open access . hence , we still have @xmath30 for the total reputation . in addition , we take into consideration the impact of an article by the scientist . the impact gives some payoff @xmath31 both for the scientist and for the publisher . further , this impact obviously will reduce by some @xmath32 , if the concerning journal is not sufficiently available . moreover , we assume that @xmath33 saying that the reduced impact in case of non - open access publishing is smaller than the loss of reputation in case of open access publishing . open access publishing will cause expenses @xmath34 that are shared equally by both players if both play their open access strategy . furthermore , @xmath35 expresses some moderate journal price scientists have to pay for library subscriptions as well as the compensation for expenses of the publisher . finally , let us introduce @xmath36 for exorbitant profit representing the prices of the most expensive journals coupled with continuing non - open access publishing and taking the tremendously increasing fees for access for granted . for our considerations , we make the realistic assumption that @xmath37 is so large that the inequality @xmath38 holds true , which is equivalent to @xmath39 we already have @xmath40 and the open access expenses @xmath41 actually can not exceed the upper bound given by this inequality . if an open access publishing scientist meets a publisher accepting open access , the scientist will get his reduced reputation , his research impact , and will pay for maintaining some open access publishing system as well as for the journal subscription . the publisher takes the price for the journal subscription and some amount given by the impact . apart from this , the publisher contributes to some extent to open access . hence , the payoff is @xmath42 for the scientist and @xmath43 for the publisher . if the scientist s strategy is open access whereas the publisher plays the non - open access strategy , the scientist s payoff will be @xmath44 . the scientists have to run the open access completely by themselves and do not have any expenses for journal subscriptions . for the publisher nothing happens and hence the payoff is @xmath45 . a non - open access publishing scientist gets his full reputation @xmath46 but only reduced research impact @xmath47 . additionally , the journal has to be paid , which reduces the payoff by @xmath48 . if the publisher s counter - strategy is open access , the payoff for the scientist will be @xmath49 . in this case the payoff for the publisher will be @xmath50 paying the total sum for open access acceptance . moreover , the payoff is @xmath51 for the scientist and @xmath52 for the publisher in case of the non - open access strategy of the publisher . altogether , we have the complete income - and - loss statement as given in table [ tabelle1 ] . .[tabelle1 ] complete payoffs [ cols="^,^,^ " , ] where the horizontal arrows picture the change of the publishers strategies and the vertical arrows illustrate the deviation in the scientists strategies . this dealing with an `` oscillating '' system requires a non - static approach . therefore , we consider the frequencies of the strategies within both populations as time - dependent quantities which are differentiable functions of @xmath53 , i.e. @xmath54 and @xmath55 for @xmath27 . the first derivatives of these functions @xmath56 describe the rate of growth of the respective frequencies . modeling the dynamics and the cyclic behavior of that system , we follow the ideas of hofbauer and sigmund developed in @xcite , chapter 10.3 . the increase of strategy @xmath23 of the scientists population is given by the per capita growth rate @xmath57 and equals the difference between its payoff @xmath58 and the average payoff @xmath59 . the same holds true for the publishers population . this leads to the system of ordinary differential equations @xmath60 for @xmath27 . let us look closer at these differential equations . we have @xmath61 and @xmath62 hence , @xmath63 and @xmath64 since @xmath12 and @xmath65 , it is admissible to introduce new variables @xmath66 and @xmath67 ( then @xmath68 and @xmath69 ) and consider the corresponding differential equations for @xmath70 and @xmath71 . these are @xmath72 and similarly @xmath73 now , let us introduce new constants by @xmath74 which are positive by lemma [ lemma - positive ] . hence , only two equations are remaining namely @xmath75 this is an adequate form for investigating the qualitative behavior of the system . the subsequent section is devoted to a further investigation of the dynamics of the system given by equations ( [ sys-1 ] ) and ( [ sys-2 ] ) . we consider the square @xmath76 . obviously , @xmath77 by @xmath78 we use symplectic techniques for investigating what happens in the interior @xmath79 of @xmath80 . more precisely , in order to prove that the orbits of the system ( [ sys-1 ] ) and ( [ sys-2 ] ) are closed and do not leave @xmath79 we describe the whole setting as a hamiltonian system . the interior of @xmath80 is the open subset @xmath81 of @xmath82 . the function @xmath83 given by @xmath84 is positive . hence , @xmath85 defines a @xmath86-form @xmath87 on @xmath88 . since we are in dimension @xmath86 the following is clear . the pair @xmath89 is a symplectic manifold . @xmath90 we are going to derive that equations ( [ sys-1 ] ) and ( [ sys-2 ] ) are the equations of motion of the hamiltonian system @xmath91 where the hamiltonian @xmath92 is defined as @xmath93 the notion of hamiltonian systems is well - known and widely used in mathematical physics to describe the behavior of systems in classical mechanics . in physics the hamiltonian @xmath94 describes the energy of the system under consideration . hence , it would be interesting to understand , what intrinsic property is determined by the `` energy '' of open access publishing . [ ham_vf ] the hamiltonian vector field of @xmath91 is given by @xmath95 generally , @xmath96 is defined by the equation @xmath97 clearly , @xmath98 if @xmath99 is any vector field on @xmath88 with functions @xmath100 and @xmath101 as components , then @xmath102 hence , @xmath103 if and only if @xmath104 which proves the assertion . let us remind that motions of the hamiltonian system @xmath91 are curves @xmath105 in @xmath88 satisfying the differential equation @xmath106 altogether , we now can prove the following proposition . the solutions of ( [ sys-1 ] ) and ( [ sys-2 ] ) are exactly the motions of the hamiltonian system @xmath91 . we have @xmath107 and @xmath108 thus , equations ( [ sys-1 ] ) and ( [ sys-2 ] ) are equivalent to @xmath109 which implies the statement by lemma [ ham_vf ] . as a consequence , the solutions in @xmath88 of the system ( [ sys-1 ] ) and ( [ sys-2 ] ) correspond to level set curves of the hamiltonian @xmath94 , i.e. the hamiltonian @xmath94 is constant along solutions . let @xmath110 give a solution of equations ( [ sys-1 ] ) and ( [ sys-2 ] ) . then @xmath111 indeed , @xmath112 by using ( [ sys-1 ] ) as well as ( [ sys-2 ] ) for replacing @xmath113 and @xmath114 . in the remainder of this section we want to discuss the shape of the level set curves of @xmath94 . in particular , we will show that these curves are closed , which implies that the solutions of ( [ sys-1 ] ) and ( [ sys-2 ] ) are periodic . first , we are going to determine the extremal points of @xmath94 . by @xmath115 and @xmath116 we obtain @xmath117 as necessary conditions for a critical point @xmath118 . this leads to @xmath119 which corresponds exactly to our nash equilibrium described in lemma [ nash - equ ] . by @xmath120 we have @xmath121 at the point @xmath118 . this implies that the matrix @xmath122 is negative definite showing that @xmath94 takes in @xmath118 a maximum as the unique extremum . second , since @xmath123 we have @xmath124 for each @xmath125 in the boundary @xmath126 of @xmath80 . hence , the graph of the hamiltonian @xmath94 looks like a hill over the square @xmath80 . with parameters @xmath127 , @xmath128 , @xmath129 , and @xmath130,width=264 ] the top of the hill represents the unique maximum . towards the boundary @xmath126 , the hill crashes into bottomless depths . with parameters @xmath127 , @xmath128 , @xmath129 , and @xmath130,width=226 ] since all solutions of our system correspond to level set curves , they are visualized as contour lines of the hill described . hence , all orbits surround the point @xmath118 and remain in @xmath88 . obviously , the model presented here is a basic approach to discuss the phenomenon of multiple arrangements of realizing open access publishing in different scientific communities . however , in contrast to the approach presented in @xcite it treats open access publishing in a canonical way as a game between scientists and publishers . the behavior of oscillation on closed orbits is a rather rough and simplified view of the whole scenario . we do not consider this model to be universally valid , but throughout a certain period of time it seems to allow a reasonably satisfying description of insights which are intuitively clear . this model can be modified . for instance , taking into account that both the parameter @xmath131 reducing the scientist s reputation in case of open access publishing and the value of @xmath132 expressing the reduced impact in case of non - open access publishing may change in the context of the game . for example , if the rate of open access publishing scientists increases , @xmath131 will be supposed to decrease etc . playing a little bit and slightly altering the model presented here may result in further refinements . moreover , the huge number of evolutionary game - theoretic approaches in the literature allows a lot of modifications and further developments of these concepts . they vary from biological game theory to concepts that are more applicable in economic contexts . these models are good sources and suitable to establish further improvements in a more fundamental manner , see e.g. @xcite and @xcite as well as the references therein . the authors have benefited from many discussions with both scientists and information professionals and particularly from the working context of the first author as a mathematician as well as a subject librarian . the idea developed in this note arose during the 5th spring school on geometry and mathematical physics at hiddensee in may 2007 . both authors thank the biological station of hiddensee for the kind hospitality as well as for the inspiring working atmosphere , the organizer of the workshop stephan block for his invitation , and the participants for the intensive discussion .	the paper presents an evolutionary game - theoretic approach to open access publishing as an asymmetric game between scientists and publishers . we show how the ordinary differential equations of the model presented can be written as a system of hamiltonian partial differential equations . the understanding of the setting as a hamiltonian system implies some properties reflecting the qualitative behavior of the system .
my task in this opening lecture is to set the stage for the subsequent lectures that develop in more detail the connections between particle physics and cosmology . to do so , i first recall the essential aspects of standard big - bang cosmology , emphasizing that the questions it raises about the early history of the universe can only be answered by particle physics . the latter is described by its own standard model , which makes successful quantitative predictions for accelerator experiments , but leaves open many fundamental questions . these include the origin of particle masses , the proliferation of different types of elementary particles and the possible unification of all the particle interactions . in combination with accelerator experiments , astrophysics and cosmology may cast important light on the solutions of these problems . according to astrophysicists and cosmologists , most of the matter in the universe has never been seen , and can not consist of ordinary matter @xcite . the formation of structures in the universe would be helped by presence of massive weakly - interacting cold dark matter particles @xcite . candidates for these include the lightest supersymmetric particle ( lsp ) @xcite , the axion @xcite and metastable superheavy particles @xcite whose decays might be responsible for ultra - high - energy cosmic rays beyond the greisen - zatsepin - kuzmin ( gzk ) cutoff @xcite , if they exist . these are just a few of the connections between the very big and the very small that are developed by other lecturers at this summer institute . according to standard big - bang cosmology @xcite , the entire visible universe is expanding homogeneously and isotropically from a very dense and hot initial state . the first direct piece of evidence for this was the discovery by hubble that distant objects in the universe are receding from each other at velocities proportional to their distances from each other : @xmath1 where @xmath2 is the recession velocity , @xmath3 is the distance and the hubble constant @xmath4 km / s / mpc where @xmath5 @xcite . observations of the universe suggest that it is indeed very homogeneous and isotropic on large scales @xmath6 mpc @xcite . the next piece of evidence for the big bang to be discovered was the cosmic microwave background ( cmb ) radiation @xcite . extrapolating the present hubble expansion back in time , the cmb is thought to have been emitted when the universe was about 3000 times smaller and hotter than it is today , with age @xmath7 y. the cmb has a dipole deviation from isotropy at the @xmath8 level , which this is believed to be due to the earth s motion relative to a machian cosmological frame . smaller - scale anisotropies have been discovered more recently by the cobe satellite and subsequent experiments , as seen in fig . [ fig : cmb ] @xcite , and may provide a window on the universe when it was much younger still , as we shall see later . the third piece of evidence for the big bang was provided by the abundances of light elements seen in fig . [ fig : bbn ] , which are thought to have been established when the universe was about @xmath9 times smaller and hotter than it is today @xcite , with age @xmath10 to @xmath11 s. this nuclear ` cooking ' must have occurred when the temperature @xmath12 of the universe corresponded to characteristic particle energies @xmath13 mev . back when the universe was @xmath14 to @xmath15 s old , it is thought to have made a transition from a plasma of quarks and gluons to hadrons at a temperature @xmath16 mev @xcite . previous to that , the electroweak transition when standard model particles acquired their masses is thought have occurred when the universe was @xmath17 to @xmath18 s old , and the temperature @xmath19 gev @xcite . during the expansion of the universe , it acts as a ` cosmic decelerator ' , whose effective temperature @xmath12 falls as it expands @xcite : @xmath20 where @xmath21 is the scale factor measuring the size of the universe . during the early history of the universe when most particle masses were negligible , the rate of expansion was such that the age @xmath22 inserting the units , one finds that the temperature of the universe would have been about @xmath23 k when it was about a second old . such high temperatures correspond to high energies for the thermalized particles : @xmath23 k @xmath10 mev ( cf the electron mass @xmath24 mev ) , @xmath25 k @xmath10 gev ( cf the proton mass @xmath10 gev ) . in general , the time - temperature relation is such that @xmath26 thus , it is clear that the very early history of the universe must have been dominated by elementary particles , and only their physics can explain how the universe got to be the way it is today . as you can see in fig . [ fig : lep ] , data from the lep accelerator are , unfortunately , in excellent agreement with the standard model . indeed , no accelerator data provide evidence for physics beyond the standard model . nevertheless , particle physicists are convinced that there must be accessible physics beyond the standard model , because it leaves many fundamental questions unanswered . we seek the _ origin of particle masses _ and the reason why they are so much smaller than the planck mass @xmath27 gev . are the masses due to a higgs boson , and is it accompanied by supersymmetric particles ? we seek a _ theory of flavour _ , because the standard model has six random - seeming quark masses , three disparate charged - lepton masses , three weak mixing angles and the cp - violating kobayashi - maskawa phase . moreover , we seek a _ grand unified theory _ , because the standard model has three independent gauge couplings and ( potentially ) a cp - violating phase in qcd . altogether , the standard model has a total of 19 parameters , without even addressing the more fundamental questions of the origins of the particle quantum numbers . beyond all these beyonds , other theorists seek a _ theory of everything _ that includes gravity , reconciles it with quantum mechanics , explains the origin of space - time and why we live in four dimensions ( if we do so ) . experiments agree perfectly with the predictions of the standard model @xcite.,width=321 ] non - accelerator neutrino experiments @xcite now provide us with the first direct evidence for physics beyond the standard model , convincing us that neutrinos oscillate and have different non - zero masses . to describe these , we need three neutrino mass parameters , three neutrino mixing angles and three cp - violating phases in the neutrino sector . moreover , we should not forget about gravity , with at least two parameters to understand : newton s constant @xmath28 gev)@xmath29 and the cosmological ` constant ' , which recent data suggest is non - zero @xcite , and may not even be constant . talking of cosmology , we would need at least one extra parameter to produce an inflationary potential , and at least one other to generate the baryon asymmetry , which can not be explained within the standard model . at what energy scale might new physics beyond the standard model appear , between the energies @xmath30 gev already explored and the planck energy @xmath31 gev ? the problem of mass must presumably be solved by new physics at some energy @xmath32 tev , whether it be just a higgs boson or some richer physics such as supersymmetry@xcite . simple ideas of grand unification suggest new physics at a scale @xmath33 gev @xcite . on the other hand , we have no good idea what energy scale might be associated with the solution to the problem of flavour , or where extra dimensions might appear . if there is a significant discrepancy between the bnl measurement of the muon anomalous magnetic moment @xcite and the standard model , which is not yet established , this could only removed by new physics at a scale @xmath32 tev . there are two circumstantial pieces of evidence in favour of grand unification , namely the existence of neutrino masses - which might have been generated at some mass scale between @xmath34 gev and @xmath35 gev - and the weak neutral - current mixing angle @xmath36 . the value of the latter could be explained by grand unification at a scale @xmath33 gev combined with supersymmetry at a scale @xmath37 gev @xcite . the lhc will be able to discover ` any ' new physics at a scale @xmath38 gev , but many of the other ideas mentioned above may not be directly testable at accelerators for the foreseeable future . astrophysics and cosmology may provide the only laboratories for testing some of these ideas . for example , the cosmic microwave background ( cmb ) and inflation may be providing a direct window on physics at the gut scale , and ultra - high - energy cosmic rays ( uhecrs ) might be due to the decays of metastable superheavy particle weighing @xmath39 gev or so @xcite . on the other hand , many particle candidates for dark matter weigh @xmath32 tev and could be detected at the lhc . it seems that accelerators , astrophysics and cosmology are condemned to symbiosis . what does the universe contain ? let us enumerate its composition in terms of the density budget of the universe , measured relative to the critical density : @xmath40 . inflation @xcite suggests that the _ total density _ of the universe is very close to the critical value : @xmath41 , and this estimate is supported by cmb data @xcite . i remind you that inflation explains why the universe is so large : the scale size @xmath42 cm , why the universe is so old : its age @xmath43 s , why its geometry is so nearly flat with a euclidean geometry , and why the universe is so homogeneous on large scales . it achieves these feats by postulating an epoch of ( near- ) exponential expansion during the very early universe , making the universe very large and giving it a long time to recollapse ( if it ever will ) . even the most distant parts of the observable universe would have been very close to each other prior to this inflationary epoch , and so could have synchronized their behaviours . this inflationary expansion would have blown the universe up like an inflated ballon , which seems almost flat to an ant living on its surface . during the inflationary expansion , quantum fluctuations in the inflaton field would have generated small density perturbations ( cf . the observations in fig . [ fig : cmb ] ) capable of growing into the structures seen in the universe today @xcite , as discussed later . big - bang nucleosynthesis suggests that the _ baryon density _ @xmath44 @xcite , an estimate that has been supported by analyses of the relative sizes of small fluctuations in the cmb at different scales @xcite . the baryons are insufficient to explain the _ total matter density _ @xmath45 , as estimated independently by analyses of clusters of galaxies and , more recently , by combining the observations of high - redshift supernovae with those of the cmb . the supernovae constrain the density budget of the universe in a way that is almost orthogonal to the cmb constraint , and is very consistent with the prior indications from galaxy clusters @xcite . observations of the structures that have formed at different scales in the universe suggest that most of the missing dark matter is in the form of non - relativistic _ cold dark matter _ , as discussed in the next session . the theory of structure formation suggests that very little of the dark matter is in the form of _ hot dark matter _ particles that were relativistic when structures started to form : @xmath46 @xcite . applying this constraint to neutrinos , for which @xmath47 this constraint tells us that @xmath48 ev , a limit that is highly competitive with direct limits @xcite . if @xmath49 and the matter density @xmath50 , how do we balance the density budget of the universe ? there must be _ vacuum energy _ @xmath51 with @xmath52 . all the available cosmological data are consistent with @xmath51 having been constant at redshifts @xmath53 , as per einstein s original suggestion of a cosmological constant . however , we can not yet exclude some slowly varying source of vacuum energy , ` quintessence ' with an equation of state parametrized by @xmath54 @xcite . measurable vacuum energy would provide a second general - relativity observable to explain , in addition to the planck mass scale @xmath55 . this would provide a tremendous opportunity for any theory of everything including quantum gravity , such as string . the ultimate challenge for theoretical physics may be to calculate @xmath51 . we have seen that big - bang nucleosynthesis @xcite and the cmb @xcite independently imply that baryons make up only a few % of the density of the universe . numerically , this corresponds to a baryon - to - photon ratio @xmath56 , raising several questions . why is there so little baryonic matter ? why is there any at all ? why is there apparently no antimatter ? astronauts did not disappear in a burst of radiation when they landed on the moon , and neither have space probes landing on mars or an asteroid . the small abundance of antiprotons in the cosmic rays is consistent with their production by primary matter cosmic rays @xcite , and no antinuclei have been seen @xcite . if there were any large concentration of antimatter in our local cluster of galaxies , we would have detected radiation from matter - antimatter annihilations at its boundary . the cmb would have been distorted by similar radiation from any matter - antimatter boundary within the observable universe @xcite . so it seems that there must be a real cosmological asymmetry betwen matter and antimatter . this could be explained if , going back to when the universe was less than @xmath57 s old , it contained about one extra quark for every @xmath58 quark - antiquark pairs in the primordial soup . as the universe expanded , most of the quarks would have annihilated with those antiquarks to produce radiation , and the few quarks left over would have survived to combine into the baryons seen today . where did this small quark - antiquark asymmetry originate ? did the big bang start off with it , or did the laws of nature generate it during the subsequent expansion ? the conditions for such cosmological baryogenesis were established by sakharov in 1967 @xcite . there has to be a difference between the interactions of matter and antimatter particles , in the form of charge - conjugation ( c ) violation , which was discovered in the weak interactions in 1957 , and cp violation , which was discovered in kaon decays in 1964 . there must also have been a departure from thermal equilibrium , which would have been possible during a phase transition , perhaps the electroweak phase transition when @xmath59 s or a gut phase transition when @xmath60 s , or at the end of inflation . finally , there must have been a violation of baryon number , which would have happened through nonperturbative weak interactions at high temperatures @xcite and is thought to be a generic feature of guts . various specific mechanisms for big - bang baryogenesis have been proposed , ranging from the out - of - equilibrium decays of gut bosons @xcite or heavy neutrinos @xcite to processes around the epoch of the electroweak phase transition . the cp violation in the standard model seems inadequate to generate the required baryon asymmetry , but this might be possible if it is extended to include supersymmetry @xcite . how have these hard - won baryons organized themselves into the structures - clusters , galaxies , stars , planets and us - that we see in the universe today ? as already mentioned , the prime candidates for the seeds of these structures are quantum fluctuations in the inflaton field , which would have caused different parts of the universe to expand differently and generated a _ gaussian random field _ of density perturbations if the inflaton energy was roughly constant during inflation , these perturbations would be _ almost scale - invariant _ , as postulated by astrophysicists . the cmb data shown in fig . [ fig : cmb ] are consistent with both these properties . accepting this scenario , the magnitude of the primordial perturbations would be related to the field energy density @xmath61 during inflation : @xmath62 inserting the magnitude of @xmath63 oberved by the cobe and subsequent experiments @xcite , one estimates @xmath64 comparable with the gut scale @xcite . these primordial perturbations would have produced embryonic potential wells into which the non - relativistic cold dark matter particles would have fallen , while relativistic hot dark matter particles would have escaped . in this way , cold matter particles would have amplified the amplitudes of the primordial density perturbations , while the baryons were still coupled to the relativistic radiation . then , when the baryonic matter and radiation ` re- ' combined to form atoms , they would have fallen into the deeper potential wells prepared by the cold dark matter . this theory of structure formation fits remarkably well the data on all scales from over @xmath65 mpc down to @xmath10 mpc @xcite . it is this agreement that provides the most stringent upper limit on the possible hot dark matter such as neutrinos . as discussed earlier , most of the dark matter is thought to be non - relativistic cold dark matter . there are almost as many candidates for this as in the californian gubernatorial election but , as in that case , some of the candidates are more favoured than others . _ lightest supersymmetric particle _ : the existence of supersymmetry at relatively low energies @xmath32 tev is motivated by the hierarchy problem , namely why is the electroweak scale @xmath66 gev , the only candidate we have for a primary mass scale in physics @xcite . alternatively , one may rephrase this question as why the fermi constant @xmath67 , the newton constant , or as why the coulomb potential @xmath68 in an atom . this can be traced to the fact that @xmath69 , which is in turn due to the fact that the masses of particles in atoms @xmath70 . you might think it be sufficient to set @xmath71 by hand and forget about the hierarchy problem . however , this is insufficient because standard model loop corrections to the higgs and/or @xmath72 mass are quadratically divergent : @xmath73 which is @xmath74 if the cutoff @xmath51 where the standard model breaks down and new physics appears @xmath75 or @xmath55 . these loop corrections can be controlled by postulating supersymmetry @xcite , which predicts that bosons and fermions appear in pairs with equal couplings . since the divergences in boson loops are positive and those in fermion loops are negative , ( [ quadratic ] ) is replaced by @xmath76 which is @xmath77 if @xmath78 thus , the loop corrections to the electroweak scale may be made naturally small by postulating small differences between supersymmetric partner particles . it is a generic feature of many supersymmetric models that the lightest supersymmetric particle ( lsp ) is stable , as a result of a particular combination of baryon and lepton number being conserved . this ensures that heavier sparticles can only decay into lighter ones , and the lsp is stable because it has no allowed decay modes . furthermore , generically the lsp is electrically neutral and has only weak interactions , making it an ideal weakly - interacting massive particle ( wimp ) @xcite . moreover , there are generic regions of the supersymmetric space where the relic lsp density falls within the range preferred by the cosmological data , as seen in fig . [ fig : bench ] @xcite . the lsp has many rivals to be the cold dark matter , including axions and the ` schwarzenegger ' candidate , an ultraheavy metastable particle @xcite . the next two sections discuss how these candidates might be elected by experiment . _ annihilations in the galactic halo _ : these would produce some antiprotons , positrons and photons that might be detectable among the cosmic rays @xcite . as already discussed , the oberved antiprotons appear completely consistent with production by primary matter cosmic rays @xcite . the prospects for detecting lsp annihilation positrons do not look bright either , at least in a set of proposed supersymmetric benchmark scenarios @xcite . the prospects for detecting lsp annihilation photons may be brighter , if the lsp density is enhanced in the core of the galaxy . as seen in fig . [ fig : gamma ] @xcite , glast might be the best - placed to detect these . rays from the galactic centre by glast and ground - based experiments may be able to test certain supersymmetric benchmark scenarios @xcite.,width=321 ] _ annihilations in the sun or earth _ : as lsps fly though the galaxy , some of them pass through the solar system on hyperbolic trajectories . if they pass through the sun or earth , they may scatter , and the deposit of recoil energy may convert their trajectories into elliptical orbits with perihelions ( perigees ) below the surface of the sun ( earth ) . they would then scatter repeatedly , losing more energy each time , until they eventually settle into a cloud in the core . there they would annihilate , and any high - energy neutrino they produce might be detectable in undergound experiments @xcite , either by interactions inside the detector or via collisons in nearby material that produce muons passing through the detector . as seen in fig . [ fig : muons ] @xcite , annihilations inside the sun would be detectable in several supersymmetric scenarios , whereas the prospects for detecting terrestrial annihilations do not look so good . detectors for energetic muons originating from the interactions of high - energy neutrinos produced by the annihilations of supersymmetric relic particles captured inside the sun may probe some supersymmetric benchmark scenarios @xcite.,width=321 ] _ direct detection of dark matter scattering _ : in many scenarios , it is also possible to detect directly the scattering of lsps on nuclei in a low - background underground laboratory @xcite , via the few kev of recoil energy deposited . this scattering is expected to have both spin - dependent and spin - independent components , with the latter seeming more promising for the relatively heavy lsps favoured by the absence of sparticles in collider experiments to date . as seen in fig . [ fig : spinind ] @xcite , dark matter may be detectable directly in this way in a number of supersymmetric scenarios , at least in some projected experiments . now for the ` arnold ' candidate @xcite . the spectrum of cosmic rays falls almost featurelessly @xmath79 until @xmath80 gev . at energies @xmath81 gev , protons or nuclei coming from more than @xmath82 mpc away would have scattered on cmb photons before reaching us , producing pions and losing energy - the gzk cutoff @xcite . the agasa experiment @xcite has reported seeing ultra - high - energy cosmic rays ( uhecrs ) beyond the gzk cutoff , but the hires experiment @xcite does not . the auger experiment @xcite now under construction in argentina should be able to tell us definitively whether such uhecrs exist . what might be their origins ? the most plausible is some ` bottom - up ' mechanism of acceleration by astrophysical sources . the upper limit on the energy attainable with such a cosmic accelerator is @xmath83 where @xmath84 is the atomic number of the accelerated nucleus , @xmath85 is the size of the cosmic accelerator and @xmath86 its magnetic field . fig . [ fig : hillas ] shows some of the possible astrophysical sources of uhecrs , including neutron stars , active galactic nuclei ( agns ) , radio - galaxy lobes and galactic clusters with gamma - ray bursters ( grbs ) being other possible sources . in any scenario based on such discrete sources , one might expect clustering of arrival directions and correlations with astrophysical objects , as has sometimes been claimed @xcite . alternatively , one might postulate some ` top - down ' model based on the decays or interactions of massive gut - scale particles @xcite . such superheavy candidates include topological defects and metastable relic particles , such as the cryptons expected as relics of the hidden sector in string models @xcite . the energy spectrum of their decays can fit the agasa spectrum of uhecrs @xcite , as seen in fig . [ fig : toldra ] , but their composition is potentially an issue . in such models , most of the uhecrs would arise from decays within the halo of our own galaxy , and their arrival directions should be anisotropic . there could be some clustering if a large fraction of the cold dark matter in our galactic halo is clumped . copernicus taught us that we do not live at the centre of the universe . modern astrophysicists teach us that we are not made of the same stuff as most of the matter in the universe , and modern cosmologists teach us that matter is not even the dominant form of energy in the universe . the challenge for coming observations is to prove these assertions and determine the nature of the missing matter and energy . the rest of this summer institute will provide you with some of the tools you will need for this task . n. a. bahcall , j. p. ostriker , s. perlmutter and p. j. steinhardt , science * 284 * , 1481 ( 1999 ) [ arxiv : astro - ph/9906463 ] . see , for example : f. r. pearce _ et al . _ , _ simulations of galaxy formation in a cosmological volume _ , arxiv : astro - ph/0010587 . j. ellis , j. s. hagelin , d. v. nanopoulos , k. a. olive and m. srednicki , _ nucl . _ b * 238 * , 453 ( 1984 ) ; see also h. goldberg , _ phys . lett . _ * 50 * , 1419 ( 1983 ) . for a review and references , see : i. g. irastorza _ et al . _ , nucl . suppl . * 114 * , 75 ( 2003 ) . v. berezinsky , m. kachelriess and a. vilenkin , _ phys . _ * 79 * , 4302 ( 1997 ) [ arxiv : astro - ph/9708217 ] . k. greisen , _ phys . * 16 * , 748 ( 1966 ) ; g. t. zatsepin and v. a. kuzmin , _ pisma zh . fiz . _ * 4 * , 114 ( 1966 ) . e. w. kolb and m. s. turner , _ the early universe _ ( addison - wesley , redwood city , usa , 1990 ) . c. l. bennett _ et al . _ , astrophys . j. * 464 * , l1 ( 1996 ) [ arxiv : astro - ph/9601067 ] . d. n. spergel _ et al . _ , astrophys . j. suppl . * 148 * , 175 ( 2003 ) [ arxiv : astro - ph/0302209 ] . for a recent review , see : k. a. olive , arxiv : astro - ph/0202486 . for a review , see : h. satz , nucl . a * 715 * , 3 ( 2003 ) [ arxiv : hep - ph/0209181 ] . see , for example : k. kajantie , m. laine , k. rummukainen and m. e. shaposhnikov , nucl . b * 493 * , 413 ( 1997 ) [ arxiv : hep - lat/9612006 ] . for a recent review , see : m. w. grunewald , nucl . suppl . * 117 * , 280 ( 2003 ) [ arxiv : hep - ex/0210003 ] . y. fukuda _ et al . _ [ super - kamiokande collaboration ] , _ phys . lett . _ * 81 * , 1562 ( 1998 ) [ arxiv : hep - ex/9807003 ] . q. r. ahmad _ et al . _ [ sno collaboration ] , _ phys . rev . lett . _ * 89 * , 011301 ( 2002 ) [ arxiv : nucl - ex/0204008 ] ; _ phys . lett . _ * 89 * , 011302 ( 2002 ) [ arxiv : nucl - ex/0204009 ] ; s. n. ahmed _ et al . _ [ sno collaboration ] , arxiv : nucl - ex/0309004 . for an early review , see : p. fayet and s. ferrara , phys . * 32 * ( 1977 ) 249 ; see also : h. p. nilles , phys . * 110 * ( 1984 ) 1 ; h. e. haber and g. l. kane , phys . * 117 * ( 1985 ) 75 . j. ellis , s. kelley and d. v. nanopoulos , _ phys . _ b * 260 * , 131 ( 1991 ) ; u. amaldi , w. de boer and h. furstenau , _ phys . _ b * 260 * , 447 ( 1991 ) ; p. langacker and m. x. luo , _ phys . _ d * 44 * , 817 ( 1991 ) ; c. giunti , c. w. kim and u. w. lee , _ mod . lett . _ a * 6 * , 1745 ( 1991 ) . g. w. bennett _ et al . _ [ muon g-2 collaboration ] , _ phys . lett . _ * 89 * , 101804 ( 2002 ) [ erratum - ibid . * 89 * , 1219903 ( 2002 ) ] [ arxiv : hep - ex/0208001 ] . j. r. ellis , j. l. lopez and d. v. nanopoulos , _ phys . _ b * 247 * , 257 ( 1990 ) ; k. benakli , j. r. ellis and d. v. nanopoulos , _ phys . _ d * 59 * , 047301 ( 1999 ) [ arxiv : hep - ph/9803333 ] . for a review , see : d. h. lyth and a. riotto , _ phys . rept . _ * 314 * , 1 ( 1999 ) [ arxiv : hep - ph/9807278 ] . see , for example : j. m. bardeen , p. j. steinhardt and m. s. turner , _ phys . d * 28 * , 679 ( 1983 ) . k. hagiwara _ et al . _ [ particle data group collaboration ] , phys . d * 66 * , 010001 ( 2002 ) . l. m. wang , r. r. caldwell , j. p. ostriker and p. j. steinhardt , astrophys . j. * 530 * , 17 ( 2000 ) [ arxiv : astro - ph/9901388 ] . t. maeno _ et al . _ [ bess collaboration ] , astropart . * 16 * , 121 ( 2001 ) [ arxiv : astro - ph/0010381 ] . m. aguilar _ et al . _ [ ams collaboration ] , phys . rept . * 366 * , 331 ( 2002 ) [ erratum - ibid . * 380 * , 97 ( 2003 ) ] . a. g. cohen , a. de rujula and s. l. glashow , astrophys . j. * 495 * , 539 ( 1998 ) [ arxiv : astro - ph/9707087 ] . a. d. sakharov , _ pisma zh . * 5 * , 32 ( 1967 ) g. t hooft , phys . d * 14 * , 3432 ( 1976 ) [ erratum - ibid . d * 18 * , 2199 ( 1978 ) ] . m. yoshimura , phys . lett . * 41 * , 281 ( 1978 ) [ erratum - ibid . * 42 * , 746 ( 1979 ) ] . m. fukugita and t. yanagida , _ phys . lett . _ b * 174 * , 45 ( 1986 ) . see , for example : m. carena , m. quiros , m. seco and c. e. wagner , nucl . b * 650 * , 24 ( 2003 ) [ arxiv : hep - ph/0208043 ] . l. maiani , _ proceedings of the 1979 gif - sur - yvette summer school on particle physics _ , 1 ; g. t hooft , in _ recent developments in gauge theories , proceedings of the nato advanced study institute , cargese , 1979 _ , eds . g. t hooft _ _ , ( plenum press , ny , 1980 ) ; e. witten , _ phys . _ b * 105 * , 267 ( 1981 ) . m. battaglia _ et al . _ , _ eur . c * 22 * , 535 ( 2001 ) [ arxiv : hep - ph/0106204 ] , as updated in m. battaglia , a. de roeck , j. r. ellis , f. gianotti , k. a. olive and l. pape , arxiv : hep - ph/0306219 . d. j. chung , p. crotty , e. w. kolb and a. riotto , phys . d * 64 * , 043503 ( 2001 ) [ arxiv : hep - ph/0104100 ] . j. ellis , j. l. feng , a. ferstl , k. t. matchev and k. a. olive , arxiv : astro - ph/0110225 . j. silk and m. srednicki , _ phys . _ * 53 * , 624 ( 1984 ) . j. silk , k. a. olive and m. srednicki , _ phys . _ * 55 * , 257 ( 1985 ) . m. w. goodman and e. witten , _ phys . _ d * 31 * , 3059 ( 1985 ) . m. takeda _ et al . _ , arxiv : astro - ph/0209422 . t. abu - zayyad _ et al . _ [ high resolution fly s eye collaboration ] , arxiv : astro - ph/0208243 and arxiv : astro - ph/0208301 . a. letessier - selvon , arxiv : astro - ph/0208526 ; j. cronin _ et al . _ , + http://www.auger.org/. p. g. tinyakov and i. i. tkachev , _ pisma zh . fiz . _ * 74 * , 3 ( 2001 ) [ arxiv : astro - ph/0102101 ] . see , for example , s. sarkar and r. toldra , nucl . b * 621 * , 495 ( 2002 ) [ arxiv : hep - ph/0108098 ] .	big - bang cosmology and ideas for possible physics beyond the standard model of particle physics are introduced . the density budget of the universe is audited , and the issues involved in calculating the baron density from microphysics are mentioned , as is the role of cold dark matter in the formation of cosmological structures . candidates for cold dark matter are introduced , with particular attention to the lightest supersymmetric particle and metastable superheavy relics . prospects for detecting supersymmetric dark matter in non - accelerator experiments are assessed , and the possible role of decays in generating ultra - high - energy cosmic rays is discussed . more details of these and other astroparticle topics are presented during the rest of this summer institute . 12m_1/2 12m_1/2 _ _ r /#1#1 _ # 1#2 _ opening lecture at the 31st slac summer institute , july 2003 : psn l01 _ + cern - th/2003 - 268 @xmath0 astro - ph/0310911
m dwarfs comprise the bulk ( @xmath375% ) of the stellar population of our galaxy . moreover , their low masses and small radii , compared to sun - like stars , make planets of a given mass , size and orbital separation much easier to detect around them via both the transit and doppler methods . in addition , the habitable zone ( hz ) of such low luminosity red dwarfs lies considerably closer to the central star than in the case of solar - types ; this further enhances the transit and doppler signatures of hz planets , and also allows a larger number of planetary orbits to be observed in a given time , making the detection and characterisation of such planets even easier . consequently , the next generation of missions investigating exoplanets are aimed at later - type stars rather than the mainly solar - type ones targeted by _ kepler _ , and discussions of exoplanet habitability increasingly focus on m dwarf systems . there are many complications pertaining to the habitability of planets around m - dwarfs , even if they possess surface temperature and pressure conditions favourable to liquid water , such as tidal locking ( e.g. * ? ? ? * ; * ? ? ? * ) , run - away greenhouse effects ( e.g. * ? ? ? * ) and water loss @xcite . the first major uncertainty , however , is whether they can possess such conditions at all , given what we currently know about exoplanets ( e.g. * ? ? ? the last decade of exoplanet discoveries has dramatically altered our understanding of a `` standard '' low - mass planet . _ kepler _ has detected thousands of planetary candidates ( e.g. * ? ? ? * ) , the majority of which are small ( @xmath5 r@xmath6 : * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) and close to their host stars . correcting for observational biases , one finds that most stars probably harbour at least one such `` kepler '' planet @xcite . masses have been now been obtained for a significant number ( albeit a small fraction ) of these exoplanets , by combining _ kepler _ transit data with either transit timing variations ( ttvs ; e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) or radial velocity measurements ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? a very surprising result to emerge from this work is that a large fraction of low - mass , close - in kepler planets are enshrouded by voluminous h / he envelopes , which contain a non - negligible fraction of the planet s mass ( as demonstrated , for instance , by the extremely low densities of some : @xmath7 g @xmath8 ; e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) : very unlike the low - mass planets in our own solar system . in particular , statistical analyses of the properties of the low - mass exoplanet population reveal that the latter are inconsistent with a completely rocky composition @xcite . instead , comparisons of the observed exoplanets with measured masses and radii to structural models @xcite indicate that the dominant structure is a solid core overlaid with a @xmath31% ( by mass ) h / he envelope . since close - in planets are subject to intense irradiation by high energy stellar photons , which can drive an outflow by heating the upper layers of the atmosphere to close to the escape temperature ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , the evaporation of such h / he envelopes becomes important for the planets evolution ( e.g. * ? ? ? * ; * ? ? ? forward modelling studies suggest that the majority of close - in exoplanets were born with h / he envelopes , but about half have subsequently lost these through evaporation @xcite . in summary , current exoplanet demographics lead us to infer that a dominant mode of low - mass planet formation produces a rocky earth - like core surrounded by a h / he envelope with a mass - fraction of @xmath31% @xcite . this observational result is supported by recent theoretical calculations within the framework of core accretion , which suggest that low - mass cores @xmath9m@xmath6 will acquire envelope mass - fractions of order a few percent @xcite . for example , the scalings of @xcite imply that a 1m@xmath6 core will accrete a h / he envelope mass - fraction of 0.1 - 1% during the gas disk s lifetime . we note that various studies of gas accretion onto earth - like cores find that the results are sensitive to a number of uncertain parameters , such as the opacity and disc properties ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , and , especially , the assumed core accretion rate . thus , these studies suggest a range of initial h / he envelope masses for a given core mass . however , this theoretical prejudice should not blind us to the _ empirical _ fact that current exoplanet data indicate that envelope mass fractions @xmath10% are preferred . for a terrestrial - mass planet , orbiting either a solar - type star or an m dwarf and located within the classical hz ( defined here as the range of orbital separations where an earth - mass planet , with roughly earth - like composition and atmosphere , can harbour liquid surface water ; eg . , see * ? ? ? * ) , a h / he envelope with mass - fraction of order a percent would preclude habitability , by yielding very high surface temperatures and pressures incompatible with liquid water . however , habitable conditions may be achieved if evaporation can reduce the h / he envelope mass - fraction to @xmath11 ( e.g. * ? ? ? * ) ; strip away this primordial envelope entirely so that it is replaced by a tenuous secondary atmosphere , such as those of the solar system terrestrial planets ( e.g. * ? ? ? * ; * ? ? ? * ) ; or separate h and he in planets with a low initial h / he fraction to leave a habitable he atmosphere @xcite . in other words , evaporation may turn a large population of uninhabitable low - mass planets with voluminous h / he envelopes into habitable ones . such an effect is unlikely in the hz of sun - like stars , since their x - ray / uv fluxes over gyr timescales at @xmath31au are too low to remove a massive h / he envelope ( though several works have shown it is possible to remove a _ small _ h / he atmosphere ; e.g. , * ? ? * ; * ? ? ? * ; * ? ? ? m dwarfs , on the other hand , are far more active , and remain so for much longer , than solar - types , leading to much higher xuv fluxes within their hzs over gyr timescales ( e.g. * ? ? ? * ; * ? ? ? * ) . moreover , low - mass planets appear abundant around these red dwarfs , which are themselves the most common stars in the galaxy ; evaporation , if efficient enough , could thus lead to a plethora of habitable planets in m dwarf systems . in this work , therefore , we consider terrestrial - mass planets in the hz of m dwarfs , with an initial h / he envelope mass - fraction of @xmath31% , and investigate whether evaporation over a gyr can remove a sufficient portion of this atmosphere to render the planet habitable at the end . we account for both stellar evolution and the thermal evolution of the planet . in order to make advances in this important area , we also abjure various simplifying and , as it turns out , erroneous assumptions about the evaporative flow made in previous studies of this subject , as described below , in order to obtain a more rigorous and realistic estimate of the mass loss . given the _ kepler _ results above , and the interest in habitable planets , several recent studies have investigated the evaporation of voluminous h / he envelopes around terrestrial - mass solid cores . however , previous studies have typically made two key simplifying assumptions : _ ( i ) _ radiative cooling in the flow is either neglected , or accounted for by assuming a fixed energy efficiency for driving the flow - either in a global or local sense ; and _ ( ii ) _ evaporation is assumed to occur in the hydrodynamic limit at all times . both assumptions are likely to lead to an overestimation of the amount of h / he a planet can lose . first , radiative cooling becomes important when the timescale for a fluid element to advect its heat outwards ( the ` flow timescale ' ) becomes comparable to the timescale for radiative losses to cool that fluid element ( e.g. * ? ? ? this lowers the temperature in the flow , which in turn pushes the sonic point to greater heights ( lower densities ) and thus reduces the mass - loss rate . second , in order to be in the hydrodynamic limit , one requires that the gas remain collisional up to the sonic point ( e.g. , * ? ? ? if this is not satisfied , the flow will collapse and switch to jeans escape , which has a much reduced mass - loss rate ( as we demonstrate later , jeans escape can not remove a significant h / he atmosphere on gyr time - scales , and other non - thermal processes are also unlikely to play a significant role ) . finally , these two points are inter - related : as radiative cooling pushes the sonic point to higher heights and lower densities , it can also trigger the transition to jeans escape earlier than in calculations that do not include cooling . @xcite and @xcite studied the evolution of terrestrial mass cores in the hz of solar - type stars , using evaporation rates calculated in the `` locally energy - limited '' approach , where a fixed fraction of the absorbed photons energy is locally deposited into heat . this fraction is calibrated to detailed calculations by @xcite that explicitly solve the micro - physics of photon absorption ; as such , it is more accurate than the standard `` energy - limited '' formalism , wherein an ad hoc fraction of the total incoming radiative flux goes into heating the gas . however , these calculations still neglect radiative cooling , and as such will overestimate the mass - loss rate when radiative cooling is important . @xcite further argue that if the core accretion rate is very large ( mass doubling times @xmath12 years ) , resulting in a high - entropy bloated atmosphere , evaporation may be able to completely remove initially massive h / he envelopes from low - mass cores and leave behind a potentially habitable planet . however , such large core accretion rates require that the time of planet formation to be fine - tuned to occur just before disc dispersal , otherwise the core would end up with a mass @xmath13 m@xmath6 . with more reasonable accretion rates , they find cores are likely to retain a large fraction of their original envelope . @xcite , expanded on this work to demonstrate that earth - mass planets at 1au around solar - type stars can lose a massive 1% h / he envelope over a gyr only if the star is an unusually fast rotator ( i.e. , unusually active , with x - ray luminosity in the upper 90th percentile of the observed spread in x - ray luminosities in these stars ) . in the majority of cases , i.e. , for solar - type stars with more standard rotation rates and actitivity levels , they find that such planets can only lose h / he envelopes with initial mass - fractions @xmath140.1% . @xcite have recently investigated photoevaporation around m dwarfs . using the standard `` energy - limited '' formalism and further assuming a hydrodynamic flow at all times , they argue that evaporation can completely strip h / he envelopes with a mass - fraction @xmath1% from earth - mass cores in the hzs of m dwarfs . if true , this would imply a potentially vast number of worlds with habitable conditions around m dwarfs . however , the problems noted above with the adopted assumptions call this result into question . specifically , below an envelope mass - fraction of @xmath31% , the radius of the planet is an extremely strong function of the remaining envelope mass - fraction @xcite ; at the same time , the flow timescale increases strongly with decreasing planetary radius ( as the atmosphere falls deeper into the planet s gravity well ) . thus , one expects radiative cooling to increase significantly as the h / he mass - fraction descends below @xmath31% , resulting first in suppressed hydrodynamic mass - loss rates , and then a transition to jeans escape with greatly diminished loss rates . this process can quench mass loss from planets with a hefty h / he envelope still remaining ; indeed , the effect has already been identified at closer separations and higher planetary masses @xcite , and is the origin of the so - called `` evaporation valley '' @xcite . in this paper , therefore , we eschew the assumptions made in earlier studies . instead , we explicitly account for the effect described above , by : _ ( i ) _ smoothly transitioning from the regime where heating is balanced by outflow ( as in the locally energy - limited calculations of * ? ? ? * ) to the regime where heating is balanced by radiative cooling , using a parametrisation calibrated to detailed monte - carlo radiative transfer simulations ; and _ ( ii ) _ using our hydrodynamic models to determine when the transition to jeans escape occurs , triggering a strong suppression in mass - loss rates . in general , with our more appropriate treatment , we find that whether or not a planet can lose enough h / he to be considered habitable depends strongly on the core mass and orbital separation . in particular , the mass loss we derive for @xmath31 cores with initially @xmath31% he / he envelopes in the hz of m dwarfs is much lower than previous estimates , implying that such planets will not be habitable . stellar xuv - driven evaporation is particularly important around m dwarfs because these stars are extremely active : for instance , an average early- to mid - m dwarf remains at saturated levels of coronal activity , with @xmath15/ @xmath16 10@xmath1710@xmath18 , for a few 100myr , compared to only a few@xmath1910myr of saturated activity in an average solar - type star ( e.g. , * ? ? ? consequently , a planet with a given equilibrium temperature receives orders of magnitude more high - energy radiation , integrated over its lifetime , around an m dwarf than around a star like the sun . in this study , we are concerned with the evaporation of low - mass planets . @xcite demonstrated that at high x - ray fluxes , mass - loss from such planets is predominantly x - ray - driven , and euv heating can be neglected . we thus confine ourselves to x - ray - driven flows here . to study the latter , we need a fiducial x - ray spectrum ; we use that of ad leo , an m3.5 dwarf of mass @xmath30.4 , and one of the most active nearby stars . th very strong coronal activity on ad leo arises not because it is anomalous , but simply because it is relatively young compared to most nearby field stars . at a median age of @xmath3100myr ( estimated age range @xmath325300myr ; * ? ? ? * ) , it is still in its saturated phase of activity , with @xmath15/ = 10@xmath20 ( @xmath15 = 7@xmath1910@xmath21ergs@xmath22 and = 9@xmath1910@xmath23ergs@xmath22 ; * ? ? ? * ) . its high levels of activity , combined with its proximity ( @xmath24pc ) , have made ad leo a touchstone for understanding the coronae of m dwarfs . nevertheless , the intrinsic faintness of m dwarfs ( ad leo included : @xmath2510@xmath26 ) means that , even when @xmath15/ is very high , the star is still very faint in x - rays , and it is not yet possible to acquire a full high - quality x - ray spectrum . instead , one usually reconstructs the spectrum using the technique of emission measure distributions ( emds ) , based upon observations of bright emission lines . we use an emd - derived synthetic x - ray spectrum of ad leo supplied by j. sanz - forcada ( pvt.comm.2014 ) . the methodology for constructing it is described in detail by @xcite and @xcite ; this spectrum is also used by @xcite , who show that it agrees very well with current x - ray ( and uv ) data for ad leo . we plot the spectrum in fig.1 , and compare it to a scaled - solar spectrum ( representing a young solar - type star ) ; the comparison solar - like spectrum was calculated in @xcite using the method of @xcite , and was designed to be representative of a young , saturated solar - type star . it is immediately clear that the ad leo spectrum is considerably softer ; thus , since x - ray heating is mainly due to photo - electrons liberated by soft @xmath30.1 - 1kev photons , we expect significantly more heating using our realistic m dwarf spectrum than with a scaled - solar proxy for it ( e.g. , as done in * ? ? ? 0.11kev ) x - ray photons in ad leo relative to the young solar - type star . ] since ad leo is currently still in the saturated regime , we assume it has maintained the same @xmath15/ = 10@xmath20 since birth to the present ( the standard behaviour of saturated low - mass stars , before stellar rotation slows sufficiently for the activity to become unsaturated ) . however , its bolometric luminosity certainly _ has _ evolved with time , as the star has contracted towards the main sequence ; we assume that its follows the theoretical evolutionary track of @xcite for a 0.4 star , and scale @xmath15 accordingly in time . finally , for ease of computation , we assume that ad leo will remain saturated , i.e. , its @xmath15/ will stay unchanged , till 1gyr ( when we terminate our evaporation calculations ) . this is not physically strictly accurate , since a 0.4 star should start spinning down after a few 100myr ( e.g. , * ? ? ? * ) , thus entering the unsaturated regime and evincing decreasing @xmath15/ ; assuming saturation at these late times then formally implies that our mass - loss rates will be overestimated here . however , as we shall see , for the core - mass and h / he envelope mass - fraction regime we study in this paper , our calculations show that the planets are either completely stripped of their envelopes well before 100myr , or have entered the jeans escape regime , with tiny mass - loss rates , by this age . as such , our simplifying assumption of saturated activity beyond a few 100 myr has no discernible impact on our results . in order to derive evaporative mass - loss rates , we must specify the planetary structure . as our starting point , we assume planets with earth - like solid cores , composed of 2/3 rock + 1/3 iron , swathed in a h / he envelope with a mass - fraction ( defined as @xmath27/@xmath28 ) up to @xmath31% . as discussed above , this is consistent with both observations and theoretical calculations of the conditions at birth of `` kepler '' planets @xcite . current theories of planet formation , though , do not strongly constrain the initial entropy ( or equivalently , radius ) of such planets , with both `` hot start '' planets ( those with short initial cooling time scales ) and `` cold start '' ones ( with long initial cooling time scales ) remaining viable . we thus choose initial planetary radii ( entropies ) corresponding to cooling timescales in the range 10@xmath2910@xmath30 yr , to comfortably span the plausible range from `` hot '' to `` cold start '' scenarios ( further discussed in 5 ) . finally , we are interested in the potential habitability , under the effects of evaporation , of low - mass planets in the hz of m dwarfs . the `` classical '' hz is defined as the range of orbital separations around a star where an earth - mass planet with a co@xmath31h@xmath31o n@xmath31 atmosphere can sustain liquid water on the surface ; the inner edge of this zone is ( conservatively ) set by the moist greenhouse effect , and the outer edge ( again , conservatively ) by the maximum greenhouse effect @xcite . the position of the hz of course changes as the star evolves in temperature and luminosity ; what we are really interested in is the fate of planets located within the stable hz of an m dwarf on the main sequence ( ms ) . using the fitting equations supplied by @xcite for the hz boundaries , and the ms values of the stellar temperature and luminosity for a 0.4 dwarf from the @xcite tracks ( @xmath16 3500k , @xmath16 1.86@xmath1910@xmath26 ; the star arrives on the ms after @xmath3500myr ) , yields an inner edge of the classical hz for our m dwarf at 0.15au , and an outer edge at 0.28au . for ease of calculation , and discussions for m dwarfs with masses around 0.4 m@xmath32 , we choose here to parametrise the hz in terms of the blackbody temperature ( @xmath33 ) of the planet instead , which we define as the equilibrium temperature of a planet with zero albedo , orbiting a star of luminosity at a radial separation of @xmath34 : @xmath35 [ /@xmath36^{1/4}$ ] . we set the inner edge of the classical hz around our m dwarf to be at a ms blackbody temperature of @xmath37k , and the outer edge at @xmath38k . these imply radii of @xmath34 = 0.12au and 0.26au respectively on the ms : very close to the actual separations from the @xcite equations . since we are interested in framing the general theory of evaporation driven habitability here , the @xmath31020% difference has no appreciable effect on our conclusions . detailed calculations of the habitability of systems with specific parameters ( i.e. future detected exoplanet systems ) , can wait until a later date . we discuss our theory of x - ray driven evaporation , in both the hydrodynamic and ballistic limits , in 3 ; our numerical implementation of this in 4 ; and our treatment of thermal evolution in 5 . we present our results in 6 , and explore the implications for habitability around m dwarfs in 7 ; our main conclusions are summarised in 8 . our interest lies in low mass planets ; in these , atmospheric heating and evaporation are dominated by x - rays @xcite . it is well known that , in radiative equilibrium , heating due to high energy photons can be described using a relationship between temperature ( @xmath39 ) and ionization parameter ( @xmath40 , where @xmath41 is the received x - ray flux and @xmath42 the number density of particles ) @xcite . @xcite used this principle to calculate solutions to the evaporation problem . however , the precise form of the @xmath43 relation is sensitive to the input x - ray spectrum and the metallicity of the gas . the x - ray spectrum of m dwarfs is quite different from that of solar type stars , being in general much softer in the @xmath44 kev range important for photo - electric heating . we must therefore calculate a new @xmath43 relation , specific to our template m dwarf , to understand the behaviour of the x - ray irradiated planetary atmosphere . we use the radiative transfer code mocassin @xcite to solve for the @xmath43 relation appropriate to the ad leo x - ray spectrum discussed in 2.2 . we irradiate a plane parallel atmosphere , with densities spanning @xmath45 g @xmath8 , the expected range in the evaporative flow . for specificity , we scale the x - ray flux to that expected at a planet orbiting at @xmath37k ( @xmath30.12au ) ; however , since we are concerned with mapping out the @xmath43 profile , rather than with the x - ray flux itself , the exact flux chosen to perform the calculation does not matter very much . the resulting @xmath43 profile is shown in fig.[fig : t - xi ] , along with a functional fit which we adopt in our further calculations . the soft m dwarf x - ray spectrum has a higher fraction of @xmath30.1 - 1kev photons compared to solar - type spectra ( see fig.[fig : spec ] ) ; consequently , the m dwarf produces considerably higher gas temperatures at low ionization parameters ( @xmath46 ) . we first attempt to solve for the evaporative flow by inserting our @xmath43 relation into the semi - analytical methodology of @xcite , which assumes radiative equilibrium in the flow . this , however , leads to an inconsistency : over much of the parameter range of interest , the mechanical luminosity of the resulting evaporative wind becomes comparable to the radiative energy input rate , implying that radiative equilibrium is not a good approximation ; instead , energy losses due to @xmath47 work may dominate over radiative losses much of the time . in hindsight , this is not too surprising , given the higher heating efficiency of our m dwarf x - ray spectrum relative to that in the original @xcite calculations . as such , we must derive a full numerical solution to the radiation - hydrodynamic problem . nevertheless , the fact that , _ when _ radiative equilibrium holds , the gas temperature should equal that given by the @xmath43 relation enables us to simplify the numerics , fully span the pertinent parameter space and , crucially , smoothly transition to the situation where radiative losses _ do _ become important ( which , as discussed in 2.1 , ultimately determines when our hydrodynamic wind is quenched ) . our specific numerical technique for accomplishing this is detailed in 4.1 . to assess the importance of _ non_-hydrodynamic mass loss , we calculate the ballistic ( jeans escape ) mass - loss rate , given by @xmath48 where @xmath49 is the planetary radius , @xmath50 the density at the exobase and @xmath51 the distribution of particle velocities . we assume that the exobase is in local thermodynamic equilibrium with the x - ray irradiation , so that its temperature is specified by our @xmath43 relation ; @xmath51 is then the maxwell - boltzmann distribution at this temperature . the exobase density is @xmath52 where @xmath53 is the scale height of the atmosphere ; @xmath54 the mean molecular weight of the gas particles , set to 1.35 for an atomic solar - abundance h / he mixture ; and @xmath55 the collision cross - section , for which we adopt @xmath56 @xmath57 @xcite . equation ( [ eqn : mdot_jeans ] ) can be solved numerically for the mass - loss rate as a function of core mass . since we expect the jeans escape rates to be low , we assume the extent of the h / he atmosphere is small , and approximate the planetary radius @xmath49 by that of the solid core . the core mass - radius relation is from @xcite , for a core comprising 2/3 rock and 1/3 iron by mass . , @xmath58 , @xmath59 , @xmath60 in 1 gyr . ] the resulting @xmath61 are plotted in fig [ fig : jeans ] as a function of core mass and are compatible with those found previously for highly irradiated terrestrial mass planets ( e.g. * ? ? ? * ; * ? ? ? these mass - loss rates are indeed very low ; in particular , they are significantly below the rates one might expect from hydrodynamic evaporation ( as we will see ) . we find that jeans escape can only remove a h / he envelope mass - fraction ( @xmath62 ) of order 10@xmath63 over a gyr timescale : completely negligible compared to the @xmath31% initial h / he mass - fractions we are concerned with here . thus , in order to strip any significant h / he envelope from solid terrestrial - mass cores in the hz of an m dwarf , efficient hydrodynamic evaporation appears essential . as discussed in 3.1 , we must resort to full numerical hydrodynamic calculations to obtain the evaporation rates for the m dwarf case . furthermore , to avoid the problems faced by @xcite in finding steady - state solutions to the wind problem , we explicitly integrate the time - dependent hydrodynamic equations until a steady - state is reached after several sound - crossing times ( e.g * ? ? ? * ) : in all our calculations , we run the simulations for 30 crossing times to ensure that steady - state is achieved . we employ the zeus hydrodynamical code @xcite , which has been successfully tested and used for planetary evaporation problems in the past ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? our modified zeus algorithm includes heating by x - rays and their radiative transfer ( c.f . * ? ? ? * ; * ? ? ? as noted in 3.1 , our evaporative flow may be in a state far from radiative equilibrium . specifically , we expect the gas to behave as follows . in steady - state , since the flow is expanding and we do not anticipate any heating by shocks , the gas will always be cooler than the temperature at radiative equilibrium ( @xmath64 , given by the @xmath43 relation ) . when the gas is considerably cooler than @xmath64 - due to a large amount of @xmath47 work - the photo - heating rate will greatly exceed the radiative cooling rate , and we may ignore radiative cooling . as the gas temperature approaches @xmath64 , however , radiative cooling will become comparable to the photo - heating rate , and the temperature will asymptote to @xmath64 . with this qualitative picture in mind , we build a simplified thermal model for x - ray heating / radiative cooling , using a `` newtonian - heating '' approach . we begin by defining a _ local _ heating time as @xmath65 , where @xmath66 is the gas internal energy per unit mass at @xmath67 , and @xmath68 is the local heating rate per unit mass due to absorption of high energy photons , given by @xmath69\frac{\sigma_\nu}{\mu}\ , { \rm d}\nu\ ] ] where @xmath70 is the optical depth ( from above ) at a distance @xmath71 from the centre of the planet : @xmath72 here , @xmath73 represents the fraction of the photon energy that is converted into heating rather than used up in ionization . we caution that this is _ not to be confused _ with the efficiency parameter @xmath74 commonly deployed in `` energy - limited '' estimates of the mass - loss rate ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , which ( arbitrarily ) sets how much of the photon energy goes into @xmath47 work instead of being radiatively lost . in our analysis , @xmath73 merely accounts for the energy diverted to ionization processes ; radiative energy losses , on the other hand , are explicitly accounted for by incorporating our derived @xmath43 relationship into the calculations ( via @xmath66 in the energy equation , as described further below ) , and _ not _ via an ad hoc efficiency parameter @xmath74 . integrating a frequency - dependent radiation - hydrodynamics problem is beyond the scope of this work . instead , we use the fact that the heating in our case is dominated by soft x - ray photons , mainly at energies of @xmath31kev @xcite . we can thus approximate our frequency - dependent integral in equation(3 ) by a monochromatic calculation at 1kev , with an absorption cross - section of @xmath75 @xmath57 . heating by @xmath31kev photons is primarily due to secondary photo - electrons ; @xmath76 then depends on the ionization state of the gas , with a value of @xmath30.1 in neutral gas and @xmath30.9 in highly ionized plasma @xcite . while we do not attempt to solve for the full ionization structure of the gas , our mocassin results indicate the gas is partially ionized ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , so we pick an intermediate value of @xmath77 consistent with the ionization fraction . our calculations should not be exceptionally sensitive to our choice , as the thermostatting of the flow towards the @xmath43 relationship will weaken the sensitivity of the flow to the choice of @xmath78 . these considerations enable us to include radiative heating and cooling explicitly in our calculations , by writing the internal energy equation as @xmath79 where @xmath80 is the specific internal energy of a fluid element and @xmath81 is the ratio of specific heats , set to 5/3 for atomic gas . the first term on the r.h.s . of equation(5 ) is the sink in internal energy due to @xmath47 cooling , while the second is the sum of the source and sink terms due to radiative heating and cooling respectively . when @xmath82 , this term is dominated by heating , with negligible cooling by radiation ; when @xmath83 , radiative heating and cooling rates begin to balance each other ( they must be equal when @xmath67 ) ; and finally , when @xmath84 , this term dominates over @xmath47 cooling and thermostats the gas temperature to @xmath67 . equation ( [ eqn : energy_evolve ] ) allows us to satisfy energy conservation , so that we do not overestimate the fraction of incoming x - ray flux that is converted to mechanical luminosity , while also accounting for the burgeoning amount of energy lost as radiation as the gas temperature draws closer to the radiative equilibrium value . this approach is much more satisfactory than the typical assumption of a constant `` heating efficiency '' everywhere ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , since it allows regions to locally account for radiative losses and thus varying the net heating efficiency in both a local and global manner . hence , in addition to a monochromatic ray - tracing calculation for 1kev photons , we solve equation ( [ eqn : energy_evolve ] ) along with the time - dependent equations of hydrodynamics . we note that our model explicitly assumes that @xmath85 , i.e. , that we approach thermal equilibrium from below . this is satisfied here since the flow is expanding ( and hence cooling via @xmath47 work ) everywhere . if this were not true e.g. if there was heating due to compression or shocks then equation ( [ eqn : energy_evolve ] ) would need to be generalised to include a cooling timescale , in addition to the heating timescale @xmath86 . we solve our evaporation problem in 1d along a streamline connecting the star and the planet , within a spherical co - ordinate system ( we thus implicitly assume that the velocity divergence is that of spherically symmetric outflow ) . we account for stellar gravity by using an effective gravitational potential of the form @xmath87 where @xmath88 and @xmath89 are the planetary and stellar masses respectively , @xmath71 is the radial distance from the centre of the planet , and @xmath34 the orbital semi - major axis . we do not however include a contribution from the coriolis force , since it is negligible in setting the mass - loss rates @xcite . while our code solves the hydrodynamic problem over the entire input parameter space , the applicability of the equations of hydrodynamics must be verified _ a posteriori _ , once the steady - state flow solution is obtained . whether a flow is hydrodynamic or ballistic depends on whether the gas particles are collisional or not on a length - scale shorter than the flow - scale ( e.g. * ? ? ? this property is characterised by the knudsen number : @xmath90 where @xmath91 is the mean free path of the gas particles . the relevant flow length - scale for the problem is the pressure scale - height , given by @xmath92 clearly , @xmath93 increases with distance from the planet ( since the particle number density @xmath42 decreases as an exponential function of the pressure scale - height ; note that the exobase is the radius at which @xmath94 ) . at the same time , our hydrodynamic flow solutions are valid only if the gas remains in the hydrodynamic limit ( @xmath95 ) all the way to the sonic surface ; otherwise , the flow loses causal contact with the planet ( and thus the ability to be influenced by upstream conditions ) before the sonic point is reached ( rendering the concept of a sonic point meaningless ) . putting these two facts together , we follow @xcite in defining the evaporation to occur in the hydrodynamic regime if @xmath95 at the sonic point . using the framework described above , we construct a grid of evaporative mass - loss rates as a function of planetary mass , planetary radius and stellar x - ray luminosity , for a specified orbital distance . in order to comfortably cover the desired parameter space , we span planet masses ranging from 0.5 to 20 ; radii ranging from a value corresponding to a density of 10g@xmath8 ( for a given planet mass ) to half a hill radius at the given orbital separation ; and x - ray luminosities ( evenly spaced logarithmically ) ranging from 10@xmath96 to 10@xmath97ergs@xmath22 . this grid is then coupled to the planetary thermal evolution code , to finally yield the detailed evolution of a given planet and its atmosphere . before proceeding to a discussion of thermal evolution ( 5 ) and the final results for individual planets ( 6 ) , however , it is instructive to examine a slice through our mass - loss grid for a fiducial orbit and stellar x - ray luminosity . fig.[fig : mdot_rates ] shows mass - loss rates as a function of planetary mass and radius , for an x - ray flux of 1.23@xmath98 erg s@xmath22 : the irradiation expected from ad leo at @xmath3100myr ( roughly its current age ) at an orbital separation of @xmath99au ( corresponding to @xmath100k , i.e. , the inner edge of the hz at 1gyr ) . over - plotted is the mass - radius relationship for a solid core comprising 2/3 rock and 1/3 iron ( _ dashed magenta line _ ) . a planet with a h / he envelope mass - fraction @xmath141% , for which the radius is dominated by the solid core , will lie only slightly to the right of this line ( @xmath3 factor 2 , @xcite ) , while a planet with a larger envelope mass - fraction , with radius dominated by the envelope , will lie much further to the right . erg s@xmath22 cm@xmath26 . the thick lines show contours of constant knudsen number at the sonic point ( kn@xmath102 ) of 10 ( _ red dot - dashed _ ) , 1.0 ( _ black solid _ ) & 0.1 ( _ blue dotted _ ) . the _ purple thin dashed line _ shows the mass - radius relation for a solid core comprising 2/3 rock and 1/3 iron , taken from @xcite . the _ white thin solid line _ shows a line of constant binding energy to thermal energy in the bolometrically heated atmosphere ; planets to the right of this line have underlying envelopes too loosely bound to be considered tethered to the core ( see text ) . ] also over - plotted are lines of constant knudsen number at the sonic point ( which we will denote by kn@xmath102 ) : kn@xmath102 = 10 ( _ red dot - dash line _ ) , 1 ( _ solid black line _ ) and 0.1 ( _ blue dotted line _ ) . for a fixed planetary mass , we see that kn@xmath102 decreases with increasing radius . thus ( recalling that hydrodynamic escape requires kn@xmath102 @xmath103 1 ) , it is easier to hydrodynamically evaporate a more distended atmosphere ; an intuitive result , since more puffed - up atmospheres have a lower gravitational binding energy for a given planetary mass . indeed , planets with radii to the right of the thin white line in the plot have atmospheres that are so loosely bound that they are unlikely to remain tethered to the planet for more than a myr : our hydrodynamic calculations indicate that the evaporative flows on these planets are driven entirely by stellar bolometric heating rather than by x - ray heating , leading to a rapid `` boil - off '' phase during which most of the atmosphere is lost shortly after disc dispersal ( see * ? ? ? * for a discussion of this process ) , see we are considering planets with relatively low envelope mass fractions this effect is only prominent in one case . finally , we also see that the lines of constant kn@xmath102 are shallower than the mass - radius relationship for our solid core , with the kn@xmath102=1 line intersecting the latter locus at @xmath104 . since hydrodynamic flows can only occur to the right of the kn@xmath102=1 line ( because they require kn@xmath102 @xmath103 1 ) , only planets less massive than @xmath32 can be plausibly stripped entirely of their h / he envelopes via such flows . higher mass planets can still undergo hydrodynamic evaporation as long as their envelopes remain very extended , but such flows will stall once the planetary radius shrinks sufficiently to hit the kn@xmath102=1 limit , which occurs while these planets still retain h / he mass - fractions @xmath1051% . subsequent mass loss only happens ballistically , which can hardly put a dent in such atmospheres over gyr timescales ( as demonstrated in 3.2 ) ; thus , planets more massive than @xmath32 will retain substantial h / he envelopes over their lifetimes . in short , it is easier to completely strip a lower mass planet of its h / he envelope than a higher mass one : not surprising , since escape velocity increases with planetary mass . the threshold mass of @xmath32 identified here , separating the envelope retention / stripping outcomes , is of course quantitatively valid only for the specific x - ray flux adopted in this snapshot plot , and only in the absence of thermal evolution ( which is not accounted for in this plot ) . the true threshold will depend on the integrated history of the x - ray irradiation ( since a higher x - ray flux , expected at earlier times , will push the sonic point deeper into the atmosphere , raising the threshold mass ) , as well as on the thermal evolution of the planet ( since the atmosphere , contracting as it cools , sinks ever deeper into the planet s gravity well , making hydrodynamic escape harder and decreasing the threshold mass ) . by coupling our full mass - loss grid ( of which fig.[fig : mdot_rates ] is only an instantaneous slice at one x - ray flux ) to our thermal evolution code , we will account for both of these effects in 5 and 6 below . nevertheless , the physical effects summarised here will remain qualitatively valid , and only particularly low - mass cores can be stripped entirely by evaporation . in order to fully model the evolution in evaporative mass loss , we must account for how the h / he envelope cools and contracts . we solve the coupled thermal evolution and evaporation problem for a planet with a solid core and h / he envelope using the mesa stellar evolution code adapted to planets @xcite . the method is described at length in @xcite ; here we only summarise the salient inputs to the code . as discussed in 2.2 , the stellar bolometric luminosity ( @xmath106 ) is assumed to follow the lyon evolutionary track @xcite for a 0.4m@xmath107 star ( appropriate for ad leo ) , while the planet s orbital distance is fixed at either @xmath108au or 0.26au , so that the blackbody temperature at the planet s location at 1gyr is @xmath37k or 200k respectively ( corresponding to the inner and outer edges of the classical hz around an m dwarf ; see 2.2 ) . for reference , the evolution in blackbody temperature at 0.12au is shown in fig.[fig : teq_time ] . ) for a planet at @xmath109 au around an m - star . ] the radius of the solid core is assumed to be constant throughout the planet s evolution , and is specified , for a given core mass and our adopted core composition of 2/3 rock + 1/3 iron , by the mass - radius relationship from @xcite . we do account for the thermal content of the core , due to both radioactive decay and thermal heat capacity , by adopting an earth - like value ( see * ? ? ? the effects of both external bolometric irradiation by the star and internal release of thermal energy from the core are accounted for in mesa , using the @xmath110 relationship formulated by @xcite for a gray atmospheric model with both internal and external heat fluxes , where one uses an incoming stream with a frequency set to the peak of the black - body spectrum for the star and an outgoing ir stream in local thermodynamic equilibrium with the upper atmosphere . current planet formation models are unable to stringently constrain the initial thermal properties of newborn planets . consequently , we consider planets with a large range of initial radii ( or , more accurately , initial entropies ) . the latter are parametrised in terms of an initial cooling ( i.e. , kelvin - helmholz ) timescale @xmath111 , defined as the ratio of a planet s initial internal energy to its initial luminosity . naively , one expects this kelvin - helmholz timescale to be of order the `` age '' of the planet at birth , i.e. , the time it takes to form , and thus comparable to the mean lifetime of primordial disks , @xmath310@xmath112 yr ( e.g. * ? ? ? a ten - fold uncertainty in this estimate is not infeasible , due to variations in disc lifetimes ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) or post - formation processes ( e.g. * ? ? ? * ; * ? ? ? ; as such , we use starting models with @xmath111 in the range 10@xmath2910@xmath30 yr , to encompass a plausible maximal difference between `` hot start '' planets ( with short initial cooling times ) and `` cold start '' ones ( with long initial cooling times ) ( c.f . * ; * ? ? ? finally , evaporation is included in our mesa calculations by coupling the code to our grid of newly calculated mass - loss rates as a function of planetary mass , radius and stellar x - ray flux ( 4.3 ) . we present here our results for mass loss ( accounting for both evaporation and thermal evolution ) in the hz of our fiducial m dwarf , for three core masses @xmath113 = 0.8 , 0.9 and 1 at two different orbital separations : @xmath108au ( roughly the inner edge of the hz ) and 0.26au ( roughly the outer edge ) . we investigate initial h / he mass - fractions ( denoted henceforth by @xmath114 ) of @xmath31% and @xmath30.04% ( i.e. , @xmath115 and 4@xmath1910@xmath18 ) . we stop our calculations either after 1gyr of evolution , or once the envelope mass - fraction falls to @xmath59 , the amount that can be efficiently removed by jeans escape on less than gyr timescales ( see [ sec : jeans ] ) . we start by considering the evolution of the atmospheric temperature , density and pressure at the planetary surface ( @xmath116 , @xmath117 and @xmath118 respectively ) , over gyr timescales , in the _ absence _ of evaporation ( i.e. , with evaporation artificially turned off ) . these results constitute a baseline from which to judge whether evaporation will aid habitability or not . we first carry out this analysis for @xmath119 = 300k ( i.e. , at @xmath120au ) . figs.6 , 7 and 8 show our results for @xmath113 = 0.8 , 0.9 and 1.0 respectively , for a h / he envelope with @xmath121 . with evaporation turned off , of course , this mass fraction remains constant over time . in each case , we present results for both a relatively long initial cooling timescale ( @xmath111 ) for the planetary atmosphere , corresponding to a `` cold start '' planet formation scenario , and a relatively short @xmath111 , corresponding to a `` hot start '' ( see 5 ) . we see that , for all three core masses , the final surface temperature after a gyr is @xmath122k , and the final pressure @xmath118 is nearly 10@xmath123 bar ( with variations in @xmath111 making negligible difference to these results ) . these are well above the critical point of water ( @xmath124k , @xmath125 bar ) ; as such , any surface water would exist as a supercritical fluid , not as a liquid ( see phase diagram of water , fig.20 ; constructed from phase transition equations provided by martin chaplin ( pvt.comm.2015 ) ) . hence , as stated earlier , removal of a large fraction of the h / he envelope via hydrodynamic escape , which will reduce @xmath116 and @xmath118 , is _ required _ for habitability when @xmath114 is of order a percent . the degree to which this can happen is the focus of our analysis in 6.1 and 6.2 below , since , as discussed , such relatively high initial mass fractions are commensurate with current data . next , in figs.9 , 10 and 11 , we plot the evolution of surface conditions for a much lower initial atmospheric mass : @xmath114 @xmath16 4@xmath126 . again , in the absence of evaporation , this fraction remains constant . now the surface conditions after a gyr are much more benign , with @xmath127k and @xmath128bar . as the phase diagram in fig.20 shows , surface water should exist as a liquid under these conditions . however , realistic levels of evaporation will in fact remove most or all of this tenuous atmosphere : as we will show below , strong hydrodynamic escape very quickly reduces the atmospheric mass - fraction to @xmath310@xmath63 ; even jeans escape can then remove the remainder on gyr timescales . any potential long - term habitability will then depend on the quantity , properties and evolution of ( outgassed ) secondary atmospheres , similar to the case on earth and mars ( e.g. * ? ? ? as such , for these low initial atmospheric mass fractions , we will plot the evolution of the mass fraction down to 10@xmath63 , to make the point that the primordial h / he atmosphere is likely to be entirely lost , but will abstain from any further discussion of the final surface conditions , which will depend on secondary atmospheres whose modeling is beyond the scope of this paper ( although several authors have begun to investigate this stage theoretically ; e.g. , * ? ? ? * ; * ? ? ? * ) . nonetheless , since secondary atmospheres are at least _ present _ on all the terrestrial planets and massive moons in the solar system , we will consider planets whose h / he envelopes have been completely stripped to still be `` potentially habitable '' . the results for @xmath38k ( @xmath129au ) without evaporation ( not plotted ) are similar : while the surface temperatures after a gyr are now somewhat lower , @xmath116 and @xmath118 are still far too high for liquid surface water when @xmath121 ; for @xmath130 , the surface conditions are again conducive to liquid water , but with the same caveats as above . we now turn the evaporation on , and once again examine the results for core masses of 0.8 , 0.9 and 1.0 at @xmath37k . * 0.8 : * fig.12 shows the evolution of the atmospheric mass - fraction and planetary radius for @xmath28 = 0.8 , for @xmath132 ( top panels ) and 4@xmath126 ( bottom panels ) . in both cases , evaporation strips the atmosphere down to a mass fraction of 10@xmath63 extremely rapidly , in @xmath1410myr . this steep decline is due to runaway evaporation , wherein the cooling timescale exceeds the evaporation timescale : the atmosphere can not contract quickly enough to avoid evaporation by falling deeper into the planet s gravitational potential well ( c.f . * ? ? ? note that the runaway occurs much earlier than the median age of 10@xmath133myr estimated for ad leo ( and for other early to mid - m field dwarfs evincing similar saturated x - ray activity ; see 2.2 ) ; consequently , this evaporation result is unaffected by our simplistic assumption that activity remains saturated beyond a few 100 myr up to a gyr ( instead of declining at later times ) . for a 0.8 core , therefore , we expect any primordial h / he atmosphere with initial mass fraction @xmath141% to be completely lost : eroded down to 10@xmath63 in @xmath1410myr by runaway hydrodynamic evaporation , as shown , and the minuscule remainder efficiently peeled off by continuing hydrodynamic and/or jeans escape over a less than gyr timescale . * 0.9 : * fig.13 shows the evolution in atmospheric mass fraction and planetary radius for @xmath28 = 0.9 . for @xmath132 , we see that the `` hot '' and `` cold '' start cases behave very differently . the atmosphere in the former is initially relatively bloated , resulting in a large fraction of it being blown off almost instantaneously when evaporation is switched on ( a phenomenon termed `` boil - off '' by owen & wu 2015 ; see 4.3 ) . for reasons of clarity , this initial boil - off , which occurs over the first few time - steps of our calculations ( in @xmath141.5 myr ) , is masked out in fig.13 and subsequent figures . essentially one can not build stable models with envelope mass fractions and cooling times as short as desired they are not globally thermodynamically stable in the presence of mass - loss . specifically , in the pressure confining environment of the parent disc the planet can acquire a @xmath10% envelope , but once this pressure confinement is removed such a massive envelope can not be gravitationally bound to the planet ( see owen & wu , 2015 ) . therefore , the bolometrically driven mass - loss allows the planet to readjust to a stable state that is in quasi thermodynamic equilibrium . the plots trace the evolution of the planet after boil - off ends . the remainder then undergoes runaway hydrodynamic escape , reducing the mass fraction to 10@xmath63 in @xmath310myr . for a `` cold start '' , on the other hand , the initial atmosphere is much more compact , and the pace of hydrodynamic evaporation is thus far more leisurely ; by @xmath330myr , evaporation is too slight to make any significant difference , and the atmospheric mass fraction settles down at @xmath310@xmath17 . jeans escape , which only becomes efficient at mass fractions @xmath1410@xmath63 , has no discernible effect on the surviving atmosphere . note as well that the @xmath330myr it takes for the mass fraction to reach approximate steady - state is at the lower limit of the 10@xmath133myr age estimate for ad leo ; our result is thus minimally affected by our naive assumption that activity continues to be saturated beyond a few 100 myr instead of declining at such late times . finally , cores with @xmath13410@xmath18 are evaporated down to 10@xmath63 in only a few myr , regardless of initial cooling timescale , just as in the 0.8 case . the evolution of surface conditions for the @xmath132 case is shown in fig.14 . for a `` cold start '' 0.9 core , we see that @xmath116 and @xmath118 after a gyr are @xmath3 530k and 630bar respectively ( note that the slight continuing evolution of these quantities beyond @xmath3100myr is predominantly due to ongoing thermal evolution of the planet , and not any appreciable evaporative mass loss ) . this [ @xmath116 , @xmath118 ] combination is conducive to liquid surface water ( fig.20 ) , and within the range found in deep - ocean hydrothermals vents on earth . fig.14 also shows the `` hot start '' core with the same @xmath132 for comparison . the [ @xmath116,@xmath118 ] @xmath3 [ 350k , 10bar ] achieved at the end of our calculations is of no particular significance , since we expect the tiny remaining atmospheric mass fraction of 10@xmath63 to be completely removed subsequently by even jeans escape ; we only note that the surface temperature is beginning to flatten out since it can not fall below the equilibrium temperature ( with albedo assumed to be zero ) of 300k at this orbital radius , while the pressure continues to plummet , which will eventually lead to any liquid surface water boiling into vapour . of more physical interest is the fact that @xmath116 in the `` hot start '' case is _ lower _ than in the `` cold start '' one , once the rapid initial boil - off stage ( which is not plotted in these figures ) has ended . this is because hydrodynamic escape requires mass continuity at the sonic point ; when a substantial fraction of the atmosphere is blown off from the top , therefore , the underlying layers swiftly expand to take its place , and the upward advection of heat in this process causes the surface to cool ( see discussion in * ? ? ? * ) . in summary , at @xmath37k , we expect liquid water to survive on gyr timescales on the surface of a `` cold start '' 0.9 core with @xmath132 ; the partial evaporation of the h / he atmosphere engenders habitable conditions in this case . a `` hot start '' 0.9 core with the same @xmath114 , or a core with much smaller @xmath114 and either a `` hot '' or a `` cold '' start , however , is unlikely to retain any of its primordial h / he atmosphere ; habitability conditions in this case will depend on any secondary atmosphere that arises . therefore , a 0.9 m@xmath6 core appears to be at the _ transition _ between a planet becoming habitable due to evaporation , and remaining uninhabitable due to retention of a significant h / he envelope . * 1.0 : * finally , fig.15 shows the the evolution in atmospheric mass fraction and planetary radius for @xmath28 = 1.0 . there is relatively little evaporation when @xmath132 , with both `` hot '' and `` cold start '' cores equilibrating at a mass fraction of @xmath32@xmath1910@xmath17 ( there is some initial boil - off in the `` hot start '' case , but hardly enough to initiate runaway evaporation ; the planet s gravity is too strong ) . when @xmath13410@xmath18 , on the other hand , runaway evaporation reduces the fraction to 10@xmath63 in just a few myr , as in the 0.8 and 0.9 cases . in fig.16 we plot the evolution of surface conditions for @xmath132 . both `` hot '' and `` cold '' start cores end up with very similar surface temperatures and pressures after a gyr , with [ @xmath116 , @xmath118 ] @xmath3 [ 800k , 2@xmath1910@xmath136bar ] . fig.20 demonstrates that these conditions are too extreme for liquid water ; any surface water here can only exist as a supercritical fluid . we now investigate the effects of hydrodynamic evaporation at @xmath138k . the single case of a 0.8 core will serve to illustrate the main results here . * 0.8 : * we plot the evolution in atmospheric mass fraction and planetary radius for this core mass in fig.17 . there is hardly any evaporation for either @xmath132 or 4@xmath126 : in both cases , the final mass fraction after a gyr is very similar to the initial value . figs.18 and 19 show the evolution in surface conditions for these two initial mass fractions . for @xmath114 of 10@xmath26 , the final [ @xmath116 , @xmath118 ] after a gyr is @xmath3 [ 800k , 4@xmath1910@xmath136bar ] ; as the phase diagram in fig.20 indicates , surface water will exist as a supercritical fluid instead of a liquid under these conditions . for @xmath114 of 4@xmath1910@xmath139 , on the other hand , the equilibrium [ @xmath116 , @xmath118 ] after a gyr is @xmath3 [ 400k , 280bar ] , which _ is _ conducive to liquid water . we can easily extrapolate from these results , and the preceding ones , to deduce the implications for 0.9 and 1.0 cores at this radial separation . first , surface temperature and pressure always increase with core mass , for a given @xmath114 and orbital radius . thus , since a 0.8 core with an initial h / he atmospheric mass fraction of a percent can not harbour liquid water at @xmath38k , higher mass cores with the same initial mass fraction can not either . second , in the _ absence _ of evaporation , @xmath116 and @xmath118 are very similar for 0.81.0 cores at any fixed @xmath114 and orbital radius ( e.g. , see the no - evaporation cases plotted in fig.20 for @xmath37k : the @xmath118 for 0.81.0 cores are nearly identical , and their @xmath116 vary by @xmath140100k , for both @xmath114 of 10@xmath26 and 4@xmath1910@xmath18 ) . we found above that there is hardly any evaporation in the 0.8 case even with @xmath13410@xmath18 at this orbital radius ; therefore , since the evaporative rate decreases with increasing core mass ( as the gravitational potential well deepens ) , we can invoke the no - evaporation results here for the entire range 0.81.0 . this in turn implies that 0.9 and 1.0 cores with @xmath13410@xmath18 will have [ @xmath116 , @xmath118 ] very close to that of the 0.8 core with the same initial mass fraction , and can thus also harbour liquid surface water on gyr timescales . these inferences are indeed what our more detailed calculations show . = 300k ( inner edge of hz ) , with a core mass @xmath113 = 0.8m@xmath141 and an initial h / he envelope mass - fraction of @xmath142 @xmath16 10@xmath26 ( the latter remains constant with time in these ` no - evaporation ' calculations ) . the _ dashed vertical line _ shows the median age of ad leo ( @xmath3100myr ) . the _ red curve _ is for a `` hot - start '' model , and the _ blue curve _ for a `` cold - start '' model ; @xmath111 for each is noted in the left panel . ] , but for @xmath113 = 0.9m@xmath141 . ] , but for @xmath113 = 1.0m@xmath141 . ] , for the same core mass @xmath113 = 0.8m@xmath141 , but now for an initial h / he envelope mass - fraction of @xmath142 @xmath16 4@xmath1910@xmath18 . ] , but for @xmath113 = 0.9m@xmath141 . ] , but for @xmath113 = 1.0m@xmath141 . ] = 300k ( inner edge of hz ) , with a core mass @xmath113 = 0.8m@xmath6 . the top row is for planets with an initial h / he envelope mass - fraction of @xmath142 @xmath16 10@xmath26 , while the bottom row is for @xmath142 @xmath16 4@xmath1910@xmath18 . the evolution of the `` hot - start '' and `` cold - start '' models is essentially indistiguishable in the bottom row . ] , but for @xmath113 = 0.9m@xmath6 . note that the `` hot - start '' model ( _ red curve _ ) , for @xmath142 @xmath16 10@xmath26 ( top panels ) , undergoes `` boil - off '' @xcite at very early times , resulting in an almost instantaneous decrease in the initial envelope mass and planetary radius compared to the `` cold - start '' model ( the initial boil - off phase is not plotted for clarity ) - see text . ] ( i.e. , planets at @xmath119 = 300k , with @xmath113 = 0.9m@xmath6 and @xmath142 @xmath16 10@xmath26 ) . ] , but for @xmath113 = 1.0m@xmath6 . ] ( i.e. , planets at @xmath119 = 300k , with @xmath113 = 1.0m@xmath6 and @xmath142 @xmath16 10@xmath26 ) . ] .[fig:1.0e_300_0.01 ] , with the same core mass @xmath113 = 0.8m@xmath141 , but now for a radial location corresponding to @xmath119 = 200k ( i.e. , outer edge of hz ) . ] ( i.e. , planets at @xmath119 = 200k , with @xmath113 = 0.8m@xmath6 and @xmath142 @xmath16 10@xmath26 ) . ] ( i.e. , planets at @xmath119 = 200k , with @xmath113 = 0.8m@xmath6 and @xmath142 @xmath16 4@xmath1910@xmath18 ) . ] the central result of the preceding analysis is that solid cores made of 2/3 rock + 1/3 iron , with mass @xmath41 , can not lose enough of their envelopes to become habitable in the classical hz of m dwarfs , if they are born with significant h / he envelopes . the fundamental reason is that hydrodynamic escape is quenched in such planets while a sizeable portion of the h / he envelope still lingers , and subsequent jeans escape which can at best extract a fractional mass of @xmath310@xmath63 over a gyr is too feeble to remove this remainder . only cores @xmath1401 ( with the precise limiting mass depending on where in the hz one is ) can undergo hydrodynamic escape all the way down to their surface in spite of having initial h / he mass - fractions of @xmath31% , and may thus be habitable ( either due to a tenuous remnant h / he envelope , or a similarly wispy secondary atmosphere ) in the classical hz of m dwarfs . to summarise : _ kepler _ data imply that small ( terrestrial - size ) planets are ubiquitous at small orbital separations , with nearly every star ( statistically speaking ) hosting one such `` kepler '' planet ; around m dwarfs , a sizeable fraction of these close - in planets reside within the classical hz ( see also further below ) . concurrently , both observations and theory indicate that such planets are usually born with h / he envelopes with a mass - fraction @xmath41% , and our calculations imply that planets with cores comprising 2/3 rock and 1/3 iron , with core masses @xmath41 and initial h / he envelope mass - fractions @xmath41% , can not lose enough of their envelopes to be habitable within the classical hz of m dwarfs . with these considerations , we propose three possible classes of solid - core habitable planets around m dwarfs : _ ( 1 ) sub - earth mass rock / iron - core planets within the classical hz _ : we found above that sub - earth mass rock / iron cores with mass @xmath140.9 near the inner edge of the classical hz of m dwarfs , and @xmath1400.8 near the outer edge can be stripped of @xmath31% initial h / he envelopes within a gyr . such @xmath14venus - mass planets may therefore be habitable within the classical hz of m dwarfs , if they acquire secondary atmospheres like the solar system terrestrial planets ( the hz results of @xcite can then be applied to such planets ) . _ ( 2 ) sub - earth to super - earth mass ice - core planets within the classical hz _ : our discussion so far has focussed on rock / iron cores . naively , the results from _ ( 1 ) _ above should apply to ice cores as well ( but see below ) , except for a higher threshold mass ( the limiting planet mass below which @xmath31% initial h / he envelopes will be stripped ) . this is easily seen by examining fig.4 , where the threshold mass is defined as the intersection of the kn@xmath102=1 line and the locus of a core with fixed density . for a given core mass , an ice core , with a density @xmath310 times smaller than a rock / iron one , will have the latter locus shifted to the right ( i.e. , to a larger radius ) by a factor of 10@xmath1432 , resulting in a threshold mass larger by a factor of @xmath32 as well from fig.4 . thus , to zeroeth order , one expects ice - core planets , with a somewhat ( factor of @xmath32 ) larger limiting mass than rock / iron core ones , to also be habitable within the classical hz of m dwarfs . however , a correct analysis of evaporation here must also account for the opacity of steam ( liberated from the core s surface ) in addition to h / he , which may strongly affect the outcome ( since oxygen in this case within h@xmath31o molecules is a strong x - ray absorber ) . this will be the subject of future work . note also that icy cores can only form beyond the ice - line , situated at @xmath31.5au for a @xmath30.4 m dwarf at an age of order a myrau ; with a stellar bolometric luminosity of @xmath144@xmath145 for a 0.4 star at an age of 1.5myr , from the evolutionary tracks of @xcite ] . thus ice - cores must migrate inwards for the above discussion to be pertinent . _ ( 3 ) planets formed after gas - disk dispersal _ : finally , our entire analysis in this paper is for planets that form while a significant amount of gas still remains in the surrounding primordial disk , allowing them to accrete relatively massive h / he envelopes ; _ kepler _ data , as discussed , imply that such planets are ubiquitous . however , the terrestrial planets in our own solar system evince signatures of having coalesced after the gas disk dispersed ( as the radiometricly determined age of the earth @xcite is estimated to be @xmath146 myr younger than that of the solar - system , e.g. @xcite , well after the gas disc would have dispersed ) . if solid - core planets around m dwarfs can form in a similar fashion , then they might well be habitable in the classical hz of m dwarfs ( and slightly interior / exterior to it depending on planetary mass ) ; the work by @xcite , examining solely secondary atmospheres , is more directly applicable in that case than this paper . we note that there is , prima facie , another interesting possibility . while we find that rock / iron cores @xmath41 with @xmath31% initial h / he envelopes retain too large a fraction of these envelopes to be habitable within the classical hz of m dwarfs , they _ will _ be stripped of such atmospheres if their orbits are much smaller . simultaneously , @xcite show that the inner boundary of the hz ( set by either the moist greenhouse or runaway greehouse effect ) for such _ super_-earth mass planets will be closer to the star than for earth - mass ones ( where the latter are used , by definition , to calculate the classical hz limits ) . therefore , one might expect an overlap in space , interior to the classical hz , between where a super - earth of given mass can be stripped of a @xmath31% h / he envelope , and where it will be habitable with a secondary atmosphere before greenhouse effects become overwhelming ; super - earths would then be habitable within this overlap region . unfortunately , such an overlap is implausible . for solar - type stars , @xcite show that the inner boundary of the hz moves inwards from 0.99au for a 1 planet to 0.94au for a 10 one ; i.e. , only a very small shift of 0.05au for a factor of 10 increase in planetary mass . the effect should be similar around m dwarfs . this tiny decrease in orbital separation will have negligible impact on the incident x - ray flux and thus on the mass - loss rates ; as such , we do not expect any hz for super - earths interior to the classical hz around m dwarfs . conversely , for sub - earth mass planets , the inner boundary of the hz moves slightly _ outwards _ , as @xcite show ; thus , we do not expect any hz for these planets interior to the classical hz either . there are two main ingredients that determine the actual frequency of habitable planets around m dwarfs . the first , as we have shown , is a rigorous treatment of evaporation , using hydrodynamic models that account for radiative cooling as well as the transition to jeans escape , and including the thermal evolution of the planet . a simplistic `` energy - limited '' formalism , that moreover assumes that the escape is always hydrodynamic ( e.g. , @xcite for m dwarfs ) , leads to a gross overestimation of evaporative mass - loss rates , and thereby an overly optimistic appraisal of the fraction of planets that can be rendered habitable by stripping their primordial h / he envelopes . our results are in direct conflict with those of @xcite . the latter authors calculations suggest that in the majority of cases , m dwarfs can completely strip envelopes with mass fraction @xmath1410% ( and even larger ones in some cases ) at the inner edge of the hz , and mass fractions @xmath141% at the outer edge of the hz , from 12 m@xmath6 cores . in other words , they conclude that evaporation can generate a plethora of potentially habitable earth - mass planets around m - dwarfs . our results indicate that this is not the case : it is the simplifications made by @xcite thet lead them to conclude otherwise . the main reason is their choice of a constant efficiency energy - limited mass - loss prescription . as we have discussed throughout this work , at late times when the xuv flux declines and the envelope mass fraction drops below @xmath31% ( i.e. , when the planet s radius shrinks rapidly with decreasing envelope mass ) , the efficiency of hydrodynamic mass - loss decreases significantly ( c.f . * ) ; furthermore , the flow transitions to non - hydrodynamic ( jeans escape ) at late times . neglecting these effects , @xcite find that their planets can undergo runaway mass - loss at ages @xmath147 gyr ; including these effects , however , we find that a @xmath31m@xmath6 planet can only ever undergo runaway mass - loss at early times ( @xmath148 gyr ) , when the flux is high enough to permit hydrodynamic evaporation . in fact , if we adopt an energy - limited mass - loss rate with no hydrodynamic cut - off at low fluxes , we indeed recover the runaway mass - loss at late times found by @xcite . this is demonstrated in fig . [ fig : el_compare ] , where we compare the results of our calculations ( solid line ) to those for the energy - limited case with efficiencies ( @xmath74 ) of 5% ( dot - dashed ) , 10% ( dashed ) & 25% ( dotted ) ; note that @xcite s default choice is @xmath74 = 30% . the plot shows that , while our planet loses negligible mass at late times due to the hydrodynamic cut - off , the energy - limited cases can undergo unphysical runaway mass - loss at these ages . we note that while an efficiency of @xmath149 produces a good match to the final envelope mass - fraction at 10 gyr in this _ particular _ example , this is _ not _ a generally applicable result : indeed , the plot shows that the 10% case ( dashed line ) is also starting to enter a run - away phase at the end of the calculation . in general , there is no single efficiency value that can mimic the trends in evaporation in our results , especially the hydrodynamic cut - off , which turns out to be extremely important for evaporation and the question of potential habitability . using an energy - limited formalism to determine evaporation rates when the flow is close to transitioning from hydrodynamic to non - hydrodynamic can lead to rather poor predictions . finally , we note that our prediction that planets @xmath41 will retain most of their initial voluminous h / he envelopes , in the classical hz of m dwarfs , should be directly testable with tess , which will target a large number of m dwarfs . planet at the inner edge of the hz . the _ solid _ curve shows the evolution with the mass - loss rates calculated in this work , while the _ dot - dashed _ , _ dashed _ and _ dotted _ curves show mass - loss rates calculated in the `` energy - limited '' framework with efficiencies of @xmath74 = 5 , 10 and 25% respectively . note that we have assumed that the x - ray flux remains saturated over the entire evolution : even in this extreme case , the flow still becomes non - hydrodynamic ( ballistic ) at late times in our model , preventing runanway mass loss ; this is not true of the energy - limited cases . ] the second ingredient that controls the actual prevalence of habitable planets around m dwarfs is the frequency of their planets as a function of planetary mass , at the small end of the planetary spectrum . a recent careful analysis by @xcite indicates that the occurrence of m dwarf planets continues to increase to planetary radii below 1 , without any strong evidence for a downturn in the planet radius function that was suggested by previous studies . if this is true , then it bodes well for the frequency of potentially habitable planets orbiting m dwarfs , since the smallest ( lowest - mass ) planets will be most easily stripped of their initial h / he envelopes . indeed , applying their improved calculations to previous work by @xcite and @xcite , @xcite already show that the frequency of earth - size planets in the hz of m dwarfs calculated by @xcite to be @xmath30.15 planets per star increases to @xmath30.250.8 , a remarkable fraction . if this frequency continues to increase towards smaller planets , as @xcite suggest , then evaporation can indeed create a bonanza of potentially habitable planets around these cool red dwarfs . however , the latter results are based on a quite limited sample ; they need to be verified and sharpened with more data from the k2 mission , as well as upcoming ones such as tess . our key results are as follow : \1 ) jeans escape can only remove a very small h / he mass - fraction from low - mass exoplanets in the hz of m dwarfs ( @xmath310@xmath63 over a gyr ) : hydrodynamic evaporation is essential for removing significant amounts of h / he . \2 ) previously derived mass - loss rates , based on the `` energy - limited '' ( or `` locally energy - limited '' ) formalism , and assuming moreover that escape is always in the hydrodynamic limit , are significant overestimations . our improved model , accounting for both radiative losses and the transition from hydrodynamic to jeans escape , and including the thermal evolution of the planet , yields much lower mass loss rates . \3 ) in particular , for cores made of 2/3 rock + 1/3 iron , we find that only sub - earth mass cores @xmath140.9 near the inner edge of the classical hz of m dwarfs , and @xmath1400.8 near the outer edge can be stripped of @xmath31% initial h / he envelopes . rock / iron cores with mass @xmath41 , and born with h / he envelope mass - fractions @xmath41% , can not lose sufficient envelope mass to become habitable within the classical hz of m dwarfs . our prediction that @xmath41 cores in the hz of m dwarfs should still be shrouded by their voluminous natal h / he envelopes over gyr timescales will be directly testable with tess . \4 ) we propose three classes of potentially habitable planets in the hz of m dwarfs : _ ( i ) _ planets with sub - earth mass rock / iron cores : since they can be stripped of 1% h / he envelopes within a gyr , they may be habitable if they can acquire suitable secondary atmospheres , like solar system terrestrial planets ; _ ( ii ) _ planets with ice - cores , with an upper limit to the core mass somewhat higher than the sub - earth mass for rock / iron cores , may also be able to lose their primordial h / he envelopes and thus become habitable ; however , steam opacity must be included in the evaporation calculations to verify this possibility ; and _ ( iii ) _ planets ( possibly like the terrestrial ones in the solar system ) that form after gas disk dispersal , and thus have only tenuous secondary atmospheres , with very little primordial h / he . we thank the referee for an instructive report . we are grateful to barbara ercolano , ren heller and renyu hu form comments on the manuscript . we are indebted to jorge sanz - forcada for supplying the synthetic xuv spectrum of ad leo , which enabled a more realistic analysis of evaporation around m dwarfs than possible with the scaled - solar spectra used in previous work , and for comments on the manuscript . we are also very grateful to martin chaplin for providing the equations for the phase diagram of water . jeo acknowledges support provided by nasa through hubble fellowship grant hst - hf2 - 51346.001-a awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc . , for nasa , under contract nas 5 - 26555 . sm acknowledges the support of stfc - uk consolidated grant st / k001051/1 .	the ubiquity of m dwarfs , combined with the relative ease of detecting terrestrial - mass planets around them , has made them prime targets for finding and characterising planets in the `` habitable zone '' ( hz ) . however , _ kepler _ finds that terrestrial - mass exoplanets are often born with voluminous h / he envelopes , comprising mass - fractions ( @xmath0 ) @xmath1% . if these planets retain such envelopes over gyr timescales , they will not be `` habitable '' even within the hz . given the strong x - ray / uv fluxes of m dwarfs , we study whether sufficient envelope mass can be photoevaporated away for these planets to become habitable . we improve upon previous work by using hydrodynamic models that account for radiative cooling as well as the transition from hydrodynamic to ballistic escape . adopting a template active m dwarf xuv spectrum , including stellar evolution , and considering both evaporation and thermal evolution , we show that : _ ( 1 ) _ the mass - loss is ( considerably ) lower than previous estimates that use an `` energy - limited '' formalism and ignore the transition to jeans escape ; _ ( 2 ) _ at the inner edge of the hz , planets with core mass @xmath2 can lose enough h / he to become habitable if their initial envelope mass - fraction is @xmath31% ; _ ( 3 ) _ at the outer edge of the hz , evaporation can not remove a @xmath31% h / he envelope even from cores down to 0.8 . thus , if planets form with bulky h / he envelopes , only those with low - mass cores may eventually be habitable . cores @xmath4 1 , with @xmath41% natal h / he envelopes , will _ not _ be habitable in the hz of m dwarfs . ps . p_surf 1t_bb , ms
networks of coupled oscillators have been extensively studied for many years , owing to their wide applicability in physics , chemistry , and biology . as examples we mention laser arrays , josephson junctions , populations of fireflies , etc . @xcite . the phase - only models have proved to provide useful models for systems with weak coupling . the best known model of this type is the kuramoto model in which the oscillators are described by phase variables @xmath0 and coupled to others through a sinusoidal function @xcite . these models exhibit a transition to collective synchronization as the coupling strength increases , a process that has been described as a phase transition . a general form of these systems is as follows : @xmath1 here @xmath2 is the natural frequency of oscillator @xmath3 , @xmath4 represents the coupling between oscillators @xmath3 and @xmath5 , @xmath6 is a phase lag and @xmath7 is the overall coupling strength . a general treatment of this system is not easy , and two types of simplifications are commonly used . one of them is to assume global coupling among the oscillators , and assign the natural frequencies @xmath2 randomly and independently from some prespecified distribution @xcite . the second tractable case arises when all the oscillators are assumed to be identical , and the coupling @xmath4 is taken to be local , eg . , nearest - neighbor coupling @xcite . the intermediate case of nonlocal coupling is harder but the phenomena described by the resulting model are much richer . in this paper we suppose that the oscillators are arranged on a ring . in the continuum limit @xmath8 the system is described by the nonlocal equation @xmath9\,dy\label{phase_eq}\ ] ] for the phase distribution @xmath10 . when the oscillators are identical ( @xmath11 is a constant ) and @xmath12 this system admits a new type of state in which a fraction of the oscillators oscillate coherently ( i.e. , in phase ) while the phases of the remaining oscillators remain incoherent @xcite . in this paper we think of this state , nowadays called a _ chimera _ state @xcite , as a localized structure embedded in a `` turbulent '' background . subsequent studies of this unexpected state with different coupling functions @xmath13 have identified a variety of different one - cluster and multi - cluster chimera states @xcite , consisting of clusters or groups of adjacent oscillators oscillating in phase with a common frequency . the clusters are almost stationary in space , although their position ( and width ) fluctuates under the influence of the incoherent oscillators on either side . recently , a new type of chimera state has been discovered , a traveling chimera state @xcite . in this state the leading edge plays the role of a synchronization front , which kicks oscillators into synchrony with the oscillators behind it , while the trailing front kicks oscillators out of synchrony ; these two fronts travel with the same speed , forming a bound state . it is natural to ask whether these states persist in the presence of spatial inhomogeneity , in particular , in the presence of spatial inhomogeneity in the natural frequency distribution ( @xmath14 ) . in @xcite laing demonstrated the robustness of the chimera state with respect to inhomogeneity in a two population model , which can be considered to be the simplest model exhibiting chimera states @xcite . however , in this model there is no spatial structure to either population . consequently , we focus in this paper on a ring of adjacent oscillators with a prescribed but nonuniform frequency distribution @xmath15 , where the continuous variable @xmath16 ( @xmath17 ) represents position along the ring . to be specific , two classes of inhomogeneity are considered , a bump inhomogeneity @xmath18 , @xmath19 , and a periodic inhomogeneity @xmath20 , where @xmath21 is a positive integer . in each case we follow @xcite and study the coupling functions @xmath22,\nonumber & & g^{(2)}_n(x)\equiv \cos(nx)+\cos[(n+1)x],\nonumber\end{aligned}\ ] ] where @xmath23 is an arbitrary positive integer . these choices are motivated by biological systems in which coupling between nearby oscillators is often attractive while that between distant oscillators may be repelling @xcite . there are several advantages to the use of these two types of coupling . the first is that these couplings allow us to obtain chimera states with random initial conditions . the second is that we can identify a large variety of new states for suitable parameter values , including ( a ) splay states , ( b ) stationary multi - cluster states with evenly distributed coherent clusters , ( c ) stationary multi - cluster states with unevenly distributed clusters , ( d ) a fully coherent state traveling with a constant speed , and ( e ) a single cluster traveling chimera state @xcite . in this paper , we analyze the effect of the two types of inhomogeneity in @xmath11 on each of these states and describe the results in terms of the parameters @xmath24 and @xmath25 or @xmath21 describing the strength and length scale of the inhomogeneity , while varying the parameter @xmath26 representing the phase lag @xmath6 . in section [ eff_eq ] , we briefly review the notion of a local order parameter for studying chimera states and introduce the self - consistency equation for this quantity . in sections [ rot ] , [ trav_coh ] and [ trav ] we investigate , respectively , the effect of inhomogeneity on rotating states ( including splay states and stationary chimera states ) , traveling coherent states and the traveling chimera state . we conclude in section [ conclusion ] with a brief summary of the results and directions for future research equation ( [ phase_eq ] ) is widely used in studies of chimera states . an equivalent description can be obtained by constructing an equation for the local order parameter @xmath27 defined as the local spatial average of @xmath28 $ ] , @xmath29 the evolution equation for @xmath30 then takes the form @xcite @xmath31 where @xmath32(x , t)$ ] and @xmath33 is a compact linear operator defined via the relation @xmath34(x , t)\equiv\int_{-\pi}^{\pi}g(x - y)u(y , t)\,dy.\ ] ] a derivation of eq . ( [ effective_eq ] ) based on the ott antonsen ansatz @xcite is given in the appendix . equation ( [ effective_eq ] ) can also be obtained directly from eq . ( [ phase_eq ] ) using the change of variable @xmath35.\label{changeofvar}\ ] ] an important class of solutions of eq . ( [ phase_eq ] ) consists of stationary rotating solutions , i.e. , states of the form @xmath36 whose common frequency @xmath37 satisfies the nonlinear eigenvalue relation @xmath38\tilde{z}+\frac{1}{2}\left[e^{-i\alpha}\tilde{z}(x)-\tilde{z}^2e^{i\alpha}\tilde{z}^*(x)\right ] = 0.\label{tw1}\ ] ] here @xmath39 describes the spatial profile of the rotating solution and @xmath40 $ ] . solving eq . ( [ tw1 ] ) as a quadratic equation in @xmath41 we obtain @xmath42 the function @xmath43 is chosen to be @xmath44^{1/2}$ ] when @xmath45 and @xmath46^{1/2}$ ] when @xmath47 . this choice is dictated by stability considerations , and in particular the requirement that the essential spectrum of the linearization about the rotating solution is either stable or neutrally stable @xcite . the coherent ( incoherent ) region corresponds to the subdomain of @xmath48 $ ] where @xmath49 falls below ( above ) @xmath50 . substitution of expression ( [ zx ] ) into the definition of @xmath51 now leads to the self - consistency relation @xmath52 here the bracket @xmath53 is defined as the integral over the interval @xmath54 $ ] , i.e. , @xmath55 in the following we write @xmath56 and refer to @xmath57 and @xmath58 as the amplitude and phase of the complex order parameter @xmath51 . as shown in @xcite , eq . ( [ phase_eq ] ) with constant natural frequency @xmath11 exhibits both stationary rotating solutions ( splay states and stationary chimera states ) and traveling solutions ( traveling coherent states and traveling chimera states ) for suitable coupling functions @xmath13 . in this section , we investigate the effect of spatial inhomogeneity in @xmath11 ( i.e. , @xmath14 ) on the splay states and on stationary chimera states , focusing on the case @xmath59 studied in @xcite . we find that when the inhomogeneity is sufficiently weak , the above solutions persist . however , as the magnitude of the inhomogeneity increases , new types of solutions are born . the origin and spatial structure of these new states can be understood with the help of the self - consistency relation ( [ self - consistency ] ) . since eq . ( [ self - consistency ] ) is invariant under the transformation @xmath60 with @xmath61 an arbitrary real constant , the local order parameter @xmath51 for the coupling function @xmath62 can be written as @xmath63 where @xmath64 is positive and @xmath65 with @xmath66 and @xmath67 both real . substituting the ansatz ( [ ansatz0 ] ) into eq . ( [ self - consistency ] ) we obtain the following pair of integral - algebraic equations @xmath68 where @xmath69 and @xmath70 . these equations may also be written in the more convenient form @xmath71 in the following we consider two choices for the inhomogeneity @xmath15 , a bump @xmath72 , @xmath19 , and a periodic inhomogeneity @xmath73 where @xmath21 is a positive integer . the resulting equations possess an important and useful symmetry , @xmath74 , @xmath75 . moreover , for @xmath15 satisfying @xmath76 with @xmath21 an integer , a solution @xmath51 of the self - consistency relation implies that @xmath77 is also a solution . here @xmath78 is an arbitrary integer . in the following we use eqs . ( [ sc - full12 ] ) and ( [ sc - full22 ] ) repeatedly to study the changes in both the splay states and the stationary chimera states as the magnitude @xmath79 of the inhomogeneity increases , and compare the resulting predictions with numerical simulation of @xmath80 oscillators evenly distributed in @xmath54 $ ] . we now consider the case where @xmath15 has a bump defect at @xmath82 . for the spatial profile of the defect we pick @xmath83 , where @xmath79 and @xmath19 are parameters that can be varied . when @xmath84 and @xmath11 is a constant , eq . ( [ phase_eq ] ) exhibits so - called splay state solutions with @xmath85 , where @xmath37 is the overall rotation frequency and @xmath86 is an integer called the twist number . the phase in this type of solution drifts with speed @xmath87 to the right but the order parameter is stationary . consequently we think of the splay states as a stationary rotating states . linear stability analysis for @xmath62 shows that the splay state is stable when @xmath88 @xcite , a result that is easily confirmed in simulations starting from randomly distributed initial phases . in the following we consider the case @xmath89 and simulate @xmath80 oscillators evenly distributed in @xmath54 $ ] for different values of @xmath24 and @xmath25 . for splay states observed with @xmath90 , @xmath91 and @xmath92 . ( a ) @xmath93 . ( b ) @xmath94 . ( c ) @xmath95 . ( d ) @xmath96 . the states travel to the right ( @xmath97 ) . , title="fig:",height=113 ] for splay states observed with @xmath90 , @xmath91 and @xmath92 . ( a ) @xmath93 . ( b ) @xmath94 . ( c ) @xmath95 . ( d ) @xmath96 . the states travel to the right ( @xmath97 ) . , title="fig:",height=113 ] for splay states observed with @xmath90 , @xmath91 and @xmath92 . ( a ) @xmath93 . ( b ) @xmath94 . ( c ) @xmath95 . ( d ) @xmath96 . the states travel to the right ( @xmath97 ) . , title="fig:",height=113 ] for splay states observed with @xmath90 , @xmath91 and @xmath92 . ( a ) @xmath93 . ( b ) @xmath94 . ( c ) @xmath95 . ( d ) @xmath96 . the states travel to the right ( @xmath97 ) . , title="fig:",height=113 ] figure [ fig : cosx_bump_splay_four ] shows how the splay solutions change as @xmath24 increases . when @xmath24 becomes nonzero but remains small , the splay states persist but their phase @xmath10 no longer varies uniformly in space ( fig . [ fig : cosx_bump_splay_four](a ) ) . instead @xmath98 where @xmath99 is a continuous function of @xmath16 with @xmath100 . we refer to this type of state as a near - splay state . as @xmath24 becomes larger an incoherent region appears in the vicinity of @xmath82 , with width that increases with increasing @xmath24 ( figs . [ fig : cosx_bump_splay_four](b , c ) ) . we refer to this type of state as a chimera splay state . as @xmath24 increases further and exceeds a second threshold , a new region of incoherence is born ( fig . [ fig : cosx_bump_splay_four](d ) ) . figure [ fig : cosx_bump_splay ] provides additional information about the partially coherent near - splay state in fig . [ fig : cosx_bump_splay_four](b ) . the figure shows the real order parameters @xmath57 and @xmath58 together with @xmath101 , the local rotation frequency averaged over a long time interval ( fig . [ fig : cosx_bump_splay](d ) ) , and reveals that in the coherent region the oscillation frequencies are identical ( with @xmath102 ) , with an abrupt but continuous change in the frequency distribution within the incoherent region . the figures also reveals that as @xmath24 increases the incoherent region develops a stronger and stronger asymmetry with respect to @xmath82 , the bump maximum . this is a consequence of the asymmetry introduced by the direction of travel , i.e. , the sign of the frequency @xmath37 in eq . ( [ splay_ansatz ] ) as discussed further below . for @xmath90 and @xmath103 . ( b ) the corresponding @xmath57 . ( c ) the corresponding @xmath58 . ( d ) the oscillator frequency @xmath101 averaged over the time interval @xmath104 ( solid line ) . in the coherent region @xmath101 coincides with the global oscillation frequency @xmath105 ( open circles ) . all simulations are done with @xmath92 and @xmath80.,title="fig:",height=113 ] for @xmath90 and @xmath103 . ( b ) the corresponding @xmath57 . ( c ) the corresponding @xmath58 . ( d ) the oscillator frequency @xmath101 averaged over the time interval @xmath104 ( solid line ) . in the coherent region @xmath101 coincides with the global oscillation frequency @xmath105 ( open circles ) . all simulations are done with @xmath92 and @xmath80.,title="fig:",height=113 ] for @xmath90 and @xmath103 . ( b ) the corresponding @xmath57 . ( c ) the corresponding @xmath58 . ( d ) the oscillator frequency @xmath101 averaged over the time interval @xmath104 ( solid line ) . in the coherent region @xmath101 coincides with the global oscillation frequency @xmath105 ( open circles ) . all simulations are done with @xmath92 and @xmath80.,title="fig:",height=113 ] for @xmath90 and @xmath103 . ( b ) the corresponding @xmath57 . ( c ) the corresponding @xmath58 . ( d ) the oscillator frequency @xmath101 averaged over the time interval @xmath104 ( solid line ) . in the coherent region @xmath101 coincides with the global oscillation frequency @xmath105 ( open circles ) . all simulations are done with @xmath92 and @xmath80.,title="fig:",height=113 ] these states and the transitions between them can be explained within the framework of the self - consistent analysis . to compute the solution branches and the transition thresholds , we numerically continue solutions of eqs . ( [ sc - full12 ] ) and ( [ sc - full22 ] ) with respect to the parameter @xmath24 . when @xmath106 , the splay state ( with positive slope ) corresponds to @xmath107 , @xmath108 , @xmath109 , with @xmath110 . the order parameter is therefore @xmath111 . we use this splay state as the starting point for continuation . as @xmath24 increases , a region of incoherence develops in the phase pattern in the vicinity of @xmath82 . from the point of view of the self - consistent analysis , the incoherent region corresponds to the region where the natural frequency exceeds the amplitude @xmath57 of the complex order parameter . the boundaries between the coherent and incoherent oscillators are thus determined by the relation @xmath112 . for @xmath94 , the left and right boundaries are thus @xmath113 and @xmath114 , respectively . these predictions are in good agreement with the values measured in direct numerical simulation ( fig . [ fig : cosx_bump_splay](a ) ) . figure [ fig : cosx_bump_omega0_continue_splay](a ) shows the overall frequency @xmath37 obtained by numerical continuation of the solution of eqs . ( [ sc - full12 ] ) and ( [ sc - full22 ] ) in the parameter @xmath24 , while figs . [ fig : cosx_bump_omega0_continue_splay](b , c ) show the corresponding results for the fraction @xmath115 of the domain occupied by the coherent oscillators and the extent of the ( first ) region of incoherence , i.e. , the interval @xmath116 , also as functions of @xmath24 . from the figure we can see a clear transition at @xmath117 from a single domain - filling coherent state to a `` splay state with one incoherent cluster '' in @xmath116 , followed by a subsequent transition at @xmath118 from this state to a `` splay state with two incoherent clusters '' . these transitions occur when the profiles of @xmath119 and @xmath57 touch as @xmath24 increases and these points of tangency therefore correspond to the locations where coherence is first lost . figure [ fig : transitions ] shows that tangencies between @xmath119 and @xmath57 occur when @xmath117 and @xmath120 , implying that intervals of incoherent oscillators appear first at @xmath82 ( i.e. , the bump maximum ) and subsequently at @xmath121 , as @xmath24 increases . these predictions are in excellent agreement with the direct numerical simulations shown in fig . [ fig : cosx_bump_splay_four ] . moreover , the critical values of @xmath24 predicted by the self - consistency analysis are fully consistent with the simulation results when @xmath122 is increased quasi - statically ( not shown ) . in each case we repeated the simulations for decreasing @xmath122 but found no evidence of hysteresis in these transitions . , ( b ) the fraction @xmath115 of the domain occupied by the coherent oscillators , and ( c ) the width of the incoherent region @xmath116 , all as functions of the parameter @xmath24 . the calculation is for @xmath123 , @xmath92 and @xmath80.,title="fig:",height=151 ] , ( b ) the fraction @xmath115 of the domain occupied by the coherent oscillators , and ( c ) the width of the incoherent region @xmath116 , all as functions of the parameter @xmath24 . the calculation is for @xmath123 , @xmath92 and @xmath80.,title="fig:",height=151 ] , ( b ) the fraction @xmath115 of the domain occupied by the coherent oscillators , and ( c ) the width of the incoherent region @xmath116 , all as functions of the parameter @xmath24 . the calculation is for @xmath123 , @xmath92 and @xmath80.,title="fig:",height=151 ] and @xmath57 at the critical values @xmath24 for the appearance of new regions of incoherence around ( a ) @xmath82 for @xmath124 and ( b ) @xmath125 for @xmath126 . the calculation is for @xmath123 , @xmath92 and @xmath127.,title="fig:",height=151 ] and @xmath57 at the critical values @xmath24 for the appearance of new regions of incoherence around ( a ) @xmath82 for @xmath124 and ( b ) @xmath125 for @xmath126 . the calculation is for @xmath123 , @xmath92 and @xmath127.,title="fig:",height=151 ] chimera states with @xmath128 evenly distributed coherent clusters are readily observed when @xmath84 and @xmath11 is a constant . these states persist when a bump is introduced into the frequency distribution @xmath15 . figure [ fig : cosx_bump_2cluster ] shows the phase distribution and the corresponding local order parameters @xmath57 and @xmath58 for @xmath90 when @xmath129 . the figure shows that the two clusters persist , but are now always located near @xmath130 and @xmath131 . this is a consequence of the fact that the presence of the bump breaks the translation invariance of the system . figures [ fig : cosx_bump_2cluster](b , c ) show that the local order parameter @xmath132 has the symmetry @xmath133 . this symmetry implies @xmath134 should take the form @xmath135 , where @xmath136 is real . the corresponding self - consistency equation takes the form @xmath137 the result of numerical continuation of the solutions of eq . ( [ bump_scequation_2cluster ] ) are shown in fig . [ fig : cosx_bump_omega0_continue_2cluster ] . the 2-cluster chimera state persists to large values of @xmath24 , with the size of the coherent clusters largely insensitive to the value of @xmath24 . this prediction has been corroborated using direct simulation of eq . ( [ phase_eq_discrete ] ) with @xmath80 oscillators . in a 2-cluster chimera state for @xmath90 and @xmath129 . ( b ) the corresponding order parameter @xmath57 . ( c ) the corresponding order parameter @xmath58 . note that the oscillators in the two clusters oscillate with the same frequency but @xmath138 out of phase . the calculation is done with @xmath123 , @xmath92 and @xmath80.,title="fig:",height=151 ] in a 2-cluster chimera state for @xmath90 and @xmath129 . ( b ) the corresponding order parameter @xmath57 . ( c ) the corresponding order parameter @xmath58 . note that the oscillators in the two clusters oscillate with the same frequency but @xmath138 out of phase . the calculation is done with @xmath123 , @xmath92 and @xmath80.,title="fig:",height=151 ] in a 2-cluster chimera state for @xmath90 and @xmath129 . ( b ) the corresponding order parameter @xmath57 . ( c ) the corresponding order parameter @xmath58 . note that the oscillators in the two clusters oscillate with the same frequency but @xmath138 out of phase . the calculation is done with @xmath123 , @xmath92 and @xmath80.,title="fig:",height=151 ] and the order parameter amplitude @xmath139 on @xmath24 . ( b ) the fraction @xmath115 of coherent oscillators as a function of @xmath24 . the calculation is done with @xmath123 , @xmath92 and @xmath127.,title="fig:",height=151 ] and the order parameter amplitude @xmath139 on @xmath24 . ( b ) the fraction @xmath115 of coherent oscillators as a function of @xmath24 . the calculation is done with @xmath123 , @xmath92 and @xmath127.,title="fig:",height=151 ] when @xmath11 is constant , finite size effects cause the phase pattern to fluctuate in location . in @xcite , we demonstrate that this fluctuation is well modeled by brownian motion in which the variance is proportional to @xmath140 , even though the original system is strictly deterministic . as mentioned above , when @xmath15 is spatially dependent , the translation symmetry is broken and the coherent cluster has a preferred location . figure [ fig : cosx_bump_2cluster ] suggests that local maxima of the order parameter @xmath141 can be used to specify the location of coherent clusters . consequently we plot in fig . [ fig : bump_pos_vs_t ] the position @xmath142 of the right coherent cluster as a function of time for three different values of the parameter @xmath25 . we see that the inhomogeneity pins the coherent cluster to a particular location , and that the cluster position executes apparently random oscillations about this preferred location , whose amplitude increases with increasing @xmath25 , i.e. , with decreasing width of the bump . figure [ fig : std_of_pos ] shows the standard deviation of the position @xmath142 of the coherent cluster as a function of the parameter @xmath25 . of the coherent cluster as a function of time when @xmath96 , @xmath92 , @xmath127 and ( a ) @xmath143 , ( b ) @xmath144 and ( c ) @xmath123.,title="fig:",height=132 ] of the coherent cluster as a function of time when @xmath96 , @xmath92 , @xmath127 and ( a ) @xmath143 , ( b ) @xmath144 and ( c ) @xmath123.,title="fig:",height=132 ] of the coherent cluster as a function of time when @xmath96 , @xmath92 , @xmath127 and ( a ) @xmath143 , ( b ) @xmath144 and ( c ) @xmath123.,title="fig:",height=132 ] as a function of the parameter @xmath25 for @xmath96 , @xmath92 and @xmath127.,height=151 ] is well approximated by a normal distribution when @xmath145 . , height=151 ] the behavior shown in figs . [ fig : bump_pos_vs_t ] and [ fig : std_of_pos ] can be modeled using an ornstein - uhlenbeck process , i.e. , a linear stochastic ordinary differential equation of the form @xmath146 where @xmath147 represents the strength of the attraction to the preferred location @xmath43 , and @xmath148 indicates the strength of the noise . models of this type are expected to apply on an appropriate timescale only : the time increment @xmath149 between successive steps of the stochastic process must be large enough that the position of the cluster can be thought of as the result of a large number of pseudo - random events and hence normally distributed , but not so large that nonlinear effects become significant . figure [ fig : hist ] reveals that for an appropriate interval of @xmath149 the fluctuations @xmath150 are indeed normally distributed , thereby providing support for the applicability of eq . ( [ eq : oe ] ) to the present system . equation ( [ eq : oe ] ) has the solution @xmath151 where @xmath152 , @xmath153 and @xmath154 . to fit the parameters to the data in fig . [ fig : bump_pos_vs_t ] we notice that the relationship between consecutive observations @xmath155 is linear with an _ i.i.d . _ error term @xmath156 , where @xmath157 denotes the normal distribution with zero mean and unit variance ( fig . [ fig : hist ] ) and @xmath158 is a constant . a least - squares fit to the data @xmath159 gives the parameters @xmath147 , @xmath43 and @xmath148 . we find that for @xmath160 , @xmath161 and @xmath92 , the choice @xmath162 works well and yields the empirical model parameters @xmath163 , @xmath164 and @xmath165 , a result that is in good agreement with the simulation results for @xmath161 summarized in fig . [ fig : std_of_pos ] . in this section we consider the case @xmath73 . when @xmath24 is small , the states present for @xmath166 persist , but with increasing @xmath24 one finds a variety of intricate dynamical behavior . in the following we set @xmath167 and @xmath168 , with @xmath24 and @xmath21 as parameters to be varied . when @xmath15 is a constant , splay states are observed in which the phase @xmath10 varies linearly with @xmath16 . when @xmath24 is nonzero but small , the splay states persist as the near - splay states described by eq . ( [ splay_ansatz ] ) ; fig . [ fig : cosx_inho_nearsplay_wam1234](a ) shows an example of such a state when @xmath169 . as @xmath24 becomes larger , incoherent regions appear and these increase in width as @xmath24 increases further ( e.g. , fig . [ fig : cosx_inho_nearsplay_wam1234](b)(d ) ) . we refer to this type of state as a chimera splay state , as in the bump inhomogeneity case . figures [ fig : cosx_inho_nearsplay_wam1234 ] show the phase distribution obtained for different values of @xmath24 . for these solutions , the slope of the coherent region is no longer constant but the oscillators rotate with a constant overall frequency @xmath37 . this type of solution is also observed for other values of @xmath21 . figure [ fig : cosx_inho_l_cluster_345 ] shows examples of chimera splay states for @xmath170 , @xmath171 and @xmath172 , with @xmath21 coherent clusters in each case . . chimera splay states for ( b ) @xmath173 , ( c ) @xmath174 and ( d ) @xmath175 . in all cases @xmath92 and @xmath80.,title="fig:",width=132 ] . chimera splay states for ( b ) @xmath173 , ( c ) @xmath174 and ( d ) @xmath175 . in all cases @xmath92 and @xmath80.,title="fig:",width=132 ] . chimera splay states for ( b ) @xmath173 , ( c ) @xmath174 and ( d ) @xmath175 . in all cases @xmath92 and @xmath80.,title="fig:",width=132 ] . chimera splay states for ( b ) @xmath173 , ( c ) @xmath174 and ( d ) @xmath175 . in all cases @xmath92 and @xmath80.,title="fig:",width=132 ] and ( a ) @xmath176 ( 3-cluster state ) , ( b ) @xmath177 ( 4-cluster state ) , and ( c ) @xmath178 ( 5-cluster state ) . in all cases @xmath92 and @xmath80.,title="fig:",height=132 ] and ( a ) @xmath176 ( 3-cluster state ) , ( b ) @xmath177 ( 4-cluster state ) , and ( c ) @xmath178 ( 5-cluster state ) . in all cases @xmath92 and @xmath80.,title="fig:",height=132 ] and ( a ) @xmath176 ( 3-cluster state ) , ( b ) @xmath177 ( 4-cluster state ) , and ( c ) @xmath178 ( 5-cluster state ) . in all cases @xmath92 and @xmath80.,title="fig:",height=132 ] to compute the solution branches and identify thresholds for additional transitions , we continue the solutions of eqs . ( [ sc - full12 ] ) and ( [ sc - full22 ] ) with respect to the parameter @xmath24 . figures [ fig : splay1 ] and [ fig : splay2 ] show the dependence of @xmath37 and of the coherent fraction @xmath115 on @xmath24 when @xmath179 and @xmath180 , respectively . the figures indicate that the coherent fraction @xmath115 falls below 1 at @xmath181 ( @xmath179 ) , @xmath182 ( @xmath183 ) , and @xmath184 ( @xmath185 , 4 and 5 ) . these values coincide with the parameter values at which an incoherent region emerges . in addition , fig . [ fig : splay1](b ) reveals the presence of a second transition , at @xmath186 , corresponding to the emergence of a 2-cluster state from a 1-cluster state . figures [ fig : cosx_cosx_tangent ] and [ fig : tangent_l2 ] show the profiles of @xmath187 and @xmath57 at the critical values of @xmath24 at which incoherent regions emerge . and ( b ) the fraction @xmath115 of the domain occupied by the coherent oscillators on @xmath24 , with @xmath188 and @xmath189.,title="fig:",height=151 ] and ( b ) the fraction @xmath115 of the domain occupied by the coherent oscillators on @xmath24 , with @xmath188 and @xmath189.,title="fig:",height=151 ] and ( b ) the fraction @xmath115 of the domain occupied by the coherent oscillators on @xmath24 , with @xmath190 and @xmath189.,title="fig:",height=151 ] and ( b ) the fraction @xmath115 of the domain occupied by the coherent oscillators on @xmath24 , with @xmath190 and @xmath189.,title="fig:",height=151 ] and @xmath57 at the critical values @xmath24 for the appearance of a new region of incoherence around ( a ) @xmath82 for @xmath191 and ( b ) @xmath192 for @xmath193 . parameters : @xmath179 and @xmath92.,title="fig:",height=151 ] and @xmath57 at the critical values @xmath24 for the appearance of a new region of incoherence around ( a ) @xmath82 for @xmath191 and ( b ) @xmath192 for @xmath193 . parameters : @xmath179 and @xmath92.,title="fig:",height=151 ] and @xmath57 at the critical value @xmath194 . parameters : @xmath183 and @xmath92.,height=151 ] to understand the @xmath21-cluster chimera states reported in fig . [ fig : cosx_inho_l_cluster_345 ] , we computed the local order parameter @xmath51 and found that @xmath57 is approximately constant while the phase @xmath58 varies at a constant rate , suggesting the ansatz @xmath195 . with this ansatz , eqs . ( [ sc - full12 ] ) and ( [ sc - full22 ] ) reduce to a single equation , @xmath196 for all positive integers @xmath197 . figure [ fig:3cluster ] shows the result of numerical continuation of a solution for @xmath185 . the figure reveals no further transitions , indicating that the @xmath185 chimera state persists to large values of @xmath24 . numerical simulation shows that these states ( @xmath185 , 4 , 5 ) are stable and persist up to @xmath198 . and ( b ) the coherent fraction @xmath115 on @xmath24 , with @xmath199 and @xmath189.,title="fig:",height=132 ] and ( b ) the coherent fraction @xmath115 on @xmath24 , with @xmath199 and @xmath189.,title="fig:",height=132 ] in a 1-cluster chimera state for @xmath90 and @xmath200 . ( b ) the corresponding order parameter @xmath57 . ( c ) the corresponding order parameter @xmath58 . figure ( a ) shows that the presence of a nearly coherent region near @xmath82 with oscillators that oscillate @xmath138 out of phase with the coherent cluster . the calculation is done with @xmath92 and @xmath80.,title="fig:",height=132 ] in a 1-cluster chimera state for @xmath90 and @xmath200 . ( b ) the corresponding order parameter @xmath57 . ( c ) the corresponding order parameter @xmath58 . figure ( a ) shows that the presence of a nearly coherent region near @xmath82 with oscillators that oscillate @xmath138 out of phase with the coherent cluster . the calculation is done with @xmath92 and @xmath80.,title="fig:",height=132 ] in a 1-cluster chimera state for @xmath90 and @xmath200 . ( b ) the corresponding order parameter @xmath57 . ( c ) the corresponding order parameter @xmath58 . figure ( a ) shows that the presence of a nearly coherent region near @xmath82 with oscillators that oscillate @xmath138 out of phase with the coherent cluster . the calculation is done with @xmath92 and @xmath80.,title="fig:",height=132 ] and @xmath37 , and ( b ) the coherent fraction @xmath115 , all as functions of @xmath24 for the chimera state shown in fig . [ fig : cosx_inho_1cluster ] . ( c ) @xmath187 and @xmath57 as functions of @xmath16 when @xmath201.,title="fig:",height=151 ] and @xmath37 , and ( b ) the coherent fraction @xmath115 , all as functions of @xmath24 for the chimera state shown in fig . [ fig : cosx_inho_1cluster ] . ( c ) @xmath187 and @xmath57 as functions of @xmath16 when @xmath201.,title="fig:",height=151 ] and @xmath37 , and ( b ) the coherent fraction @xmath115 , all as functions of @xmath24 for the chimera state shown in fig . [ fig : cosx_inho_1cluster ] . ( c ) @xmath187 and @xmath57 as functions of @xmath16 when @xmath201.,title="fig:",height=151 ] figure [ fig : cosx_inho_1cluster ] shows an example of a 1-cluster chimera state when @xmath202 ; 1-cluster chimera states of this type have not thus far been reported for @xmath90 , @xmath166 , where computations always result in 2-cluster states @xcite . to understand the origin of this unexpected 1-cluster state we use fig . [ fig : cosx_inho_1cluster](b ) to conclude that @xmath51 is of the form @xmath203 , and hence that eqs . ( [ sc - full12 ] ) and ( [ sc - full22 ] ) reduce to the single equation @xmath204 figure [ fig : romegae ] shows the result of numerical continuation of the solution of this equation as a function of @xmath24 . the figure reveals a transition at @xmath201 ( figs . [ fig : romegae](a , b ) ) . when @xmath205 , the solution is a 2-cluster chimera ; as @xmath24 increases through @xmath206 the coherent region around @xmath82 disappears , leaving a single cluster chimera state ( fig . [ fig : romegae](c ) ) . for @xmath90 and constant @xmath11 simulations always evolve into either a 2-cluster chimera state or a splay state @xcite . it is expected that when @xmath24 is small , the 2-cluster chimera is not destroyed . figures [ fig : cosx_inho_2cluster](a , b ) gives examples of this type of state with @xmath183 and @xmath185 inhomogeneities , respectively . in both cases the clusters are located at specific locations selected by the inhomogeneity . and ( a ) @xmath207 and ( b ) @xmath208 . in both cases @xmath92 and @xmath80.,title="fig:",height=132 ] and ( a ) @xmath207 and ( b ) @xmath208 . in both cases @xmath92 and @xmath80.,title="fig:",height=132 ] the order parameter profile with @xmath185 corresponding to the numerical solution in fig . [ fig : cosx_inho_2cluster](b ) has the same symmetry properties as that in figs . [ fig : cosx_bump_2cluster](b , c ) . it follows that we may set @xmath209 , @xmath210 in the order parameter representation ( [ ansatz0 ] ) , resulting once again in the self - consistency relation ( [ bump_scequation_2cluster ] ) . we have continued the solutions of this relation for @xmath185 using the simulation in fig . [ fig : cosx_inho_2cluster](b ) with @xmath211 to initialize continuation in @xmath24 . figure [ fig : chimera2 - 1 ] shows the result of numerical continuation of the order parameter for this state . no further transitions are revealed . there are three preferred locations for the coherent clusters , related by the translation symmetry @xmath212 , @xmath213 , as shown in fig . [ fig : cosx_inho_wa01_wp3_2cluster ] . ( solid line ) and @xmath139 ( dashed line ) , and ( b ) the coherent fraction @xmath115 on @xmath24 when @xmath185 and @xmath92 . , title="fig:",height=132 ] ( solid line ) and @xmath139 ( dashed line ) , and ( b ) the coherent fraction @xmath115 on @xmath24 when @xmath185 and @xmath92 . , title="fig:",height=132 ] . the simulation is done with @xmath92 and @xmath127.,title="fig:",height=132 ] . the simulation is done with @xmath92 and @xmath127.,title="fig:",height=132 ] . the simulation is done with @xmath92 and @xmath127.,title="fig:",height=132 ] we now turn to states with a spatially structured order parameter undergoing translation . with @xmath214 and constant @xmath11 the system ( [ phase_eq ] ) exhibits a fully coherent but non - splay state that travels with a constant speed @xmath215 when @xmath216 @xcite . figure [ fig : phase_cos12x](a ) shows a snapshot of this traveling coherent state while fig . [ fig : phase_cos12x](b ) shows its position @xmath155 as a function of time . the speed of travel is constant and can be obtained by solving a nonlinear eigenvalue problem @xcite . in this section , we investigate how the inhomogeneities @xmath217 and @xmath73 affect the dynamical behavior of this state . is constant . ( b ) the position @xmath155 of this state as a function of time . the simulation is done for @xmath214 with @xmath218 and @xmath127.,title="fig:",height=113 ] is constant . ( b ) the position @xmath155 of this state as a function of time . the simulation is done for @xmath214 with @xmath218 and @xmath127.,title="fig:",height=113 ] in this section we perform two types of numerical experiments . in the first , we fix @xmath122 and start with the traveling coherent solution for @xmath106 . we then gradually increase @xmath24 gradually at fixed values of @xmath25 and @xmath122 . in the second we fix @xmath220 , @xmath19 , and vary @xmath122 . for each @xmath216 we find that the coherent state continues to travel , albeit nonuniformly , until @xmath24 reaches a threshold value that depends of the values of @xmath122 and @xmath25 . the case @xmath218 provides an example . figure [ fig : hidden1](a ) shows the traveling coherent state in the homogeneous case ( @xmath106 ) while fig . [ fig : hidden1](b ) shows the corresponding state in the presence of a frequency bump @xmath221 with @xmath222 and @xmath161 . in this case the presence of inhomogeneity leads first to a periodic fluctuation in the magnitude of the drift speed followed by , as @xmath24 continues to increase , a transition to a new state in which the direction of the drift oscillates periodically ( fig . [ fig : hidden1](b ) ) . we refer to states of this type as direction - reversing waves , by analogy with similar behavior found in other systems supporting the presence of such waves @xcite . with increasing @xmath24 the reversals become localized in space ( and possibly aperiodic , fig . [ fig : hidden2](a ) ) and then cease , leading to a stationary pinned structure at @xmath223 ( fig . [ fig : hidden2](b ) ) . figure [ fig : xvst ] shows the position @xmath155 of the maximum of the local order parameter @xmath224 of the coherent state as a function of time for the cases in figs . [ fig : hidden1 ] and [ fig : hidden2 ] , showing the transition from translation to pinning as @xmath24 increases , via states that are reflection - symmetric _ on average_. the final state is a steady reflection - symmetric pinned state aligned with the imposed inhomogeneity . in fact , the dynamics of the present system may be more complicated than indicated above since a small group of oscillators located in regions where the order parameter undergoes rapid variation in space may lose coherence in a periodic fashion even when @xmath166 thereby providing a competing source of periodic oscillations in the magnitude of the drift speed . as documented in @xcite this is the case when @xmath225 . for @xmath226 , however , the coherent state drifts uniformly when @xmath166 and this is therefore the case studied in greatest detail . of the coherent state in figs . [ fig : hidden1 ] and [ fig : hidden2 ] as a function of time for ( a ) @xmath227 , ( b ) @xmath228 , ( c ) @xmath229 , and ( d ) @xmath230 . in all cases @xmath123 , @xmath218 and @xmath127.,title="fig:",height=113 ] of the coherent state in figs . [ fig : hidden1 ] and [ fig : hidden2 ] as a function of time for ( a ) @xmath227 , ( b ) @xmath228 , ( c ) @xmath229 , and ( d ) @xmath230 . in all cases @xmath123 , @xmath218 and @xmath127.,title="fig:",height=113 ] of the coherent state in figs . [ fig : hidden1 ] and [ fig : hidden2 ] as a function of time for ( a ) @xmath227 , ( b ) @xmath228 , ( c ) @xmath229 , and ( d ) @xmath230 . in all cases @xmath123 , @xmath218 and @xmath127.,title="fig:",height=113 ] of the coherent state in figs . [ fig : hidden1 ] and [ fig : hidden2 ] as a function of time for ( a ) @xmath227 , ( b ) @xmath228 , ( c ) @xmath229 , and ( d ) @xmath230 . in all cases @xmath123 , @xmath218 and @xmath127.,title="fig:",height=113 ] the profile of the pinned coherent state can also be determined from a self - consistency analysis . for @xmath231 , the local order parameter @xmath51 can be written in the form @xmath232 owing to the reflection symmetry of the solution @xmath233 ; on applying a rotation in @xmath234 we may take @xmath64 to be real . the self - consistency equation then becomes @xmath235 when @xmath236 , @xmath161 and @xmath226 the solution of these equations is @xmath237 , results that are consistent with the values obtained from numerical simulation . for these parameter values , @xmath238 for @xmath239 $ ] , indicating that all phases rotate with the same frequency @xmath37 . as we decrease @xmath24 to @xmath240 the profiles @xmath241 and @xmath242 start to touch ( fig . [ fig : romega_omega0_011 ] ) and for yet lower @xmath24 the stationary coherent state loses stability and begins to oscillate as described in the previous paragraph . we have also conducted experiments at a fixed value of @xmath79 and @xmath19 while changing @xmath122 . for example , at fixed @xmath236 , @xmath161 and @xmath226 the system is in the pinned state shown in fig . [ fig : hidden2](b ) . since the speed of the coherent state with @xmath166 gradually increases as @xmath122 decreases , we anticipate that a given inhomogeneity will find it harder and harder to pin the state as @xmath122 decreases . this is indeed the case , and we find that there is a critical value of @xmath122 at which the given inhomogeneity is no longer able to pin the structure , with depinning via back and forth oscillations of the structure @xcite . for yet smaller values of @xmath122 , this type of oscillation also loses stability and evolves into near - splay states . when we increase @xmath122 again we uncover hysteresis in each of these transitions . figure [ fig : hidden3 ] shows two distinct states at identical parameter values : @xmath236 , @xmath161 and @xmath243 generated using different protocols : fig . [ fig : hidden3](a ) shows a direction - reversing state evolved from a traveling coherent state when we increase @xmath24 from 0 to 0.12 at @xmath243 , while fig . [ fig : hidden3](b ) shows a pinned state generated from the pinned state at @xmath218 when we change @xmath122 from 0.75 to 0.76 at fixed @xmath223 . of a coherent solution when @xmath245 . ( b ) the position @xmath155 of the coherent solution as a function of time . in both cases @xmath179 , @xmath92 and @xmath127.,title="fig:",height=113 ] of a coherent solution when @xmath245 . ( b ) the position @xmath155 of the coherent solution as a function of time . in both cases @xmath179 , @xmath92 and @xmath127.,title="fig:",height=113 ] we now turn to the case @xmath246 . similar to the bump inhomogeneity case , when @xmath122 is fixed , the traveling coherent state will continue to travel until @xmath24 reaches certain threshold . however , we did not find the pinned state as in the previous subsection around @xmath218 . here we focus on the case @xmath247 for which the traveling coherent state has a reasonable speed when @xmath166 and take @xmath179 . when @xmath24 is small , the state remains coherent ( fig . [ fig : cos12x_l1_omega0_0028](a ) ) and continues to drift , albeit no longer with a uniform speed of propagation . figure [ fig : cos12x_l1_omega0_0028](b ) shows that the speed executes slow , small amplitude oscillations about a well - defined mean value @xmath248 shown in fig . [ fig : cos12x_omegav](b ) ; the corresponding time - averaged oscillation frequency @xmath249 is shown in fig . [ fig : cos12x_omegav](a ) . when @xmath24 is increased in sufficiently small increments the oscillations grow in amplitude but the solution continues to travel to the left until @xmath250 where a hysteretic transition to a near - splay state takes place . figure [ fig : cos12x_omegav](b ) shows that prior to this transition the average speed first decreases as a consequence of the inhomogeneity , but then increases abruptly just before the transition owing to the loss of coherence on the part of a group of oscillators and the resulting abrupt increase in asymmetry of the order parameter . and ( b ) the mean drift speed @xmath251 , both as functions of @xmath24 when @xmath179 , @xmath252.,title="fig:",height=132 ] and ( b ) the mean drift speed @xmath251 , both as functions of @xmath24 when @xmath179 , @xmath252.,title="fig:",height=132 ] in addition to the states discussed in the previous sections , eq . ( [ phase_eq ] ) with constant @xmath11 admits traveling one - cluster chimera states when @xmath253 $ ] and appropriate values of the the phase lag @xmath122 . this state consists of a single coherent cluster that drifts through an incoherent background as time evolves at more or less a constant speed . figure [ fig : phase_cos34x](a ) shows a snapshot of such a state when @xmath254 . the direction of motion is determined by the gradient of the phase in the coherent region : the cluster travels to the left when the gradient is positive and to the right when the gradient is negative . figure [ fig : phase_cos34x](b ) shows the position @xmath155 of the coherent cluster as a function of time and confirms that the cluster moves to the right at an almost constant speed . in @xcite , we use numerical simulations to conclude that for @xmath254 the traveling chimera is stable in the interval @xmath255 . we therefore focus on the effects of spatial inhomogeneity on the traveling chimera state when @xmath256 . in fact the traveling chimera state is more complex than suggested in fig . [ fig : phase_cos34x](a , b ) : unlike the states discussed in the previous sections , the profile of the local order parameter fluctuates in time , suggesting that the state does not drift strictly as a rigid object . for a traveling chimera state in a spatially homogeneous system . ( b ) the position @xmath155 of the coherent cluster as a function of time . the simulation is done for @xmath257 with @xmath256 and @xmath127.,title="fig:",height=113 ] for a traveling chimera state in a spatially homogeneous system . ( b ) the position @xmath155 of the coherent cluster as a function of time . the simulation is done for @xmath257 with @xmath256 and @xmath127.,title="fig:",height=113 ] in this section we investigate the effect of a bump - like inhomogeneity @xmath258 on the motion of the traveling chimera state . starting with the traveling chimera state for @xmath106 , we increase @xmath24 in steps of @xmath259 . to describe the motion of the coherent cluster , we follow the method in @xcite and determine the instantaneous position @xmath155 of the cluster by minimizing the function @xmath260 ^ 2 $ ] , where @xmath261 is a reference profile , and using the minimizer @xmath262 as a proxy for @xmath142 . we find that even small @xmath24 suffices to stop a traveling chimera from moving : fig . [ fig : kappa_omega0 ] shows that the threshold @xmath263 for @xmath264 and that it increases monotonically to @xmath265 for @xmath266 . the resulting pinned state persists to values of @xmath24 as large as @xmath267 . on @xmath25 when @xmath256.,height=151 ] of the pinned coherent cluster in a traveling chimera state as a function of time when ( a ) @xmath228 , ( c ) @xmath229 , ( e ) @xmath230 . the average rotation frequency @xmath101 for ( b ) @xmath228 , ( d ) @xmath229 , ( f ) @xmath230 . in all cases @xmath256 and @xmath143 , @xmath127.,title="fig:",height=113 ] of the pinned coherent cluster in a traveling chimera state as a function of time when ( a ) @xmath228 , ( c ) @xmath229 , ( e ) @xmath230 . the average rotation frequency @xmath101 for ( b ) @xmath228 , ( d ) @xmath229 , ( f ) @xmath230 . in all cases @xmath256 and @xmath143 , @xmath127.,title="fig:",height=113 ] of the pinned coherent cluster in a traveling chimera state as a function of time when ( a ) @xmath228 , ( c ) @xmath229 , ( e ) @xmath230 . the average rotation frequency @xmath101 for ( b ) @xmath228 , ( d ) @xmath229 , ( f ) @xmath230 . in all cases @xmath256 and @xmath143 , @xmath127.,title="fig:",height=113 ] of the pinned coherent cluster in a traveling chimera state as a function of time when ( a ) @xmath228 , ( c ) @xmath229 , ( e ) @xmath230 . the average rotation frequency @xmath101 for ( b ) @xmath228 , ( d ) @xmath229 , ( f ) @xmath230 . in all cases @xmath256 and @xmath143 , @xmath127.,title="fig:",height=113 ] of the pinned coherent cluster in a traveling chimera state as a function of time when ( a ) @xmath228 , ( c ) @xmath229 , ( e ) @xmath230 . the average rotation frequency @xmath101 for ( b ) @xmath228 , ( d ) @xmath229 , ( f ) @xmath230 . in all cases @xmath256 and @xmath143 , @xmath127.,title="fig:",height=113 ] of the pinned coherent cluster in a traveling chimera state as a function of time when ( a ) @xmath228 , ( c ) @xmath229 , ( e ) @xmath230 . the average rotation frequency @xmath101 for ( b ) @xmath228 , ( d ) @xmath229 , ( f ) @xmath230 . in all cases @xmath256 and @xmath143 , @xmath127.,title="fig:",height=113 ] figures [ fig : xvst_travel_bump](a , c , e ) show the position @xmath155 of the coherent cluster as a function of time obtained using the above procedure for @xmath268 , respectively , i.e. , in the pinned regime . the figures show that the equilibrium position of the coherent region is located farther from the position @xmath82 of the inhomogeneity peak as @xmath24 increases . the bump in @xmath15 thus exerts a " repelling force on the coherent cluster , whose strength increases with the height of the bump . we interpret this observation as follows . the coherent cluster can only survive when the frequency gradient is sufficiently small , and is therefore repelled by regions where @xmath15 varies rapidly . in the present case this implies that the coherent cluster finds it easiest to survive in the wings of the bump inhomogeneity , and this position moves further from @xmath82 as @xmath24 increases . this interpretation is confirmed in figs . [ fig : xvst_travel_bump](b , d , f ) showing the average rotation frequency , @xmath269 , of the oscillators . the plateau in the profile of @xmath270 indicates frequency locking and hence the location of the coherent cluster ; the fluctuations in the position of the coherent cluster are smoothed out by the time - averaging . we now turn to the effects of a periodic inhomogeneity @xmath272 . when @xmath179 and @xmath24 increases the coherent cluster initially travels with a non - constant speed but then becomes pinned in place ; as in the case of the bump inhomogeneity quite small values of @xmath24 suffice to pin the coherent cluster in place ( for @xmath273 the value @xmath274 suffices ) . as shown in figs . [ fig : xvst_travel_periodic_l1](a , c ) the position @xmath155 of the pinned cluster relative to the local maximum of the inhomogeneity ( i.e. , @xmath82 ) depends on the value of @xmath275 . as shown in figs . [ fig : xvst_travel_periodic_l1_2](a , b ) the coherent cluster travels to the right and does so with a speed that is larger when @xmath276 than when @xmath277 . this effect becomes more pronounced as @xmath24 increases . this is because the coherent structure is asymmetric , with a preferred direction of motion , and this asymmetry increases with @xmath24 . evidently , the speed of the synchronization front at the leading edge is enhanced when @xmath278 but suppressed when @xmath279 and likewise for the desynchronization front at the rear . of the pinned coherent cluster in a traveling chimera state as a function of time when ( a ) @xmath280 , ( c ) @xmath281 . the average rotation frequency @xmath101 for ( b ) @xmath280 , ( d ) @xmath281 . in all cases @xmath256 , @xmath273 and @xmath127.,title="fig:",height=113 ] of the pinned coherent cluster in a traveling chimera state as a function of time when ( a ) @xmath280 , ( c ) @xmath281 . the average rotation frequency @xmath101 for ( b ) @xmath280 , ( d ) @xmath281 . in all cases @xmath256 , @xmath273 and @xmath127.,title="fig:",height=113 ] of the pinned coherent cluster in a traveling chimera state as a function of time when ( a ) @xmath280 , ( c ) @xmath281 . the average rotation frequency @xmath101 for ( b ) @xmath280 , ( d ) @xmath281 . in all cases @xmath256 , @xmath273 and @xmath127.,title="fig:",height=113 ] of the pinned coherent cluster in a traveling chimera state as a function of time when ( a ) @xmath280 , ( c ) @xmath281 . the average rotation frequency @xmath101 for ( b ) @xmath280 , ( d ) @xmath281 . in all cases @xmath256 , @xmath273 and @xmath127.,title="fig:",height=113 ] of the coherent cluster in a traveling chimera state as a function of time when ( a ) @xmath282 , ( b ) @xmath283 . in all cases @xmath256 , @xmath273 and @xmath127.,title="fig:",height=113 ] of the coherent cluster in a traveling chimera state as a function of time when ( a ) @xmath282 , ( b ) @xmath283 . in all cases @xmath256 , @xmath273 and @xmath127.,title="fig:",height=113 ] for @xmath284 we observe similar results . the coherent cluster is pinned in space already at small values of @xmath24 . since the inhomogeneous system has the discrete translation symmetry @xmath285 the coherent cluster has @xmath21 possible preferred positions . figure [ fig : xvst_travel_periodic_l2 ] shows an example for @xmath286 . panels ( a , c ) show snapshots of the phase distribution @xmath10 for @xmath287 in the two preferred locations ( separated by @xmath288 ) , while panels ( b , d ) show the corresponding average rotation frequency @xmath101 . of pinned traveling chimera states when @xmath281 , @xmath256 and @xmath286 . ( b , d ) the corresponding average rotation frequencies @xmath101 . in both cases @xmath127.,title="fig:",height=113 ] of pinned traveling chimera states when @xmath281 , @xmath256 and @xmath286 . ( b , d ) the corresponding average rotation frequencies @xmath101 . in both cases @xmath127.,title="fig:",height=113 ] of pinned traveling chimera states when @xmath281 , @xmath256 and @xmath286 . ( b , d ) the corresponding average rotation frequencies @xmath101 . in both cases @xmath127.,title="fig:",height=113 ] of pinned traveling chimera states when @xmath281 , @xmath256 and @xmath286 . ( b , d ) the corresponding average rotation frequencies @xmath101 . in both cases in this paper we have investigated a system of non - identical phase oscillators with nonlocal coupling , focusing on the effects of weak spatial inhomogeneity in an attempt to extend earlier results on identical oscillators to more realistic situations . two types of inhomogeneity were considered , a bump inhomogeneity in the frequency distribution specified by @xmath289 and a periodic inhomogeneity specified by @xmath20 . in each case we examined the effect of the amplitude @xmath24 of the inhomogeneity and its spatial scale @xmath290 ( @xmath291 ) on the properties of states known to be present in the homogeneous case @xmath166 , viz . , splay states and stationary chimera states , traveling coherent states and traveling chimera states @xcite . we have provided a fairly complete description of the effects of inhomogeneity on these states for the coupling functions @xmath90 , @xmath292 and @xmath293 employed in @xcite . specifically , we found that as the amplitude of the inhomogeneity increased a splay state turned into a near - splay state , characterized by a nonuniform spatial phase gradient , followed by the appearance of a stationary incoherent region centered on the location of maximum inhomogeneity amplitude . with further increase in @xmath24 additional intervals of incoherence opened up , leading to states resembling the stationary multi - cluster chimera states also present in the homogeneous system . these transitions , like many of the transitions identified in this paper , could be understood with the help of a self - consistency analysis based on the ott - antonsen ansatz @xcite , as described in the appendix . the effect of inhomogeneity on multi - cluster chimera states was found to be similar : the inhomogeneity trapped the coherent clusters in particular locations , and eroded their width as its amplitude @xmath24 increased , resulting in coalescence of incoherent regions with increasing @xmath24 . more significant are the effects of inhomogeneity on traveling coherent and traveling chimera states . here the inhomogeneity predictably pins the traveling structures but the details can be complex . figures [ fig : hidden1][fig : hidden2 ] show one such complex pinning transition that proceeds via an intermediate direction - reversing traveling wave . these waves are generated directly as a consequence of the inhomogeneity and would not be present otherwise , in contrast to homogeneous systems undergoing a symmetry - breaking hopf bifurcation as described in @xcite . many of the pinning transitions described here are hysteretic as demonstrated in fig . [ fig : hidden3 ] . the traveling chimera states are particularly fragile in this respect , with small amplitude inhomogeneities sufficient to arrest the motion of these states . in all these cases the coherent regions are found in regions of least inhomogeneity , an effect that translates into an effective repulsive interaction between the coherent cluster and the inhomogeneity . in future work we propose to explore similar dynamics in systems of more realistic nonlocally coupled oscillators and compare the results with those for similar systems with a random frequency distribution . this work was supported in part by a national science foundation collaborative research grant cmmi-1232902 . we suppose that an oscillator at position @xmath16 has intrinsic frequency @xmath15 , and assume that the frequency distribution @xmath15 is continuous . the model equation is @xmath294 we next introduce the probability density function @xmath295 characterizing the state of the system . this function must satisfy the continuity equation @xmath296 where @xmath297 satisfies the relation @xmath298 and @xmath299 . we also define the local order parameter @xmath300 then @xmath301.\ ] ] the above equations can be recast in a more convenient form using the ott - antonsen ansatz @xcite @xmath302.\ ] ] here @xmath303 represents the distribution of natural frequencies at each @xmath16 . matching terms proportional to different powers of @xmath304 , we obtain @xmath305,\ ] ] where the complex order parameter @xmath306 is given by @xmath307 if we take @xmath308 set @xmath309 and perform the implied contour integration , we obtain @xmath310 in the limit @xmath311 the distribution function @xmath312 reduces to a delta function . the corresponding quantity @xmath27 satisfies @xmath313,\label{zdef}\ ] ] where @xmath306 is given by ( [ zdef ] ) . equations ( [ zdef])([zdef ] ) constitute the required self - consistency description of the nonlocally coupled phase oscillator system with an inhomogeneous frequency distribution @xmath15 .	chimera states consisting of domains of coherently and incoherently oscillating nonlocally - coupled phase oscillators in systems with spatial inhomogeneity are studied . the inhomogeneity is introduced through the dependence of the oscillator frequency on its location . two types of spatial inhomogeneity , localized and spatially periodic , are considered and their effects on the existence and properties of multi - cluster and traveling chimera states are explored . the inhomogeneity is found to break up splay states , to pin the chimera states to specific locations and to trap traveling chimeras . many of these states can be studied by constructing an evolution equation for a complex order parameter . solutions of this equation are in good agreement with the results of numerical simulations . pacs numbers : : may be entered using the ` \pacs{#1 } ` command .
over the years , the study of current noise and noise correlations has become a respected and useful diagnosis for transport measurements on mesoscopic conductors . theoretically , noise was first computed mostly for non interacting systems @xcite . however , it soon became clear that low frequency noise could be used to isolate the quasiparticle charge @xcite and to study the statistical correlations @xcite in specific quasi one dimensional correlated electron systems , such as the edge waves in the quantum hall effect . in these chiral luttinger liquids , the charge of the collective excitations along the edges corresponds to the electron charge multiplied by the filling factor . attention is now turning towards conductors individual nano - objects which occur naturally , and which can be connected to current / voltage probes in order to perform a transport experiment . the crucial advantage of such nano objects is that they are essentially free of defects and in some circumstances they have an inherent one dimensional character . carbon nanotubes constitute the archetype of such 1d nano - objects : single wall armchair nanotubes have metallic behavior , with two propagating modes at the fermi level . incidentally , electronic correlations are known to play an important role in such systems . carbon nanotubes seem to constitute good candidates to study luttinger liquid behavior . in particular , their tunneling density of states and thus the tunneling @xmath0 characteristics is known to have a power law behavior @xcite in accordance with luttinger liquid theory . 8 cm luttinger models for nanotubes differ significantly from their quantum hall effect counterpart , because of their non - chiral character . forward and backward fields describing collective excitations effectively mix , because the interactions between electrons are spread along the whole length of the nanotube . for this reason , a straightforward transposition of the results obtained for chiral edge system proves difficult . nevertheless , non chiral luttinger liquids can be described with chiral fields @xcite . such chiral fields correspond to excitations with anomalous ( non - integer ) charge , which has eluded detection so far . in the present work , we propose an experimental geometry which allows to probe directly the underlying charges of the collective excitations . the setup consists of a nanotube whose bulk is contacted by a scanning tunneling microscope ( stm ) tip which injects electrons , while both extremities of the nanotube collect the current ( fig . [ fig1 ] ) . the current , the noise and the noise correlations are computed , and the effective charges are determined by comparison with the schottky formula @xcite for an `` infinite '' nanotube , the striking result is that noise correlations contribute to second order in the electron tunneling , in sharp contrast with a fermionic system which requires fourth order . the noise correlations are then positive , because the tunneling electron wave function is split in two counter propagating modes of the collective excitations in the nanotube . we conjecture that in the presence of 1d fermi liquid leads , modeled as in ref . @xcite , the absence of renormalization / interaction effects of the nanotube is recovered . a recent two terminal experiment studied the current current fluctuations in ropes of nanotubes @xcite . there , it is pointed out that the strong reduction of the low frequency noise can not be understood within the context of scattering theory @xcite . naive comparison with existing non - chiral luttinger liquid models @xcite would imply an interaction parameter much inferior to the free electron case . also , we mention that other multi - terminal geometries where a nanotube or a one dimensional wire is attached to more than two leads , have been considered @xcite . our proposal deals with the same geometry as ref @xcite , where a renormalization analysis identified the exponents of the current voltage characteristics . however , here the emphasis is put on the low frequency current fluctuation spectrum , both for the autocorrelation and the cross correlations between the two ends of the nanotube . the paper is organized as follows : the hamiltonian of our setup is specified in the next section , followed by a general non - equilibrium scheme based on the keldysh formalism to study transport in this device , which is independent of the type of leads chosen ( sect . results for a nanotube connected to leads are then presented in sect . 4 . a connection with the effective charges of refs . @xcite is established in sect . the transport geometry ( fig . [ fig1 ] ) implies tunneling from the tip ( normal or ferromagnetic metal ) to the nanotube , and subsequent propagation of collective excitations along the nanotube . in the absence of tunneling , the hamiltonian is thus simply the sum of the nanotube hamiltonian , described by a two mode luttinger liquid , together with the tip hamiltonian . using the standard conventions @xcite , the operator describing an electron with spin @xmath1 moving along the direction @xmath2 , from mode @xmath3 is specified in terms of a bosonic field : @xmath4 with @xmath5 a short distance cutoff , @xmath6 the fermi momentum , @xmath7 the momentum mismatch associated with the two modes , and the convention @xmath8 , @xmath9 and @xmath10 are chosen for the direction of propagation , for the nanotube branch , and for the spin orientation . it is convenient to express this bosonic phase in terms of the conventional non - chiral luttinger liquid fields @xmath11 and @xmath12 , with @xmath13 identifying the charge / spin and total / relative fields : @xmath14 with @xmath15 , @xmath16 , @xmath17 et @xmath18 . @xmath11 and @xmath12 are dual non - chiral fields . a plausible alternative would have been to express @xmath19 in terms of the chiral luttinger liquid fields . however , the present choice will be simpler later on when dealing with inhomogeneous luttinger liquids ( in order to include the leads ) , as the green s functions for @xmath11 , @xmath12 are known . the hamiltonian which describes the collective excitations in the nanotube has the standard form : @xmath20 with an interaction parameter @xmath21 and velocity @xmath22 . for the stm tip , one assumes for simplicity that only one electronic mode couples to the nanotube . the tip can thus be described by a semi - infinite luttinger liquid , as in kondo type problems . this turns out to be convenient in this problem where both bosonized nanotube fermions operators and tip fermions operators intervene . for the sake of generality , we allow the two spin components of the tip fields to have different fermi velocities @xmath23 , which allows to treat the case of a ferromagnetic metal . the fermion operator at the tip location @xmath24 is then : @xmath25 here , @xmath26 is the chiral luttinger liquid field , whose keldysh green s function at @xmath24 is given by @xcite : @xmath27 u_f^{\sigma}(t_1-t_2)/2a\right\}~ , \end{aligned}\ ] ] where @xmath28 refer to the upper or lower branch of the keldysh contour . the tunneling hamiltonian is a standard hopping term : @xmath29^{(\varepsilon ) } ~.\label{tunnel_hamiltonian}\end{aligned}\ ] ] here the superscript @xmath30 leaves either the operators in bracket unchanged ( @xmath31 ) , or transforms them into their hermitian conjugate ( @xmath32 ) . the voltage bias between the tip and the nanotube is included using the peierls substitution : the hopping amplitude @xmath33 acquires a time dependent phase @xmath34 , with the bias voltage identified as @xmath35 . we will use the convention @xmath36 . similarly , the tunneling current is defined as : @xmath37^{(\varepsilon ) } ~.\end{aligned}\ ] ] in eqs . ( [ fermion_nanotube ] ) and ( [ fermion_stm ] ) , we have omitted the klein factors which guarantee the anti - commutation of the 3 types of fermions operators written in terms of bosonic fields for this problem : the two nanotube branches and the stm single mode . it has been established @xcite that klein factors are in principle necessary to treat multi - luttinger system , as illustrated in the computation of noise correlations between three edge states in the fqhe . in the present work , klein factors can be dropped because we intend to work with lowest order perturbation theory . to order @xmath38 , statistical correlations between the three luttinger systems do not occur . however , they should show up when calculating higher order corrections ( @xmath39 ) . for this problem which implies propagation along the nanotube , it is also necessary to compute the ( total ) charge and ( total ) spin currents using the bosonized fields of eq . ( [ fermion_nanotube ] ) : @xmath40 similarly , we consider the spin current in the @xmath41 direction : @xmath42 note that the contribution from terms containing @xmath43 oscillations has been dropped . this is equivalent to requiring that the current measurement along the nanotube is effectively a spatial average over a length scale larger than @xmath44 . in practice , @xmath43 terms are necessary in order to establish a connection between current fluctuations and density fluctuations . in this section , the general approach used to calculate the tunneling current and noise , as well as the current and noise in the nanotube is described . all quantities are computed at zero temperature for simplicity . the calculation of the tunneling current and noise is quite similar to the perturbative results in ref . @xcite for the fqhe . here it is summarized in order to compare with the nanotube transport quantities . the keldysh technique is used to compute the average tunneling current and noise . we adopt the convention that the coefficients @xmath45 identify the upper / lower branch of the keldysh contour : @xmath46 which applies in typical tunneling situations where the product of the current averages is of order @xmath39 . in order to collect the lowest order contribution in the tunneling amplitude , the exponential is expanded to first order for the current , and to zeroth order for the noise : @xmath47 where the last factor in eqs . ( [ intermediate_tunneling_current ] ) and ( [ intermediate_tunneling_noise ] ) is the tip fermion green s function . next the nanotube and tip fields are specified in terms of the bosonized fields ( nonchiral and chiral ) , and the two keldysh ordered exponential products are computed : @xmath48 as expected , the stationary current and the real time current correlator call for the time differences @xmath49 , @xmath50 only . integrating over time , the zero frequency noise is introduced . further using the symmetry properties of the green s functions @xmath51 and @xmath52 ( similarly for @xmath53 , @xmath54 and @xmath55 ) , only @xmath56 is retained for the current : @xmath57 the tunneling current and noise imply the knowledge of the green s functions at the tunneling location only . the operator averages along the nanotube require a perturbative calculation up to second order in the tunneling hamiltonian for the tunneling current and for the noise . tunneling of an electron from the stm tip is followed by propagation of the collective excitations of the luttinger liquid towards both ends of the nanotube . @xmath58 where the contribution to the noise coming from @xmath59 has been dropped because it contributes to order @xmath39 . expressing the hamiltonian in terms of the fields , the limit @xmath60=\partial_x \phi_{c+}$ ] is used in order to cast the time ordered averages into correlators of exponentials only : @xmath61 where the contribution from the stm tip is the same as before . the two time ordered products ( one for the tip and one for the nanotube ) are expressed in terms of luttinger liquid green s functions . taking the spatial derivative , one obtains an expression with green s functions as prefactors implying propagation as well as exponentiated green s functions at the tunneling location . operating variable changes in the integrals and noticing that only @xmath62 contributes , the current and noise become : @xmath63 note the temporal decoupling ( which occurs after operating variable changes ) in these expressions . the integral over @xmath64 contains information on electron tunneling at @xmath24 , while the remaining integrals involve propagation , thus the spatial dependence in the green s functions arguments . in the previous section , general expressions were derived for the current and noise , which are independent of the form of the green s functions @xmath65 , @xmath66 , @xmath66 and @xmath67 . the green s functions are described in appendix a and are used to compute the tunneling noise and current as well as the nanotube noise and current . after substitution of the green s function of a nanotube , the tunneling current and noise read : @xmath68 with the exponent : @xmath69 @xmath70 is the bulk tunneling exponent of the current voltage characteristics @xmath71 the integrals are computed in appendix b , we obtain : @xmath72 where we used the definition of the gamma function @xmath73 . only electrons can tunnel from the tip to the nanotube , so one can check that the classical schottky formula holds always : @xmath74 some of the time integrals in eq . ( 22 ) has already been encountered when computing the tunneling current and noise . the current and noise thus become : @xmath75\nonumber\\ & & = -\frac{2e^2v^2_f\gamma^2}{\pi a}\left(\sum_{\sigma}\frac{1}{u_f^{\sigma}}\right ) \left(\frac{a}{v_f}\right)^{\nu}\frac{\|\omega_0\|^{\nu}}{{\bf \gamma}(\nu+1)}(i^{\phi\phi}(x , x')+i^{\phi\theta}(x , x'))~ , \nonumber\\\end{aligned}\ ] ] where the last factors are computed in appendix b. the standard assumptions of the calculation of the tunneling current and noise are recalled , as the same expressions appear in both results . we obtain : @xmath76 current conservation @xmath77 discussion ---------- one accepted diagnosis to detect effective or anomalous charges is to compare the noise with the associated current with the schottky formula in mind . a striking result is that despite the fact that electrons are tunneling from the stm tip to the bulk of the nanotube , the zero frequency current fluctuations are proportional to the current for @xmath78 : @xmath79 with an anomalous effective charge for an infinite nanotube . more can be learned from a measurement of the noise correlations . noise correlations have been proposed to detect statistical correlations in quantum transport @xcite . indeed , our geometry can be considered as a hanbury - brown and twiss correlation device . such experiments have now been completed for photons and more recently for electrons in quantum waveguides . here the novelty is that electronic excitations do not represent the right eigenmodes of the nanotube . for @xmath80 the noise correlations read : @xmath81 this is a priori negative . however , if the current direction is chosen to be positive from the tip to the extremities of the nanotube , the sign of the cross correlations is positive . recall that the fermionic version of the hanbury - brown and twiss experiment yields negative noise correlations @xcite . so far , positive noise correlations have been attributed in priority to bosonic systems @xcite . nevertheless , there are at least two other situations where they are encountered . first , when the source of particle is a superconductor , noise correlations can also be positive depending on the junction configuration@xcite . second , they also occur in systems with floating voltage probes @xcite . in the case of a superconductor , the emission of electron pairs through separate quantum dots guarantee that the noise correlations are always positive : a ( singlet ) entangled electron pair is generated outside the superconductor @xcite . note that the prefactors in eq . ( [ positive_noise_correlations ] ) can readily be interpreted using the language of ref . @xcite . a tunneling event to the bulk of a nanotube is accompanied by the propagation of two counter - propagating charges @xmath82 . recall that the subscript @xmath83 identifies the charge ( as opposed to spin ) excitation given by the total ( rather than relative ) contribution of the two modes propagating in the nanotube . each charge is as likely to go right or left . according to ref . @xcite electron injection in a luttinger liquid is characterized by chiral charges @xmath84 and chiral spin charges @xmath85 which describe the elementary excitations of the nanotube . @xmath86~,\ ] ] with integers @xmath87 ( @xmath88 ) . in particular , the addition of an electron with spin @xmath1 corresponds to the choice @xmath89 and @xmath90 . the current noise and noise correlations can be interpreted as an average over the two types of excitations : @xmath91 8 cm a drawing where the two types of charges `` flow away '' from the tip while propagating along the nanotube is depicted in the lower part of fig . both charges @xmath84 are equally likely to go right or left , and they are emitted as a pair with opposite labels . the noise correlations of eq.([correlations_quasiparticle ] ) are rendered positive if one adopts the standard convention for measuring the current in multi - terminal conductors @xcite . here these `` positive '' noise correlations resulting from charges moving toward both extremities of the nanotube have the added particularity that they occur to second order in a perturbative tunneling calculation . in superconducting - normal systems , the two electrons which emanate from the same cooper pair and which propagate in the two luttinger liquids provide a manifestation of the non local character of quantum mechanics . in the present case , only one electron is injected , but it is split into left and right excitations , unless one imposes one dimensional fermi liquid leads . here , we are dealing with entanglement between collective excitations of the luttinger liquid . written in terms of the chiral quasiparticle fields , the addition of an electron with given spin @xmath1 on a nanotube in the ground state @xmath92 gives : @xmath93 \label{fermion_nanotube_entanglement}\ ] ] with @xmath94 the chiral bosonic fields of the ( nonchiral ) luttinger liquid . this wave function is characterized by right and left movers @xmath8 whose fields appear explicitly in the phase operator of this many - particle wave function . these fields are independent of each other , therefore the exponential can be written as a product of fields : @xmath95 \label{entangled_wave_function}\ ] ] where for each sector ( charge / spin , total / relative mode ) the charges @xmath96 have been introduced , and chiral fractional operators are defined as : @xmath97~. \label{quasiparticle_chiral}\ ] ] the wave function described by eq . ( [ entangled_wave_function ] ) has all the characteristics of an entangled state . because the two types of excitations travel towards opposite ends of the nanotube , the time evolution of this `` injected electron '' state is simply obtained with the substitution @xmath98 . consequently , quantum mechanical non - locality is quite explicit here . the detection of a charge @xmath84 in one arm is necessarily accompanied by the simultaneous detection of a charge @xmath99 in the other extremity of the nanotube . this entanglement is the direct consequence of the correlated state of the luttinger liquid . when additional electrons are injected , these break up into the specific modes which can propagate in either direction in the nanotube . it therefore differs significantly from its analogs which use superconductors as electron injectors , where two electrons from the same cooper pair are dissociated . when considering only one sector , such as @xmath100 , it is interesting to note that the wave function has the same structure of say , a triplet spin state ( a symmetric combination of `` up '' and `` down '' states , or `` plus '' and `` minus '' charges ) for electrons , with the electrons being replaced by chiral quasiparticle operators . indeed , one has to recognize that each chiral field @xmath94 can be written as a superposition of boson operators : @xmath101 where @xmath102 creates a boson with nanotube mode @xmath3 , spin @xmath1 and momentum @xmath103 , and characterizes the collective modes of the one dimensional liquid . according to the state written in eq . ( [ quasiparticle_chiral ] ) , this linear superposition of boson operators appears in an exponential . this expresses that non - local `` many boson '' correlations are created when an electron is injected in a nanotube , and these many - body states are entangled in the present geometry . effects similar to the detection of effective charges show up in the spin sector when time reversal symmetry ( @xmath104 ) does not hold . the spin current and spin noise are obtained in a similar manner : @xmath105 so that at large distances : @xmath106 in practice , when time reversal symmetry holds ( @xmath107 ) , spin noise correlations vanish to order @xmath38 independently from the presence or the nature of the leads . in the case where the tip is non magnetized , the spin current and spin noise correlations also vanish . in the presence of one - dimensional fermi liquid leads , where the leads are considered to be luttinger liquids whose interaction parameters are set to @xmath108 , quasiparticles suffer andreev type reflections @xcite at both exterminites of the nanotubes . multiple reflections of quasiparticles in the fabry - perot geometry fermi liquid / nanotube / fermi liquid are expected to lead to a concellation of the interaction effects in the nanotube , as in the two terminal calculations of conductance and noise @xcite . although the detailed calculation is not presented here , dimensional analysis of the time integrals suggest that , the nanotube current and noise read : @xmath109 for @xmath110 , this whould give the classical schottky formula , in the very same spirit as in ref . for @xmath111 and @xmath112 on opposite ends of the nanotube , this noise correlator should vanish , to this order : the scattering theory result has a lowest non vanishing contribution of order @xmath39 . this low voltage result is modified by a higher power law behavior at higher voltage , with a threshold voltage specified by the size of the system @xmath113 as in ref . in summary , a diagnosis for detecting the chiral excitations of a luttinger liquid nanotube has been presented , which is based on the knowledge of low frequency current fluctuation spectrum in the nanotube . typical transport calculations either address the propagation in a nanotube , or compute tunneling i(v ) characteristics . here , both are addressed because they constitute the key for obtaining the quasiparticle charges . both the noise ( autocorrelation ) and the noise correlations ( cross - correlations ) are needed to identify the charges @xmath114 . independently , note that this measurement could also be confronted to other diagnoses of the nanotube interaction parameter using tunneling current voltage characteristics . this result relies on the assumption that one dimensional fermi liquid leads are avoided . such leads have been treated in different approaches @xcite , and are also labelled radiative contacts . radiative contacts imply equilibration with the electrons . in ref . @xcite , both radiative contacts and equilibration with dressed eigenmodes were studied , with the obvious result that luttinger liquid renormalization shows up in the conductance in the latter case . in special circumstances such as the case of ref . @xcite the nanotube is embedded in the metallic contacts , and it is suspended by its ends . here , the absence of a screening gate is explicit . electron transport between these two entities likely occurs in multiple electron scattering processes as studied in ref . @xcite . in these , or other contacts fabricated by growing techniques @xcite , quantities such as current and noise may not be affected by the presence of the contacts . standard fermion results should be recovered when the system is connected on one dimensional fermi liquid leads . the auto - correlation noise in one end of the nanotube should be related to the charge current with the standard schottky formula . the noise correlation signal should also vanish as expected and the next order correction @xmath115 then needs to be computed . a crucial test of the contacts is in order . it should be possible in practical situations to analyze the type of contacts which one has between the nanotube and its connections . if the ratio of the cross - correlations to the current @xmath116 does not depend on the tunneling distance ( @xmath117 ) , both contributions are of order @xmath38 and this constitutes an indication that the contacts do not affect this quasiparticle entanglement . if we are dealing with a fermi liquid behavior , the noise correlation current ratio should behave like @xmath38 , rather than a constant . finally , we have remarked that the many - body wave function which describes a luttinger liquid with an added electron has necessarily epr @xcite entangled degrees of freedom . both electrons chiralities contribute to the emission of quasiparticle pairs moving in opposite direction . this entanglement involves many particle states , unlike its electron counterpart . a suggestion for detection of such luttinger liquid entanglement without perturbing the system with leads is nevertheless needed . the issue how to detect this many - body entanglement should be addressed while taking into account different models for the leads , possibly involving multiple reflections within one contact @xcite . multiple reflections of the quasiparticles from one contact to the other kill this entanglement in one dimensional fermi liquid leads , which is implicit in the vanishing of the noise correlations to this order . at any rate , this is the first time that collective excitations entanglement is discussed in a condensed matter setting . discussions with a. lebedev , i. safi on the keldysh green s function and with m. bttiker are gratefully acknowledged . in this appendix , the green s functions are computed , assuming a luttinger liquid with an homogeneous interaction parameter @xmath21 and velocity @xmath22 . the product @xmath118 corresponds to the fermi velocity @xmath119 . the finite temperature action associated with this problem has the general form : @xmath120 which implies that the fourier transform of the time ordered green s functions @xmath121 and @xmath122 , define as : @xmath123 where t is the time ordered operator , satisfies the differential equations@xmath124 : @xmath125 the green s function @xmath121 is continuous everywhere , and @xmath126/k_{j\delta}$ ] has a discontinuity at @xmath110 . the similarity between eq . ( [ green_theta ] ) and ( [ green_phi ] ) results from the duality properties of the underlying fields . all information on @xmath122 is obtained by dividing @xmath121 by @xmath127 . according to eqs . ( 14 ) and ( 15 ) , there are additional green s functions in our problem which involve the fields @xmath128 and @xmath129 . for instance : @xmath130 using the action ( [ action ] ) one can show that ( @xmath131 is a real time variable ) : @xmath132 and similarly for @xmath133 . from these real time green s functions , we further specify the keldysh matrix elements which two times @xmath134 are assigned to the upper / lower branch ( @xmath135 ) . given an arbitrary real time green s function @xmath136 a general procedure@xmath137 for obtaining these elements is as follows : @xmath138 where : @xmath139 the same applies to @xmath140 for which we have : @xmath141 the mixed correlators read : @xmath142 where : @xmath143 the same applies to @xmath144 for which we have : @xmath145 we now compute the integrals involved in the tunneling current and noise . the general integrals which will be required to compute the current and noise read : @xmath146 we now write the integral which appears in the nanotube current , which refer to propagation along the nanotube : @xmath147 using the expressions for the green s functions ( appendix [ green_appendix ] ) : @xmath148 where the approximate sign holds at large distances . m. blanter and m. bttiker , phys . rep . * 336 * , 1 ( 2000 ) . c. kane and m.p.a . fisher , phys . lett . * 72 * , 724 ( 1994 ) ; c. de c. chamon , d.e . freed , and x.g . wen , phys . b * 51 * , 2363 ( 1995-ii ) ; p. fendley , a.w.w . ludwig , and h. saleur , phys . lett . * 75 * , 2196 ( 1995 ) . l. saminadayar , d. c. glattli , y. jin , and b. etienne , phys . 79 * , 2526 ( 1997 ) ; r. de - picciotto _ et al . _ , nature * 389 * , 162 ( 1997 ) . h. saleur and u. weiss , phys . rev . b * 63 * , 201302 ( 2001 ) . i. safi , p. devillard , and t. martin , phys . lett . * 86 * , 4628 ( 2001 ) . c. kane , l. balents , and m. p. a. fisher , phys . lett . * 79 * , 5086 ( 1997 ) . m. brockrath , d.h . cobden , j. lu , a.g . rinzler , r.e . smalley , l. balents , and p. mceuen , nature * 397 * , 598 ( 1999 ) . r. egger , a. bachtold , m.s . fuhrer , m. bockrath , d. cobden , and p. mceuen , in _ interacting electrons in nanostructures _ , edited by r. haug , and h. schoeller ( springer , 2001 ) . i. safi , ann . fr . * 22 * , 463 ( 1997 ) . k .- v . pham , m. gabay , and p. lederer , phys . b * 61 * , 16397 ( 2000 ) . w. schottky , ann . ( leipzig ) * 57 * , 541 ( 1918 ) . i. safi and h. schulz , phys . b * 52 * , 17040 ( 1995 ) ; d. maslov and m. stone , phys . rev . b * 52 * , 5539 ( 1995 ) ; v.v . ponomarenko , phys . b * 52 * , 8666 ( 1995 ) . roche , m. kociak , s. guron , a. kasumov , b. reulet , and h. bouchiat , eur . j. b * 28 * , 217 ( 2002 ) . t. martin and r. landauer , phys . b * 45 * , 1742 ( 1992 ) ; m. bttiker , phys . lett . * 65 * , 2901 ( 1990 ) ; phys . b * 46 * , 12485 ( 1992 ) . ya . m. blanter , f. hekking , and m. bttiker , phys . lett . * 81 * , 1925 ( 1998 ) . imura , k .- v . pham , p. lederer , and f. pichon , phys . b * 66 * , 035313 ( 2002 ) . b. trauzettel , r. egger , and h. grabert , phys . rev . lett . * 88 * , 116401 ( 2002 ) . s. chen , b. trauzettel , and r. egger , phys . lett . * 89 * , 226404 ( 2002 ) . pham , p. pichon , k .- imura , and p. lederer , cond - mat/0207294 . r. egger and a. gogolin , eur . j. b * 3 * , 781 ( 1998 ) . c. chamon and d.e . freed , phys . b * 60 * , 1842 - 1853 ( 1999 ) . r. guyon , p. devillard , t. martin , and i. safi , phys . rev . b * 65 * , 153304 ( 2002 ) . m. henny _ et al . _ , science * 296 * , 284 ( 1999 ) ; w. oliver _ et al . _ , science * 296 * , 299 ( 1999 ) . r. hanbury - brown and q.r . twiss , nature * 177 * , 27 ( 1956 ) . j. torrs and t. martin , eur . j. b * 12 * , 319 ( 1999 ) . t. gramespacher and m. bttiker phys . b * 61 * , 8125 ( 2000 ) . j. brlin , w. belzig , and c. bruder phys . 197001 ( 2002 ) . m. schechter , y. imry , and y. levinson phys . b * 64 * , 224513 ( 2001 ) ; f. taddei and r. fazio , phys . b * 65 * , 134522 ( 2002 ) . c. texier and m. bttiker phys . rev . b * 62 * , 7454 ( 2000 ) . p. recher , e.v . sukhorukov , and d. loss , phys . b * 63 * , 165314 ( 2001 ) . lesovik , t. martin , and g. blatter , eur . j. b * 24 * , 287 ( 2001 ) . p. recher and d. loss , phys . b * 65 * , 165327 ( 2002 ) . c. bena _ et al_. , phys . rev . 89 * , 037901 ( 2002 ) . v. bouchiat , n. chtchelkatchev , d. feinberg , g.b . lesovik , t. martin , and j. torrs , nanotechnology * 14 * , 77 ( 2003 ) . i. safi , eur . j. b * 12 * , 451 ( 1999 ) . r. egger and h. grabert , phys . * 79 * , 3463 ( 1997 ) ; phys . b * 58 * , 10761 ( 1998 ) . c. de c. chamon and e. fradkin , phys . rev . b * 56 * , 2012 ( 1997 ) ; v. ponomarenko and d. averin , jetp letters , * 74 * , 87 ( 2001 ) . a. einstein , b.podolsky , and n. rosen , phys . rev . * 47 * , 777 ( 1935 ) . i. safi , private communication .	transport through a metallic carbon nanotube is considered , where electrons are injected in the bulk by a scanning tunneling microscope tip . the charge current and noise are computed both in the absence and in the presence of one dimensional fermi liquid leads . for an infinite homogeneous nanotube , the shot noise exhibits effective charges different from the electron charge . noise correlations between both ends of the nanotube are positive , and occur to second order only in the tunneling amplitude . the positive correlations are symptomatic of an entanglement phenomenon between quasiparticles moving right and left from the tip . this entanglement involves many body states of the boson operators which describe the collective excitations of the luttinger liquid .
for the past two decades , charge transfer ( ct ) phenomena in biosystems have been intensively studied due to their role in biological functions as well as in potential applications related to nanosensors and molecular optoelectronics . oxidative damage in cells @xcite as well as the possibility of using dna as a biomolecular nanowire @xcite have inspired studies related to the hole migration through the dna nucleobases on a sequence of radical cation states . a major target for oxidants is guanine ( g ) , the nucleobase with the lowest ionization potential of the four dna bases . oxidation of guanine leads to a radical cation @xmath1 , the hole may transfer to a neutral @xmath2 , restoring neutrality in the former and creating a radical cation in the latter . especially the ct in dna oligomers has been extensively studied both experimentally and theoretically with many techniques . experiments have initially yielded different results , such as dna being a conductor @xcite , semiconductor @xcite , and insulator @xcite . also the theoretical understanding has been divergent and results in two points of view @xcite . one claiming that dna conduction occurs via a polaron picture by which the charge ( e.g. a hole ) is delocalized over a few nucleobases , and the resulting polaron hops along the dna @xcite . the other view is rooted in a conservative interpretation of the experiments by giese _ @xcite according to which charge transport over short distances occurs through superexchange - type conduction , while a long - range charge transfer is achieved by a combination of superexchange tunneling events as well as multistep hopping @xcite . in this interpretation , the excess charges are assumed to be localized on single nucleobases , and no polarons are invoked . recently @xcite , a hybrid mechanism that can not be classified as hopping nor as coherent superexchange has been theoretically proposed for the long range ct in dna . incoherent mechanisms are also significant in dna charge transport @xcite . however , in this work we will only consider coherent transport . decoherence can be taken into account theoretically by carrying out a simultaneous quantum dynamics of the electrons and the nuclei @xcite , by ad - hoc corrections of the kinetic constant expression @xcite or , at least partially , by accounting for the influence of a dissipative environment ( phonon bath ) which can support or restrict charge motion @xcite . theoretically , it has been difficult to provide a clear picture of dna conductivity due to charge transfer dynamics being inherently non - adiabatic , and its modeling must involve going beyond the born - oppenheimer approximation . this was attempted using model hamiltonians @xcite as well as full electron - nuclear non - adiabatic dynamics simulations @xcite . however , the excess charge localization is dramatically affected by the specific theory employed for handling non - adiabaticity ( e.g. ehrenfest or surface hopping ) @xcite . charge dynamics carried out with model hamiltonians has been successful in elucidating certain regimes of ct in dna . however , it has not offered a breakthrough . as an example , depending on the specific hamiltonian used , the modeled long - range charge dynamics may @xcite or may not @xcite agree with the experimental observations and generally only semiquantitative agreement with the experiments has been obtained . it is generally recognized that the key to obtaining even semiquantitative agreement with the experiment is the accounting of structural fluctuations in the calculation of the parameters of the model hamiltonians @xcite , as well as a proper accounting of the effect of the counterstrand on the nucleobases ionization potentials @xcite . the effect of polarization by the molecular environment surrounding a nucleobase on its ionization potential has been reported to be up to 0.4 ev @xcite . thus , accounting for these environmental effects is a major component of the modeling of charge transport in biological systems . current computational methods do not allow the fully quantum mechanical description of a biological molecule , such as solvated dna , as the computational scaling of even mainstream density - functional theory ( dft ) methods goes roughly as @xmath3 , with @xmath4 being the number of atoms in the system . as the scale of model systems considered in theoretical simulations increases , the computational complexity quickly escalates to unreachable cpu times , even by the most powerful supercomputer on earth . thus , theoreticians are forced to make approximations ( such as qm / mm treatments ) to recover the environmental effects that are so important for the correct description of ct phenomena in biosystems . approximate qm / mm methods are not ideal , as they need semiempirically determined parameters to function . in addition , even if the most accurate polarizable force field available is used @xcite the wavefunction of the electrons belonging to the environment is completely lost , and with that typical quantum mechanical effects , such as electronic exchange and correlation interactions with the environment are generally not contemplated . one of us has recently contributed to the development of an accurate and linear - scaling dft method that allows to include environmental effects in the ionization potentials ( site energies ) as well as in the electronic hamiltonian coupling matrix elements @xcite by exploiting ideas and techniques of subsystem dft @xcite . we will refer to this method as frozen density embedding ( fde ) throughout this paper . this method has allowed us to approach system sizes that were simply not reachable before , especially for calculations of electronic couplings . this work focuses on dna and specifically on the effects of the molecular environment surrounding the nucleobases ( particularly sugar groups and partner strand ) involved in the hole transfer . besides the solvating effects of the water molecules surrounding the dna @xcite , the environment affects the hole transfer in dna in several ways . for example , we will show that the sugar groups and the counterstrand stabilize the hole * unevenly * when it is localized on different nucleobases ( e.g. g is stabilized differently from t or a , also the 3@xmath5 position is stabilized differently from 5@xmath5 ) . counterions also play a role . coupling of the hole motion with the motion of the counterions has been considered in the literature , and ion - gated charge transport has been proposed @xcite . because in this work we do not consider dynamical effects , we can not infer on the effects of the counterions . we instead focus on completely desolvated dna model systems employing an accurate and all - electron electronic structure method : the frozen density embedding ( fde ) formulation of subsystem dft . in the most computationally demanding simulation , we calculated site energies and hamiltonian couplings for the hole transfer in a dna pentamer model system containing 308 atoms and 1322 interacting electrons . experimental works report the so - called decay factor @xmath0 by fitting the charge transfer kinetics against the marcus equation , which is derived from the fermi golden rule electron transfer rate equation @xcite , namely @xmath6 where the frank - condon coefficient that within the marcus theory is approximated as @xcite : @xmath7 where @xmath8 is the energy gap between the the initial and final state of electron transfer , @xmath9 is the reorganization energy , and @xmath10 is the boltzman constant . the above equation is appropriate for the superexchange regime @xcite , i.e. when donor acceptor energy levels are non - resonant to the bridge levels . under those energetic conditions , the kinetic constant takes the following form , @xmath11 where @xmath12 is an effective donor acceptor separation distance . kinetic constants in the hopping regime , instead , follow a power law decay @xcite . in this work , we calculate effective donor acceptor hole transfer couplings and fit them to [ m - l - j ] . we first focus on simple donor - acceptor systems , such as @xmath13-stacks dimers of nucleobases guanine ( @xmath2 ) , thymine ( @xmath14 ) , adenine ( @xmath15 ) . we also consider several other biologically relevant molecules , such as histidine , tyrosine , tryptophan , as well as other molecules rich in @xmath13 electrons that may be active in biological hole transfer . this first set of calculations will allow us to establish that fde can calculate electronic couplings and site energies for biologically relevant dyads over a wide range of intermolecular separations . this is not a trivial achievement , as when the molecular fragments are separated by large distances , numerical inaccuracies can creep in and undermine the fidelity of the calculations . we propose effective solutions to such inaccuracies so that the calculated couplings are numerically stable over a wide range of intermolecular separations . once the simple donor - acceptor couplings are presented , we also treat superexchange in the b - dna sequences and with @xmath160 - 3 , where the hole transfer is restricted to run on one strand . we analyzed the influence of the presence of the second strand and the sugar groups on the couplings , site energies and the overall decay factor , @xmath0 , in the superexchange regime . regular dft is also known as kohn sham dft ( ks - dft ) , and can be summarized by the following equation , the ks equation in canonical form , @xmath17 \phi_k(\br ) = \varepsilon_k \phi_k(\br),\ ] ] where @xmath18 is the effective potential that the one - particle ks orbitals , @xmath19 , experience , and @xmath20 are the ks orbital energies . the spin labels have been omitted for sake of clarity . the electron density for singlets is simply @xmath21 . the effective potential , @xmath18 , is given by @xmath22 where @xmath23 is the electron nucleus attraction potential , @xmath24 the hartree potential , and @xmath25 the exchange correlation ( xc ) potential@xcite . subsystem dft , instead , is based on the idea that a molecular system can be more easily approached if it is partitioned into many smaller subsystems . in mathematical terms , this is done by partitioning the electron density as follows @xcite @xmath26 with @xmath27 being the total number of subsystems . self - consistent solution of the following coupled ks - like equations ( also called ks equations with constrained electron density @xcite ) yields the set of subsystem ks orbitals , i.e. @xmath28 \phi^i_k(\br ) = \varepsilon^i_k \phi^i_k(\br),\mathrm{~with~}i=1,\ldots , n_s\ ] ] with the effective subsystem potential given by @xmath29 in fde @xcite , @xmath30 appearing above is called embedding potential and is given by @xmath31 + \\ & + \frac{\delta t_{\rm s}[\rho]}{\delta\rhor}-\frac{\delta t_{\rm s}[\rhoi]}{\delta\rhoir}+\frac{\delta e_{\rm xc}[\rho]}{\delta\rhor}-\frac{\delta e_{\rm xc}[\rhoi]}{\delta\rhoir}.\end{aligned}\ ] ] where @xmath32 , @xmath33 and @xmath34 are kinetic and exchange - correlation energy functionals , and the nuclear charge , respectively . at a first glance , the presence of the coulomb potential in the above equation suggests that the method is not linear - scaling . however , two - electron integrals are never evaluated in our method as we employ slater - type orbitals for which no analytical two - electron integral formula is known . instead , an appropriate choice of fitting functions for the calculation of the coulomb potential makes this integral easy to compute@xcite . two additional simplifications are present in the fde implementation : first , fde uses subsystem - centered integration grids ( significantly reducing the intgration time ) , and secondly , as the density of the non - active subsystems remain frozen during a subsystem s scf procedure , their coulomb potential is obtained once at the beginning of the scf procedure and it is stored in memory . ultimately , it is this that makes the method linear - scaling with respect to the number of subsystems@xcite . the density of the supersystem is found using [ fde1 ] and [ ks1 ] , for singlets @xmath35 . the above equations have not explicitly taken into account spin , however the full subsystem local spin density approximation equations can be retrieved elsewhere @xcite . it has been shown @xcite that fde generates charge - localized states if one subsystem density is constrained to integrate to a number of electrons different from what would yield charge neutrality . in this way , charge - localized broken symmetry states can be constructed by localizing an excess charge on a single molecular fragment ( subsystem ) . this level of localization can also be achieved by the constrained dft method of van voorhis @xcite . however , as opposed to the linear - scaling fde method , constrained dft scales identically to regular ks - dft , i.e. as @xmath36 . fde can construct an electron density featuring a hole on the donor ( with corresponding `` initial '' wavefunction @xmath37 ) and one density where the hole is localized on the acceptor ( with corresponding `` final '' wavefunction @xmath38 ) . these states are not the solution of the full ks - dft equations , thus a non - diagonal hamiltonian matrix should be expected in the basis of @xmath37 and @xmath38 . the off - diagonal elements of such hamiltonian can be approximated by the following formula @xcite : @xmath39.\ ] ] here @xmath40 is the molecular electronic hamiltonian , and @xmath41 is the transition density @xmath42 , with @xmath43 being the total number of electrons in the system . the donor acceptor overlap matrix elements are found by computing the following determinant : @xmath44,\ ] ] where @xmath45 is the transition overlap matrix and it is defined in terms of the occupied orbitals ( @xmath46 ) making up the determinants @xmath37 and @xmath38 @xcite @xmath47 the hamiltonian coupling is not @xmath48 , but it is generally reported as the coupling between the lwdin orthogonalized @xmath37 and @xmath38 . for only two states this takes the form , @xmath49 for more information about the fde formalism applied to charge transfer states , we refer the reader to refs .. in a previous work @xcite , it has been shown that the fde method is comparable in accuracy with _ ab - initio _ methods for the calculation of charge transfer excitation energies for the test cases considered in that work . generally speaking , semiempirical methods are extremely powerful and computationally cheap . however , we should mention that in a recent work@xcite these methods , in particular dft - based semiempiricals as the density functional tight binding ( dftb ) , produce an error of the electronic coupling around 40% which reflects an overestimation of @xmath0 values by around 12% ( see for example tables vi - xii in adam et al.@xcite ) . regarding the computational cost , we mention that fde scales linearly with the number of subsystems , which for a full - electron electronic structure method is a great advantage . all calculations ( effective electronic couplings , excitation energies , and site energies ) were performed with a modified version of the amsterdam density functional ( adf ) package @xcite . gga semilocal functionals , pw91 and pw91k , were used throughout for the exchange correlation @xcite and kinetic energy @xcite functionals , respectively . this combination of functionals ( xc and kinetic parts ) is known to reproduce @xmath13-stacking interaction energies in a subsystem dft environment @xcite . we used the tzp slater type orbital basis set throughout this work , except when noted . the molecular systems chosen for the simulations are the following dimer combinations : dna nucleobase ( adenine , guanine and thymine)dna , dna aminoacid ( histidine , tyrosine , tryptophan ) , aminoacid aminoacid and aromatic ( benzene , anthracene , indole , porphyrin , phenyl - porphyrin)aromatic . finally , all dna oligomer structures were obtained with the nab program from ambertools package @xcite in the standard b - form . in this section we present calculations of the coupling matrix element ( @xmath50 ) of hole transfer from a donor to an acceptor molecule through the vacuum . this means that the initial state of hole transfer is the donor molecule ( @xmath51 ) , and the final state the acceptor molecule ( @xmath15 ) , and no intermediate bridge states are considered . for this purpose , we chose 23 biologically relevant @xmath13-stacks dyads , and [ forsys ] shows a representative subset of them . in order to analyze the distance dependence of the coupling , donor - acceptor separations of 3 - 20 are considered and 276 coupling calculations were ran in total . in tables s1s4 of the supplementary information , we collect the results for the donor acceptor overlap , coupling , and site energies at variable donor acceptor separations for the 23 dyads . the results show that at long ranges our couplings have a linear relation with the diabatic overlaps ( see [ ag ] ) . this behavior is expected , as it is reported in the literature@xcite , where the electonic coupling and the overlap are found to be proportional to each other and the @xmath13-electron binding energy enters the proportionality constant . the fde calculations were carried out similarly to refs . . in short , first the density of the two charge - localized states ( either a hole on the donor or on the acceptor ) are obtained by self - consistent fde calculations ( i.e. freeze and thaw@xcite ) . secondly , the total energy and hamiltonian coupling among the resulting diabatic states are calculated in a post - scf analysis according to [ coulow ] . the calculated couplings and the excitation energies ( obtained solving the @xmath52 eigenvalue problem involving the hamiltonian and the overlap matrix elements ) of dna nucleobases are compared against benchmark calculations @xcite and calculations carried out with the previous version of the code @xcite , see [ tabcom ] . the purpose of the set of donor acceptor calculations is to test the numerical stability of our method for calculating long - range couplings which are at the base of coupling calculations in biosystems . in carrying out these initial tests , we identified two weaknesses of the previous implementation of our method which we have now completely cured . one was related to the density fitting routine , and one related to the overlap matrix inversion of [ ivth ] . let us first discuss the new density fitting routine . we have implemented in our code a new density fitting method @xcite for evaluating the coulomb potential associated with the transition density in [ ivth ] , we now use a spline - based fit for the radial part of the density , and sets of atom - centered spherical harmonics for the angular dependence ( zlm fit , hereafter ) . zlm fit offers the possibility to calculate the coulomb potential to very high precision @xcite . .hole transfer excitations and couplings for @xmath13-stacked dna nucleobases at 3.38 @xmath53 separation . all values in ev . [ cols="^,^,^,^,^,^,^",options="header " , ] the long range behavior of the couplings was characterized by a fit against the exponential decay law in [ j - m - l ] . the decay parameter , @xmath0 , is defined , according to [ marcus ] , in terms of a fit of the distance dependence of the coupling to the following exponential function : @xmath54 by taking the natural logarithm , [ v_vlm ] takes a linear form : @xmath55 as we can see in the supplementary information figures s1s4 and in [ tabbeta ] , the electronic couplings decay exponentially as expected @xcite , and the fits to the exponential function deliver excellent standard deviations for the fit parameters . turning to the excitation energies in tables s1s4 , we notice that they are within the thermally accessible range of the donor - acceptor ionization potential difference . for instance , let us consider the ag and aa systems . in table s1 , these two systems display a similar excitation energy . this can be explained by considering that the ionization potentials of the involved nucleobases are similar ( ip@xmath56 ev and ip@xmath57 ev @xcite ) , with a difference of only about @xmath58 ev indicating that the excitation energies expected at room temperature ( considering a maximum of 0.2 ev fluctuation @xcite ) should range between 0 - 0.4 ev for ag and 00.2 ev for aa . vs. @xmath59 in a logarithmic scale . color online . , scaledwidth=100.0% ] despite of the good fit to the exponential decay law , the calculated decay factors are not comparable to the experimentally determined ones ( @xmath60 for dna nucleobases @xcite , and @xmath61 for aminoacids @xcite ) . such an overestimation of the decay factors by our calculations is expected , as the hole in this first set of calculations tunnels from donor to acceptor through the vacuum . in the experiments , however , the tunneling barrier is always reduced from the vacuum level by the existence of other molecules populating the space in between donor and acceptor . these molecules are often called `` bridge '' and the resulting tunneling is termed bridge - mediated tunneling or superexchange . in the following sections , we will evaluate bridge - mediated superexchange couplings in two dna oligomers employing the effective two - state approximation . however , the two - state approximation is bound to break down when near - resonant bridge states are present in the oligomer . this section presents calculations of electronic couplings of guanine - to - guanine hole transfer in the dna oligomer , with @xmath160 - 3 . we will use a shorthand notation , hereafter , for this system . by including in the computations increasingly complex molecular environments , the effect of the surrounding can be determined by analyzing the trend in the decay factor , @xmath0 , when going from interacting nucleobases of a single strand to a dephosphorilated b - dna that includes the counter strand and the sugar groups . the starting point of the calculations is a completely dry b - dna structure of . the structures considered lack water molecules , metal counterions and phosphate linker groups . this is because the applicability of fde is restricted to non - covalently bound molecular fragments . consequently , appropriate modifications to the b - dna structure had to be made : we have removed the phosphate groups and capped the dangling bonds with hydrogen atoms at 1.09 from the bonding atom . the resulting structure of the modified is depicted in [ orbostrand ] . it is important to point out that the calculations were carried out with a fixed nuclear structure , i.e. thermal dissipation , as well as vibronic effects were not contemplated at all in this work , and will be the focus of a future study . the bridge separating donor and acceptor assists the hole transfer by lowering the effective tunneling barrier . this effect can be taken into account by considering the full hamiltonian and overlap matrices of the hole pseudoparticle defined in [ cou][eq : s ] . namely , @xmath62 in the matrices above , a clear distinction between matrix elements between states where the hole is localized on the bridge molecules or on the donor and acceptor molecule has been labeled . taking into account that during the hole tunneling the bridge states are virtually occupied by the hole , an effective coupling can be obtained by reducing the generalized eigenvalue problem constructed with the above hamiltonian and overlap matrices to a 2@xmath632 effective eigenvalue problem @xcite . a lwdin orthogonalization of the basis set yields a transformed hamiltonian matrix , @xmath64 , from which the following bridge - mediated effective hole coupling is derived @xcite : @xmath65 where the superscript @xmath14 stands for transpose , @xmath66 is the green s operator , defined as @xmath67 and @xmath68 is the row vector of the transformed hamiltonian collecting the couplings between the donor / acceptor with the bridge states . generally , @xmath69 appearing above is the energy at which the tunneling event occurs ( i.e. at the crossing seam of the marcus parabolas ) . as we only considered static geometries of the dna oligomers , the hamiltonian eigenvalues corresponding to the donor - acceptor energies do not coincide . with that , the tunneling energy , @xmath69 , is not well defined @xcite . a natural choice of @xmath69 is to place it between @xmath70 and @xmath71 , with a common choice being @xmath72 . for example , this choice is invoked by several works in the literature @xcite . others@xcite , have favored the choice @xmath73 , which is non - symmetric ( i.e. forward ct become not equivalent to backward ct ) , however , it is still a valid choice . it was beratan @xcite and marcus @xcite who showed early on the weak dependence of the coupling with respect to the choice of tunneling energy within @xmath74 and @xmath75 . thus , we adopt @xmath76 in all calculations . in the following section , we will use [ efcou ] to estimate the hole superexchange coupling matrix element in dna oligomers . it is convenient to consider the two components of the coupling separately : the component related to the tunneling through the vacuum [ @xmath77 in [ efcou ] also called `` through space '' coupling ] , and the superexchange contribution from the bridge states ( @xmath78 ) . this distinction offers insight to the factors influencing the overall superexchange coupling . however , it is known that [ efcou ] is valid only when the donor , bridge and acceptor states are weakly mixed . hatcher et al . @xcite found that if a is present between two gs , there is a high chance that dynamical fluctuations will lead to a situation where the energy levels of a are in resonance with g , thus , undermining the two - state model and the validity of [ efcou ] ( formally valid only for non - resonant tunneling through the bridge ) . non - resonant tunneling is indeed the case for g(t)@xmath79 g . for g(a)@xmath79 g , near degeneracies of the donor / acceptor bridge system can arise ( see subsection 4.2 and figure 5 ) . for the particular conformations that we adopt in this study , we indeed encountered near - degeneracies for g(a)@xmath80 g and g(a)@xmath81 g oligomers . c , acts as a solvating environment ( in yellow , labeled as `` spectators '' ) and no hole is allowed to localize on it.,scaledwidth=80.0% ] in this section , we present and analyze the distance decay rates as well as the effect of the molecular environment on the effective hole transfer coupling @xmath50 for four model systems : single / double strand , and with / without ribose groups . when the ribose groups are included in the calculations , they are attached covalently to the nucleobases . this effectively increases the size ( in terms of number of atoms ) of all the participating charge transfer states . the largest system considered is the double strand with ribose groups and counts 308 atoms and 1322 electrons . a summary of the coupling values is given in [ tabcm ] . lrrrr & @xmath77 ( mev ) & @xmath78 ( mev ) & e@xmath82 ( ev ) & e@xmath83 ( ev ) + + gg & 78.13 & & & + gtg & 0.76 & 12.46 & 0.71 & 0.50 + g(t)@xmath80 g & 0.01 & 1.13 & 0.79 & 0.66 + g(t)@xmath81 g & & 0.09 & 0.79 & 0.77 + + gg & 92.6 & & & + gtg & 0.65 & 7.66 & 0.93 & 0.96 + g(t)@xmath80 g & 0.01 & 0.47 & 1.11 & 0.94 + g(t)@xmath81 g & & 0.02 & 0.99 & 1.16 + + gg & 71.38 & & & + gtg & 0.18 & 25.01 & 0.43 & 0.37 + g(t)@xmath80 g & 0.02 & 1.70 & 0.58 & 0.37 + g(t)@xmath81 g & & 0.21 & 0.41 & 0.41 + + gg & 91.07 & & & + gtg & 0.02 & 7.35 & 0.62 & 0.87 + g(t)@xmath80 g & 0.02 & 0.61 & 0.93 & 0.60 + g(t)@xmath81 g & & 0.02 & 0.50 & 0.82 + it is interesting to notice that the two components to the overall coupling , @xmath84 and @xmath78 , follow an inverted trend . let us first analyze the embedding effects due to the ribose groups on the direct @xmath84 couplings . it is arguable that the presence of any environment induces some slight charge delocalization in the adiabatic states simply because additional quantum states become accessible . however , we see that the ribose groups have the effect of localizing further the hole s wavefunction to the nucleobases in turn decreasing the direct nearest - neighbor coupling . we notice such trend for all the oligomers . the non - nearest - neighbor couplings are only marginally affected ( differences of a few tenths of mev ) . the picture is inverted for the superexchange component of the coupling , @xmath78 . there , we notice that the ribose groups increase this coupling term consistently . we rationalize these contrasting trends by generalizing that the through - space effect is due to a further localization of the hole s wavefunction , while the through - bridge effect is due to an uneven solvation by the ribose of the g and t nucleobases and due to the h - bonding interaction with the partner strand . we remind the reader that the effects related to the phosphates groups are not contemplated here , as these groups were not included in the structures used for coupling and site energy calculations . a closer inspection of the data in [ tabcm ] reveals that the double strand environmental effects overpower completely the effects due to the ribose groups . this is an important finding , as it indicates that the wavefunction and energetics of the hole in dna is almost completely determined by the interactions with the counterstrand rather than by the interactions with the covalently bound ribose groups . we will further analyze this aspect later on when considering the decay factors , @xmath0 . the full picture of the environmental effects on the hole - transfer couplings is captured by [ gtgbeta ] . there , the logarithmic decay of the effective couplings with the donor acceptor distance is plotted . the trends are explained by inspecting the relative energy levels of donor / acceptor and the bridge states in the single strand and in the double strand oligomers presented in [ etun ] . the figure shows that the energy levels of the bridge states change in going from the 5@xmath5 to the 3@xmath5 side of the dna oligomer . this behavior has been characterized before both experimentally @xcite and theoretically @xcite ( although using semiempirical methods ) . our calculations reproduce the experimental trend that the nucleobases in 3@xmath5 ( right hand side in [ etun ] ) have lower ionization potential than the same nucleobases placed at the 5@xmath5 position @xcite . for with @xmath160 - 3,scaledwidth=100.0% ] considering the average tunneling barrier height , calculated by taking the average of all energy differences between each bridge state with d / a states ( see [ tabcm ] ) , a dependency of the electronic coupling with the barrier height is evident . the taller the barrier , the larger the @xmath0 . this result is in line with the concept that the square of the coupling is related to the transmission probability of tunneling through the wall of potential energy separating the donor and the acceptor @xcite . position ) was taken as the zero of energy.,scaledwidth=100.0% ] to further characterize the solvating effect of the dna backbone on the electronic couplings in , we included the deoxyribose groups on all the nucleobases present in the model system ( including the spectator counterstrand ) . as before , we analyzed the differences in the hole transfer parameters for the single and double stranded cases . [ gtgbeta ] and [ tabcm ] show that the presence of the sugar groups does not change the picture significantly and we recover a similar trend of couplings and site energies as for the structures without sugars . however , from [ tbeta ] we notice that while the couplings in the single stranded are affected by the presence of the sugar groups ( @xmath0 values of 1.34 to 1.19 , respectively ) , the double stranded system is not appreciably affected by the sugar groups . our calculations show that while the site energies are generally lowered by the presence of the sugars , the stabilizing effect of the counterstrand and the effect of the sugar groups on the localization of the hole wavefunction have opposite effects on the coupling . this results in a small variation of the calculated couplings when the sugar groups are included . the calculated @xmath0 values are in broad agreement with the accepted values for dna , however overestimate the measurements of giese et al . @xcite ( @xmath85 ) . lrr & @xmath0 & @xmath86 + + single strand & 1.34 & 0.08 + double strand & 1.63 & 0.05 + single strand with sugars & 1.19 & 0.14 + double strand with sugars & 1.64 & 0.09 + + single strand & 1.01 & 0.13 + double strand & 1.56 & 0.33 + lrrrr & @xmath77 ( mev ) & @xmath78 ( mev ) & e@xmath82 ( ev ) & e@xmath83 ( ev ) + + gg & 78.13 & & & + gag & 0.67 & 5.43 & 0.32 & 0.17 + g(a)@xmath80 g & 0.01 & 1.43 & 0.37 & 0.15 + g(a)@xmath81 g & & 0.44 & 0.22 & 0.29 + + gg & 92.6 & & & + gag & 0.69 & 0.78 & 0.35 & 0.25 + g(a)@xmath80 g & 0.01 & 0.07 & 0.32 & 0.24 + g(a)@xmath81 g & & 0.04 & 0.27 & 0.28 + for with @xmath160 - 3,scaledwidth=100.0% ] in this work , we also analyzed the system with @xmath160 - 3 , for single and double stranded dna . although , as we mentioned above the dna oligomers with bridging adenines invalidate the two - state approximation , when we apply [ efcou ] we assume a non - resonant behavior between guanines and adenines which is an approximation . borrowing from the system that the effect of the sugars is not significant , for this final test we omitted the calculations with the ribose groups . all calculations were carried out in the same way as the systems , however , we used the transition overlap inversion threshold of @xmath87 for the simulations involving @xmath88 and @xmath89 pairs as we had determined these thresholds in the corresponding calculations carried out in the vacuum ( section [ vac ] ) . [ tab_ade ] shows that , once again , the couplings follow an exponential decay rate when the distance is increased , and similar to , single strands sequences have lower tunneling wall height which results in a lower @xmath0 ( see [ allade ] ) . the @xmath0 values of 1.01 @xmath53 and 1.57 @xmath53 ( see [ tbeta ] ) are in agreement with experimental data @xcite . turning to the energy levels , we see on [ etun ] that the second strand system features bridge energy values decaying much faster than the single strand ones . the general trend of the site energies for the bridge is in agreement with previous theoretical predictions @xcite , however , the near degeneracy experienced by the 3@xmath5-most adenine could not be predicted and it is likely the result of the specific nuclear geometry chosen in this work . we believe this near - degeneracy is the reason for the fact that the coupling does not follow an exact exponential decay law in [ allade ] . this is problematic and may undermine the applicability of [ efcou ] as previously discussed by hatcher et al.@xcite . we also notice that the uneven stabilization of the bridge states and donor / acceptor states is much more pronounced in the system than in the system . after inspection of the overall electrostatics of the interaction between g : c and t : a @xcite , we notice that t has a strong permanent dipole pointing towards a , similarly to c : g . instead , a has a much weaker dipole compared to c or t and thus upon contact of the gtg strand with the cac strand the cytosines will stabilize much more the holes on gs than the adenines can stabilize the holes on ts , hence the tunneling wall increases from single strand to double strand . for the system this effect is much dampened by the fact that donor , acceptor , and bridge states are similarly solvated by the counterstrand . characterizing the effects of the molecular environment on the through space and through bridge hole transfer couplings in biological systems is currently out of reach of standard all - electron electronic structure methods due to the large system sizes needed in the simulations . subsystem dft offers a way to include the environmental effects from first principles with no need to parametrize the interactions between subsystems . in this work , we have showed that the frozen density embedding formulation of subsystem dft is capable of tackling biosystems of realistic sizes , such as double stranded dna pentamers . the simulations focused on the hole transfer couplings of 23 dyads relevant to biology as well as two dna oligomers , and . while the calculations on the dyads were carried out for benchmark purposes , and are fairly standard not revealing any unexpected results . the calculations on the dna oligomers , instead , uncovered new paradigms related to the interactions of the hole in dna with its molecular environment . our calculations of hole transfer in dna reproduce experimental findings regarding the preference of the hole for the 3@xmath5 position rather than the 5@xmath5 . in agreement with simple arguments based on electrostatics , our calculations show that the bridge states ( thymines ) in experience a dramatically different environmental effect than the donor and acceptor ( guanines ) . again , in accordance with the simple electrostatic picture , the same effect is not noticed in the system . when the through space and through bridge couplings are inspected , our calculations show that the effects of the ribose groups and the nucleobases in the counterstrand are opposite and different in magnitude depending on the oligomer size . we conclude , however , that the effect of the counterstrand completely overpowers any effect due to the presence of the ribose groups . our calculated decay factors ( @xmath0 ) feature excellent standard deviation , and satisfactory agreement with the experimentally determined ones . the major limitation of our calculations rests in the absence of nuclear dynamics . as dynamics plays a major role in modulating the couplings and energies in biological hole transfer , we commit to investigate such dynamical effects in a follow up work . we thank the nsf - xsede program ( award tg - che120105 ) for computational resources , and the office of the dean of fasn of rutgers - newark for startup funds . we also thank agostino migliore and alexander voityuk for enlightening discussions . this information is available free of charge via the internet at http://pubs.acs.org/. genereux , j. c. ; barton , j. k. _ chem . rev . _ * 2010 * , _ 110 _ , 16421662 nunez , m. e. ; hall , d. b. ; barton , j. k. _ chem . biol . _ * 1999 * , _ 6 _ , 85 ratner , m. _ nature _ * 1999 * , _ 397 _ , 480481 kasumov , a. y. ; kociak , m. ; gueron , s. ; reulet , b. ; volkov , v. t. ; klinov , d. v. ; bouchiat , h. _ science _ * 2001 * , _ 291 _ , 280 porath , d. ; bezryadin , a. ; de vries , s. ; dekker , c. _ nature _ * 2000 * , _ 403 _ , 635638 braun , e. ; eichen , y. ; sivan , u. ; ben - yoseph , g. _ nature _ * 1998 * , _ 391 _ , 775778 venkatramani , r. ; keinan , s. ; balaeff , a. ; beratan , d. n. _ coord . chem . rev . _ * 2011 * , _ 255 _ , 635648 conwell , e. m. _ proc . natl . usa _ * 2005 * , _ 102 _ , 87958799 giese , b. _ acc . _ * 2000 * , _ 33 _ , 631636 giese , b. ; amaudrut , j. ; kohler , a .- k . ; spormann , m. ; wessely , s. _ nature _ * 2001 * , _ 412 _ , 318320 lewis , f. d. ; liu , j. ; weigel , w. ; rettig , w. ; kurnikov , i. v. ; beratan , d. n. _ proc . usa _ * 2002 * , _ 99 _ , 1253612541 lewis , f. d. ; wu , t. ; zhang , y. ; letsinger , r. l. ; greenfield , s. r. ; wasielewski , m. r. _ science _ * 1997 * , _ 277 _ , 673676 lewis , f. d. ; wu , t. ; liu , x. ; letsinger , r. l. ; greenfield , s. r. ; miller , s. e. ; wasielewski , m. r. _ j. am . _ * 2000 * , _ 122 _ , 28892902 renaud , n. ; berlin , y. a. ; lewis , f. d. ; ratner , m. a. _ j. am . soc . _ * 2013 * , _ 135 _ , 39533963 troisi , a. ; nitzan , a. ; ratner , m. a. _ j. chem . phys . _ * 2003 * , _ 119 _ , 57825788 skourtis , s. s. ; waldeck , d. h. ; beratan , d. n. _ annu . rev . chem . _ * 2010 * , _ 61 _ , 461485 , j. _ faraday discuss . _ * 1998 * , _ 110 _ , 407419 landry , b. r. ; subotnik , j. e. _ j. chem . phys . _ * 2011 * , _ 135 _ , 191101 gutirrez , r. ; caetano , r. a. ; woiczikowski , b. p. ; kubar , t. ; elstner , m. ; cuniberti , g. _ phys . lett . _ * 2009 * , _ 102 _ , 208102 kubar , t. ; elstner , m. _ phys . phys . _ * 2013 * , _ 15 _ , 57945813 grozema , f. c. ; berlin , y. a. ; siebbeles , l. d. a. _ int . j. quantum chem . _ * 1999 * , _ 75 _ , 10091016 grozema , f. c. ; berlin , y. a. ; siebbeles , l. d. a. _ j. am . _ * 2000 * , _ 122 _ , 1090310909 grozema , f. c. ; tonzani , s. ; berlin , y. a. ; schatz , g. c. ; siebbeles , l. d. a. ; ratner , m. a. _ j. am . * 2008 * , _ 130 _ , 51575166 senthilkumar , k. ; grozema , f. ; guerra , c. ; bickelhaupt , f. ; lewis , f. ; berlin , y. ; ratner , m. ; siebbeles , l. _ j. am . soc . _ * 2005 * , _ 127 _ , 1489414903 kubar , t. ; woiczikowski , p. b. ; cuniberti , g. ; elstner , m. _ j. phys . b _ * 2008 * , _ 112 _ , 79377947 kubar , t. ; kleinekathofer , u. ; elstner , m. _ j. phys . chem . b _ * 2009 * , _ 113 _ , 1310713117 grozema , f. c. ; siebbeles , l. d. a. ; berlin , y. a. ; ratner , m. a. _ chemphyschem _ * 2002 * , _ 3 _ , 536539 troisi , a. ; orlandi , g. _ j. phys . chem . b _ * 2002 * , _ 106 _ , 20932101 grozema , f. c. ; tonzani , s. ; berlin , y. a. ; schatz , g. c. ; siebbeles , l. d. a. ; ratner , m. a. _ j. am . chem . soc . _ * 2008 * , _ 130 _ , 51575166 voityuk , a. a. _ j. chem . phys . _ * 2008 * , _ 128 _ , 045104 voityuk , a. a. _ j. chem . phys . _ * 2008 * , _ 128 _ , 115101 saito , i. ; nakamura , t. ; nakatani , k. ; yoshioka , y. ; yamaguchi , k. ; sugiyama , h. _ j. am . * 1998 * , _ 120 _ , 1268612687 voityuk , a. a. ; jortner , j. ; bixon , m. ; rosch , n. _ chem . _ * 2000 * , _ 324 _ , 430434 wang , l .- p . ; chen , j. ; van voorhis , t. _ j. chem . theory comput . _ * 2013 * , _ 9 _ , 452460 pavanello , m. ; van voorhis , t. ; visscher , l. ; neugebauer , j. _ j. chem . phys . _ * 2013 * , _ 138 _ , 054101 pavanello , m. ; neugebauer , j. _ j. chem . phys . _ * 2011 * , _ 135 _ , 234103 senatore , g. ; subbaswamy , k. r. _ phys . b _ * 1986 * , _ 34 _ , 57545757 cortona , p. _ phys . rev . b _ * 1991 * , _ 44 _ , 8454 wesolowski , t. a. ; warshel , a. _ j. chem . phys . _ * 1993 * , _ 97 _ , 8050 hong , g. ; rosta , e. ; warshel , a. _ j. phys . chem . b _ * 2006 * , _ 110 _ , 1957019574 pavanello , m. ; adamowicz , l. ; volobuyev , m. ; mennucci , b. _ j. phys . chem . b _ * 2010 * , _ 114 _ , 44164423 wolter , m. ; elstner , m. ; kubar , t. _ j. chem . phys . _ * 2013 * , _ 139 _ , 125102 schuster , g. ; landman , u. the mechanism of long - distance radical cation transport in duplex dna : ion - gated hopping of polaron - like distortions . in _ longe - range charge transfer in dna i _ ; schuster , g. , ed . ; springer berlin heidelberg , 2004 ; vol . 236 , pp 139161 voityuk , a. a. ; rsch , n. ; jortner , j. ; bixon , m. _ j. phys . * 2002 * , _ 1 _ , 75997606 marcus , r. a. ; sutin , n. _ biochim . acta _ * 1985 * , _ 811 _ , 265322 moser , c. c. ; keske , j. m. ; warncke , k. ; farid , r. s. ; dutton , p. l. _ nature _ * 1992 * , _ 355 _ , 796802 marcus , r. a. _ j. chem . phys . _ * 1956 * , _ 24 _ , 966978 jortner , j. _ j. chem . phys . _ * 1976 * , _ 64 _ , 48604867 nitzan , a. _ chemical dynamics in condensed phases _ ; oxford university press : oxford , 2006 bixon , m. ; jortner , j. _ j. am . soc . _ * 2001 * , _ 123 _ , 1255612567 renger , t. ; marcus , r. a. _ j. phys . * 2003 * , _ 107 _ , 84048419 bixon , m. ; giese , b. ; wessely , s. ; langenbacher , t. ; michel - beyerle , m. e. ; jortner , j. _ proc . usa _ * 1999 * , _ 96 _ , 1171311716 berlin , y. a. ; ratner , m. a. _ radiat . * 2005 * , _ 74 _ , 124131 , w. ; sham , l. j. _ phys . rev . _ * 1965 * , _ 140 _ , 11331138 wesolowski , t. a. one - electron equations for embedded electron density : challenge for theory and practical payoffs in multi - level modeling of complex polyatomic systems . in _ computational chemistry : reviews of current trends _ ; leszczynski , j. , ed . ; world scientific : singapore , 2006 ; vol . 10 , pp 182 , g. ; bickelhaupt , f. m. ; baerends , e. j. ; van gisbergen , s. j. a. ; fonseca guerra , c. ; snijders , j. g. ; ziegler , t. _ j. comput . chem . _ * 2001 * , _ 22 _ , 931967 jacob , c. r. ; neugebauer , j. ; visscher , l. _ j. comput . chem . _ * 2008 * , _ 29 _ , 10111018 solovyeva , a. ; pavanello , m. ; neugebauer , j. _ j. chem . phys . _ * 2012 * , _ 136 _ , 194104 goodpaster , j. d. ; barnes , t. a. ; miller , iii , t. f. _ j. chem . phys . _ * 2011 * , _ 134 _ , 164108 kaduk , b. ; kowalczyk , t. ; van voorhis , t. _ chem . * 2012 * , _ 112 _ , 321370 wu , q. ; van voorhis , t. _ phys . rev . * 2005 * , _ 72 _ , 024502 wu , q. ; van voorhis , t. _ j. chem . _ * 2006 * , _ 125 _ , 164105 thom , a. j. w. ; head - gordon , m. _ j. chem . phys . _ * 2009 * , _ 131 _ , 124113 mayer , i. _ int . j. quantum chem . _ * 2002 * , _ 90 _ , 6365 adam , k. ; hoffmann , f. ; heck , a. ; oberhofer , h. ; elstner , m. ; blumberger , j. _ j. chem . phys . _ * 2014 * , _ 140 _ , 104105 perdew , j. p. ; chevary , j. a. ; vosko , s. h. ; jackson , k. a. ; pederson , m. r. ; singh , d. j. ; fiolhais , c. _ phys . b _ * 1992 * , _ 46 _ , 66716687 lembarki , a. ; chermette , h. _ phys . rev . * 1994 * , _ 50 _ , 5328 wesolowski , t. a. ; weber , j. _ int . j. quantum chem . _ * 1997 * , _ 61 _ , 303 wesolowski , t. a. ; ellinger , y. ; weber , j. _ j. chem . phys . _ * 1998 * , _ 108 _ , 6078 wesolowski , t. a. ; tran , f. _ j. chem . phys . _ * 2003 * , _ 118 _ , 2072 gtz , a. ; beyhan , s. ; visscher , l. _ j. chem . theory comput . _ * 2009 * , _ 5 _ , 31613174 case , d. et al . _ amber 12 _ , university of california , san francisco , 2012 migliore , a. _ j. chem . theory comput . _ * 2011 * , _ 7 _ , 17121725 mulliken , r. s. _ j. phys . chem . _ * 1952 * , _ 56 _ , 295311 blancafort , l. ; voityuk , a. a. _ j. phys . chem . * 2006 * , _ 110 _ , 64266432 franchini , m. ; philipsen , p. h. t. ; van lenthe , e. ; visscher , l. _ j. chem . theory comput . _ * 2014 * , submitted delley , b. _ j. chem . phys . _ * 1990 * , _ 92 _ , 508517 giese , b. _ annu . rev . biochem . _ * 2002 * , _ 71 _ , 5170 mishra , d. ; pal , s. _ j. mol . spec . ( theochem ) _ * 2009 * , _ 902 _ , 96102 voityuk , a. a. _ j. chem . phys . _ * 2007 * , _ 128 _ , 115101 giese , b. hole injection and hole transfer through dna : the hopping mechanism . in _ longe - range charge transfer in dna i _ ; schuster , g. , ed . ; springer berlin , heidelberg , 2004 ; vol . 236 , pp 2744 winkler , j. r. _ curr . _ * 2000 * , _ 4 _ , 192198 mcweeny , r. _ methods of molecular quantum mechanics _ ; academic press : san diego , 1992 skourtis , s. s. ; beratan , d. n. ; onuchic , j. n. _ chem . phys . _ * 1993 * , _ 176 _ , 501520 evenson , j. w. ; karplus , m. _ j. chem . phys . _ * 1992 * , _ 96 _ , 52725278 newton , m. d. _ chem . * 1991 * , _ 91 _ , 767792 lwdin , p .- o . spectrosc . _ * 1963 * , _ 10 _ , 1233 larsson , s. _ j. am . chem . soc . _ * 1981 * , _ 103 _ , 40344040 priyadarshy , s. ; skourtis , s. s. ; risser , s. m. ; beratan , d. n. _ j. chem . phys . _ * 1996 * , _ 104 _ , 94739481 marcus , r. _ chem . * 1987 * , _ 133 _ , 471477 hatcher , e. ; balaeff , a. ; keinan , s. ; venkatramani , r. ; beratan , d. n. _ j. am . soc . _ * 2008 * , _ 130 _ , 1175211761 voityuk , a. a. _ phys . chem . phys . _ * 2012 * , _ 14 _ , 1378913793 beratan , d. n. ; hopfield , j. j. _ j. am . soc . _ * 1984 * , _ 106 _ , 15841594 miller , k. j. _ j. am . soc . _ * 1990 * , _ 112 _ , 85338542	in the past two decades , many research groups worldwide have tried to understand and categorize simple regimes in the charge transfer of such biological systems as dna . theoretically speaking , the lack of exact theories for electron nuclear dynamics on one side , and poor quality of the parameters needed by model hamiltonians and nonadiabatic dynamics alike ( such as couplings and site energies ) on the other , are the two main difficulties for an appropriate description of the charge transfer phenomena . in this work , we present an application of a previously benchmarked and linear - scaling subsystem dft method for the calculation of couplings , site energies and superexchange decay factors ( @xmath0 ) of several biological donor acceptor dyads , as well as double stranded dna oligomers comprised of up to 5 base pairs . the calculations are all - electron , and provide a clear view of the role of the environment on superexchange couplings in dna they follow experimental trends and confirm previous semiempirical calculations . the subsystem dft method is proven to be an excellent tool for long - range , bridge - mediated coupling and site energy calculations of embedded molecular systems .
[ fig1 ] the gaia mission has been approved to provide for the first time a clear picture of the formation , structure , evolution , and future of the entire milky way . in addition , as secondary goals , gaia will contribute to many other branches of astrophysics , especially stellar and solar system minor body astrophysics , with a valuable contribution to cosmology and fundamental physics . this clear scientific prioritisation must drive the design , and all compromises . understanding the structure and evolution of the galaxy requires three complementary observational approaches : * carrying out a full census of all the objects in a large , representative , part of the galaxy ; * mapping quantitatively the spatial structure of the galaxy ; * measuring the motions of objects in three - dimensions to determine the gravitational field and the stellar orbits . in other words , what is required are complementary measurements of distances ( astrometry ) , photometry to determine both extinction and intrinsic stellar properties , and the radial velocities along our line of sight . detailed studies of the local group are the key specific test of our understanding of the formation and growth of structure in the universe . the most significant questions here relevant for gaia include the history of most of the stars in the galaxy : the nature of the inner galactic disk and the galactic bulge : what are their age , abundance and assembly histories ? stellar studies locally of course are of critical general significance : only locally can we determine the stellar initial mass function directly . this function directly controls the chemical and luminosity evolution of the universe . galactic satellite galaxies are proving the most suitable environs to quantify the nature and distribution of dark matter , and to test the small scale predictions of hierarchical galaxy formation models . the galactic disk itself , of course , is a key test of angular momentum distributions ( figure 1 ) , chemical evolution , and merger histories , and must eventually provide the most robust information on cold dark matter . such issues as the number of local dwarf galaxies and the inner cdm profiles of the dsphs are well known demanding challenges for cdm models . other local group information is also important : what is the stellar initial mass function , what is the distribution of chemical elements , what is the age range in the galactic bulge , when did the last significant disk merger happen ... ? all these challenge our appreciation of galaxy formation and evolution , and in turn provide the information needed to refine the models . the partnership between local observations and _ ab initio _ theory is close , and developing well , and must culminate in gaia . there are clear similarities and distinctions between fundamental properties of the different galactic stellar populations , such as age , metallicity , star formation history , angular momentum ( figure 1 ) which allow study of their individual histories . this is arguably one of the greatest advantages of studies of local group galaxies : one is able to disentangle the many different histories which have led to a galaxy typical of those which dominate the luminosity density of the universe . ninety percent of the matter in the universe is of an unknown nature ( to say nothing here of the even larger amount of dark energy ) . this matter dominates gravitational potential wells in the early universe on all scales , and everywhere except in the centres of large galaxies and dense stellar clusters today . the smallest scale lengths on which this dark matter is dominant are an important constraint on its nature : for example , if dark matter is concentrated on very small scale lengths it can not be relativistic . the only known way in which we can study dark matter is through its dynamical effects on test particles orbiting in a potential which it generates . the simplest dark - matter dominated systems known , with scale lengths of interest to constrain the physical nature of the dark matter , and with a dynamical structure simple enough to be amenable to understanding , are the galactic satellite dwarf spheroidal galaxies . very extensive studies , using gravitational lensing , galaxy rotation curves , stellar and gas kinematics , x - ray profiles , and so on , have initiated mapping dark matter on larger scales . for the dwarf galaxies , stellar kinematics are uniquely the method of choice . this dark matter dominates the matter in the universe , but remains weakly constrained : what is the nature of the dark matter ? a variety of theoretical studies suggest that the small scale structure of dark matter is the key test of its nature ( navarro , frenk & white 1997 ; klypin etal 1999 ; moore etal 1999 ) . many extant studies of dark matter on many different scale lengths are available , using techniques such as gravitational lensing , x - ray luminosity and temperature profiles , hi and h@xmath0 rotation curves , and stellar dynamics . none has been able to provide a reliable study of dark matter on the crucial small length scales , because the only small dark - matter dominated relatively simple dynamical systems , the milky way satellite dwarf galaxies , can be studied in detail only with very accurate kinematics of faint stars . to be confined on small scales , a system must be cold . for self - gravitating systems the virial velocity , and equivalently virial temperature , of a system in equilibrium is simply related to its mean density , @xmath1 , and characteristic radius , @xmath2 , as @xmath3 baryonic material , which can radiate away energy while gaseous , can form systems of a wide range of virial velocities , with the lower limit just set by the cooling law . in particular , baryonic dark matter could form dark haloes of small scalelength . one can relate the ` temperature ' of non - baryonic dark matter ( assumed in thermal equilibrium in the early universe ) with its velocity dispersion at the epoch in the early universe when it decouples from the ordinary matter ( bond & szalay 1983 ) . ` hot ' dark matter , for example a massive neutrino , which is relativistic at this epoch , has a large free - streaming length and can not be confined on small scales ( bond , szalay & white 1983 ) . in contrast , ` cold ' dark matter , which is non - relativistic when it decouples from ordinary matter , can form small - scale structure . one currently popular non - baryonic cold dark matter candidate , the axion , is not produced through processes in thermal equilibrium , has an extremely low ` temperature ' , and can cluster on scales below that of even a dwarf galaxy . thus the study of dark matter in systems with the smallest characteristic radius provides a thermometer of dark matter and can be used to determine its nature . dwarf spheroidal galaxies have ( stellar ) scale - lengths of @xmath4 pc ( irwin & hatzidimitriou 1995 ) , comparable to the scale height of the galactic thin disk and bulge . numerical simulations of cosmological structure formation do not have the dynamic range necessary to study such scales . irrespective of the future developments of computing power and techniques , the simulation of a dwarf galaxy requires the inclusion of the unknown physics of star formation and feedback , expected to be particularly important in these low velocity - dispersion systems ( e.g. dekel & silk 1986 ; wyse & silk 1985 ; lin & murray 1991 ) . improved understanding must be led by observations . what are the smallest systems for which there is evidence of dark matter ? galactic disks are thin , of small scale - height , and we are located in the middle of one . dark matter confined to the disk ( as distinct from the extended , galactic - scale dark halo ) would be ` cold ' . however , the local neighbourhood is not a good place to study cold disk dark matter , since it is now accepted there is no evidence that any exists ( kuijken & gilmore 1989 , 1991 ; gilmore , wyse & kuijken 1989 ; flynn & fuchs 1994 ; creze etal . 1998 ; holmberg & flynn 2000 ) . globular clusters are fairly nearby , and have very small scalelengths , tens of parsecs , but again there is no evidence for dark matter associated with globular clusters ( e.g. meylan & heggie 1997 ) . note that in fact in these two cases there is not just lack of evidence for cold dark matter , there is actual evidence for a lack of dark matter , which in itself is significant . in contrast , the dark matter content of low - luminosity dwarf galaxies , as inferred from analyses of their internal stellar and gas kinematics , makes them the most dark - matter - dominated galaxies ( e.g. mateo 1998 ; carignan & beaulieu 1989 ) . of these small galaxies , the low - luminosity gas - free dwarf spheroidals ( dsph ) are the most extreme . available stellar kinematic studies provide strong evidence for the presence of dominant dark matter ( e.g. review of mateo 1998 ) , confirming earlier inferences from estimates of the dsph s tidal radii and a model of the milky way gravitational potential ( faber & lin 1983 ) . substantial analytic methodologies have been developed over the last few years by several groups to interpret colour - magnitude data and stellar kinematics . these studies provide an existence theorem proving that efficient analysis of the gaia and complementary data is possible . adequate kinematic data can determine the potential well locally on a variety of different scales , parsecs up to 1mpc . the motivation is that the _ smallest _ scale is set by a fundamental characteristic of the dark matter , its velocity dispersion , analogously its temperature , so that study of the dynamics on these scales provides unique and strong constraints on the nature of dark matter . such a determination is possible and is possible only using full 6-dimensional phase space data , such as will be determinable by gaia . the dynamical structure of a dwarf galaxy , while simple , is not trivial . the galaxies are probably triaxial , can sustain complex families of anisotropic orbits , and live in a time - varying galactic tidal field . given this complexity , considerable information is required to derive a robust measurement of the 3-dimensional gravitational potential , and hence the generating ( dark ) mass . several groups have devised methods to explore the possible dynamical structure of a galactic satellite , and used these to simulate observational approaches to determination of the mass distribution . these studies show that determination of at least five phase space coordinates two projected spatial coordinates , and all three components of the velocity are necessary and sufficient to provide a robust and reliable determination of a dark matter mass distribution . acquisition of these data are possible only with a combination of radial velocity data and proper motion data for a well - selected sample of stars distributed across the whole face of the target galaxy . in the nearest appropriate target galaxies , draco and ursa minor , the stars of interest are at magnitude v@xmath5 , and the required proper motion accuracy , corresponding to 1 - 2 km / s in velocity , is 36@xmath6yr@xmath7 . thus this specific test of the nature of matter on small scales is better with sim rather than gaia . nonetheless , the analysis principles apply in general , and can be generalised to gaia studies of the local galaxy , where sim can not contribute . an important simplification in any dynamical analysis is that the potential be time - independent over an internal dynamical timescale , the tracer stellar distribution be in equilibrium with the potential , and be well - mixed ( relaxed ) . while dsph galaxies show every possible star formation history , it is a defining characteristic that star formation ceased at least a few gyr ago ( e.g. hernandez , gilmore & valls - gabaud 2000 ) . the internal dynamical times of these galaxies are typically only @xmath8 yr ; thus star formation indeed ceased many hundreds of crossing times ago and the systems should be well - mixed . this assumption can of course be directly tested by the data . most studies of galactic dynamics have been impeded by the degeneracy between anisotropy and mass . this degeneracy , that one can not distinguish between gravitational potential gradients and gradients in orbital anisotropy from measurements of just one component of the velocity ellipsoid ( radial velocity ) , is the fundamental limitation in extant analyses , and the key justification for extending observation to 5 phase space coordinates . the large observed central radial velocity dispersion is compatible with either a massive halo , and a low central density ; or no halo , and a large central density . to demonstrate our ability to discern between these possibilities , wilkinson etal ( 2002 ) constructed a set of models that span the space of halo mass and anisotropy , and perform monte carlo recoveries of the input anisotropy and halo mass . they assume that the luminosity density of a dsph is given by a plummer model @xmath9 where @xmath10 is determined by the total observed luminosity . next , assume that the potential of the system has the form @xmath11 for this dark matter potential , @xmath12 corresponds to a mass follows light plummer potential , keplerian at large radii ; @xmath13 yields , for large @xmath14 , a flat rotation curve ; and @xmath15 gives a harmonic oscilator potential . as the parameter @xmath0 decreases , the dsph becomes more and more dark matter dominated . finally , assume that the distribution function ( df ) of the stars in the model potential is a function entirely of the energy @xmath16 and norm of the angular momentum @xmath17 . under this assumption , one can derive a two parameter set of dfs , parameterised by @xmath0 and an anisotropy parameter @xmath18 , namely @xmath19 this formula holds for @xmath20 and @xmath21 ; similar expressions can be derived for all the other cases . these dfs generalise earlier calculations by dejonghe ( 1987 ) , which are restricted to the case where there is no dark matter . in terms of binney s anisotropy parameter @xmath22 , the radial and tangential velocity dispersions @xmath23 and @xmath24 vary as @xmath25 the potential normalisation @xmath26 in eqn 2 is a well defined function of the observed central radial velocity dispersion , and the assumed @xmath0 and @xmath18 . figure 2 shows the variation of the central and large scale mass to light ratio as a function of @xmath0 and @xmath18 . to simulate our ability to resolve among values of @xmath18 and @xmath0 , we generate 6d phase space coordinates @xmath27 drawn from the df @xmath28 , discard the @xmath29 ( geocentric radial position ) coordinate , and attempt to recover @xmath18 and @xmath0 using a bayesian likelihood technique ( see e.g. little & tremaine 1987 , kochanek 1996 , wilkinson & evans 1999 , van der marel etal 2000 ) . specifically , we scan a grid of @xmath30 , and at each point compute the probability of observing the input data set @xmath31 by bayes s theorem , and the assumption of uniform prior probabilities of @xmath30 , the most likely @xmath30 are given by maximising eqn 5 . confidence regions are obtained by applying 2d @xmath32 statistics to the logarithm of eqn 5 . figure 3 ( left ) shows the result of a monte carlo simulation to recover the input @xmath0 and @xmath22 of two model distributions ( @xmath33 , and @xmath34 , respectively ) . these values of @xmath0 and @xmath18 lie along the direction of the usual mass / anisotropy indeterminacy , and thus illustrate our ability to break this degeneracy . both model fits involve 800 reconstructions , each of which contains an ensemble of 250 simulated stars . the right panel of figure 3 depicts constraints placed on the central @xmath35 , and the total @xmath35 within four plummer radii . although the two models evaluated have a similar total mass within @xmath36 , our method clearly recovers the significant difference in central @xmath35 caused by a halo . the top two panels of figure 4 illustrate the effect of anticipated velocity measurement errors which are 10% and 20% of the internal ( local ) velocity dispersion of the ( sub-)system under study , assuming a 250 star sample . these errors cause a systematic displacement of the best fit @xmath0 in the small halo ( @xmath37 ) case we examine , because they create a significant increase of the apparent velocity dispersion at the maximum radius ( @xmath38 ) considered . however , the width of the distribution remains narrow , and the systematic @xmath0 shift will be rigorously incorporated into the analysis by convolving the velocity errors with the df . larger errors would not be acceptable , however , because resolving among @xmath0 values requires a precise determination of the falloff of the velocity dispersion . the third panel of figure 4 illustrates the recovery of @xmath30 using only 25 velocities ; the halo and anisotropy are essentially unconstrained by this small sample . the bottom panel shows the effect of using 750 purely radial velocities , with zero errors . although we have the same number of one dimensional velocities as in the 250 star cases , we are unable to recover @xmath0 and @xmath18 ; this failure indicates that the analysis requires all three velocity components , and this work could not be performed using ground based radial velocities . a triaxial spheroid seen projected on the plane of sky mimics a rotationally symmetric ellipse . triaxial systems can support much more complex orbital structures than can 2-d systems ; this allows a degeneracy between radially - dependent orbital complexity and system shape , potentially invalidating dynamical analyses . more complex modelling is required in this case , and has been partially developed . the ( very obvious ! ) conclusion from extensive numerical simulations based on those triaxial models is that it is vital to have full kinematics available if we are to investigate triaxial ( more generally : not isolated spherical ) systems . the discussion above illustrates the considerable progress in quantitative analyses of stellar kinematics which are currently under development . the introduction recalled that gaia is funded to study galactic evolution . what data must gaia provide so these analysis techniques can deliver this science . the quantitative analysis techniques need distances , 2-d ( better 3-d ) kinematics , and some astrophysical information on the distribution function of stellar properties . highly detailed information on a few stars is not what is needed . can gaia deliver ? simulations presented at this meeting show gaia radial velocity data will be unable to provide useful scientific data at low galactic latitudes . low galactic latitudes , while `` only '' a few percent of sky , contain most of the stars , and contain all of the inner disk and bulge . it is worth emphasising that these parts of the milky way are the only parts of high redshift galaxies which we can see . if gaia fails to provide adequate data on these low latitudes , it fails a primary science goal . however , as the simulations show , bad data are useless . the lesson for gaia spectra is clear : do something excellently , not several things badly , optimise for the faintest possible stars in uncrowded regions : leave low latitudes for the other instruments . what other instruments ? the medium band photometr ( mbp ) unfortunately has been degraded to poor ground - based quality spatial resolution , 1arcsec . thus , as long experience has establsihed , mbp will also be of little or no value in the inner galaxy . again , the implication is clear : optimise mbp for relatively low extinction higher - latitude regions , with somewhat metal - poor old stars . can gaia meets its primary science goals at all : only if the broad band filters are optimised to deliver the essential minimum astrophysical data . that is , the limited sensitivity of rvs and the poor resolution of mbp require that the broad band filters be optimised for astrophysical analyses , and not other considerations . without these choices , gaai will fail to meet its design science goals . with them , it will revolutionise our understanding of galaxy formation and evolution .	gaia has been approved to provide the data needed to quantify the formation and evolution of the milky way galaxy , and its near neighbours . that requires study of all four key galactic stellar populations : bulge , halo , thick disk , thin disk . the complex analysis methodologies required to model gaia kinematic data are being developed , and in the interim applied to the relatively simple cases of the satellite dsph galaxies . these methodologies , illustrated here , show that we will be able to interpret the gaia data . they also quantify what data gaia must provide . it is very unlikely in the present design that gaia will be able to provide either radial velocities or worthwhile photometry for study of two of the key science goals : the galactic bulge and the ( inner ) galactic old disk . the implication is that the radial velocity spectrometer and the medium band photometer should be optimised for study of low density fields suitable for their low spatial resolution and the broad band photometry must be optimised for inner galaxy astrophysical studies . # 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in
gallager invented low - density parity - check ( ldpc ) codes @xcite in 1963 . ldpc codes are linear codes defined by sparse bipartite graphs . luby et al . introduced the _ peeling algorithm _ ( pa ) @xcite for the binary erasure channel ( bec ) . pa is an iterative algorithm which is defined on tanner graphs . pa and brief propagation ( bp ) decoder have the same decoding result . as pa proceeds , edges and nodes are progressively removed . the residual graphs consist of nodes and edges that are still unknown at each iteration . the decoding successfully halts if the graph vanishes . amraoui @xcite showed that distributions of the number of check nodes of degree one in the residual graph convergences weakly to a gaussian as blocklength tends to infinity . amraoui also showed that block and bit error probability of finite - length ldpc codes are derived by the average and the variance of the number of check nodes of degree one in the residual graph . the average number of check nodes of degree one in the residual graph is determined from a system of differential equations , which was derived and solved by luby et al . the variance of the number of check nodes of degree one in the residual graph is also determined from a system of differential equations called _ covariance evolution _ , which was derived by amraoui et al . since analytical solution of covariance evolution has not been known so far , we had to resort to numerical computations to solve the covariance evolution . an alternative way to determine the variance of the number of check nodes of degree one was proposed in @xcite . the variance of the number of check nodes of degree one in the residual graph can be computed by determining the variance of the number of erased messages of bp for bec with parameter @xmath0 where @xmath0 is the threshold of the ensemble under bp decoding . this method is a valid approximation for the erasure probability close to @xmath1 . moreover , ezri et al . extended to this method to more general channels @xcite . however , if we solve the covariance evolution analytically , we can derive the variance of the number of check nodes of degree one in the residual graph for all @xmath2 where @xmath2 is the channel parameter for the bec . in this paper , we show an analytical solution of the covariance evolution for regular ldpc code ensembles . in this section , we briefly review the covariance evolution and _ initial covariance _ in @xcite . we consider the transmission over the bec with channel erasure probability @xmath3 using ldpc codes in a ( @xmath4 , @xmath5)-regular ldpc code ensemble . let @xmath6 denote the iteration round and @xmath7 be the total number of edges in the original graph . we define that @xmath8 define a parameter @xmath9 such that @xmath10 and @xmath11 when @xmath12 . let @xmath13 denote a random variable corresponding to the number of edges connecting to variable nodes of degree @xmath4 in the residual graph at the iteration round @xmath6 . let @xmath14 denote a random variable corresponding to the number of edges connecting to check nodes of degree @xmath15 in the residual graph at the iteration round @xmath6 . those random variables depends on the choice of the graph from ( @xmath16)-regular ldpc code ensemble , the channel outputs and the random choices made by pa . we define @xmath17 to simplify the notation , we drop the subscript @xmath6 . for @xmath18 , we define @xmath19 by @xmath20}{\xi}. \notag\end{aligned}\ ] ] we also define @xmath21 by the covariance of @xmath22 and @xmath23 ( @xmath24 ) divided by the total number of edges in the original graph i.e. @xmath25}{\xi}. \notag\end{aligned}\ ] ] in @xcite , amraoui showed these parameters satisfy the following system of differential equations in the limit of the block length . this system is referred to as covariance evolution . @xmath26 , \label{ce}\end{aligned}\ ] ] where @xmath27 @xmath28 @xmath29 @xmath30 where @xmath31 , @xmath32 , @xmath33 , @xmath34 , @xmath35 , @xmath36 and @xmath37 is the indicator function which equals to 1 if @xmath38 and 0 otherwise . the initial conditions of the covariance evolution are given by initial covariances . the initial covariances are the covariances of the number of edges of each degree at the start of the decoding divided by the total number of edges in the original graph . for @xmath39 , initial covariances are derived in @xcite , as follows . we show in the following theorem the analytical solution of the covariance evolution , for a ( @xmath4,@xmath5)-regular ldpc code ensemble . the proof ) , ( [ res_l_r ] ) and ( [ res_r_r ] ) with respect to @xmath9 , we can check that those equations fulfill ( [ ce ] ) . ] is given in section [ sec : derive ] . [ theorem_cov ] let @xmath41 be the normalized iteration round of pa as defined in ( [ eq : tau ] ) . for a ( @xmath4,@xmath5)-regular ldpc code ensemble and @xmath42 , in the limit of the code length , we obtain the following . @xmath43 where @xmath44 and @xmath9 is defined by @xmath10 with @xmath45 when @xmath12 . in @xcite , _ scaling parameter _ @xmath46 is given by @xmath47 where @xmath0 is the threshold of the ensemble under bp decoding , @xmath48 is the non - zero solution of @xmath49 at the threshold , @xmath50 is the blocklength and @xmath7 is the total number of edges in the original graph . we define @xmath51 and @xmath52 . since @xmath53 and @xmath54 , we see that @xmath55 and @xmath56 . using those equations , we have from ( [ res_r_r ] ) @xmath57 ( note that @xmath58 ) . recall that @xmath59 . we see that @xmath60 from ( [ alpha ] ) , we can obtain @xmath61 this is the same result as in @xcite for regular ldpc code ensembles . figure [ fig_cov36 ] shows the solution of the covariance evolution @xmath62 , @xmath63 , as a function of @xmath9 for ( 3,6)-regular ldpc code ensemble . figure [ fig_cov24 ] shows the solution of the covariance evolution @xmath62 , @xmath64 , as a function of @xmath9 for ( 2,4)-regular ldpc code ensemble . , @xmath65 , as a function of the parameter @xmath9 for the ( 3,6)-regular ldpc code ensemble . the channel parameter is @xmath66.,width=302 ] , @xmath67 , as a function of the parameter @xmath9 for the ( 2,4)-regular ldpc code ensemble . the channel parameter is @xmath68.,width=302 ] from ( [ ce ] ) , we get @xmath70 . from initial covariance , we have @xmath71 . this leads to ( [ res_l_l ] ) . in order to solve @xmath72 , we define @xmath73 , which gives @xmath74 . from ( [ ce ] ) , we see that @xmath75 where @xmath76 this equation is a first order linear differential equation and the solution is given by @xmath77 with a constant @xmath78 determined from the initial covariance where @xmath79 . note that @xmath80 . we see that @xmath81 . we have @xmath82 from ( [ ce ] ) and ( [ res_a ] ) , we get for @xmath83 @xmath84 where @xmath85 those equations are first order linear differential equations . the solutions are given by @xmath86 with constants @xmath87 determined from the initial covariances . those equations can be solved by mathmatical induction for @xmath88 . we show that @xmath89 fulfill ( [ res_l_r ] ) . from ( [ res_a ] ) , we can write @xmath90 using the same way in the induction step , we can obtain @xmath91 we show that if @xmath92 fulfill ( [ res_l_r ] ) , then also @xmath72 fulfill ( [ res_l_r ] ) . assume @xmath93 . using the induction hypothesis , we can write @xmath94 using @xmath95 , we see that @xmath96 similarly , we have @xmath97 where @xmath98 note that @xmath99 from ( [ eq_a_3 ] ) and ( [ eq_a_4 ] ) , we have @xmath100 where @xmath101 from ( [ eq_a_1]),([eq_a_2 ] ) and ( [ eq_a_5 ] ) , we have @xmath102 from initial covariance , we have @xmath103 hence we obtain @xmath104 this lads to ( [ res_l_r ] ) for @xmath88 . note that @xmath105 and that @xmath106 . we have @xmath107 hence we obtain ( [ res_l_r ] ) . in order to solve @xmath109 , we define @xmath110 and @xmath111 . from ( [ ce ] ) , we get for @xmath112 @xmath113 where @xmath114 the solution of ( [ diff_b ] ) is given by @xmath115 with a constant @xmath116 which determined from the inital covariance . from the initial covariance , we get @xmath117 we see that @xmath118 . hence we have @xmath119 from ( [ diff_br ] ) , we get @xmath120 with constants @xmath121 . for @xmath88 , those equation are solved by mathmatical induction as the proof for @xmath72 . from the initial covariances , note that @xmath122 we have @xmath123 using @xmath124 , we have @xmath125 from ( [ ce ] ) , we get for @xmath127 @xmath128 where @xmath129 the solutions of those differential equations are given by @xmath130 with constants @xmath131 . using ( [ res_br ] ) , we can solve those equations by mathmatical induction for @xmath132 . we have @xmath133 note that @xmath134 . we have for @xmath135 @xmath136 since @xmath137 , we have @xmath138 thus , we can obtain ( [ res_r_r ] ) . in this section , we consider the relationship between the _ stability condition _ @xcite and @xmath139 . for a @xmath140-regular ldpc code ensemble ( @xmath141 ) , we see from ( [ res_l_r ] ) and ( [ res_r_r ] ) that @xmath142 for a @xmath143-regular ldpc code ensemble , we see from ( [ res_l_r ] ) and ( [ res_r_r ] ) that @xmath144 if we define the correlation coefficient for @xmath145 by @xmath146 we obtain @xmath147 note that @xmath148 agree with the stability condition for regular ldpc code ensembles . figure [ fig_cov24_0 ] shows the solution of covariance evolution @xmath149 , @xmath150 , as a function of the channel parameter @xmath2 for the ( 2,4)-regular ldpc code ensemble . from figure [ fig_cov24_0 ] , we see that @xmath151 and @xmath152 when @xmath153 . also we see that @xmath154 and @xmath155 when @xmath156 . , @xmath150 , as a function of the channel parameter @xmath2 for the ( 2,4)-regular ldpc code ensemble.,width=302 ] in this paper , we have solved analytically the covariance evolution for regular ldpc code ensembles . moreover we have derived the relationship between stability condition . t. j. richardson , m. a. shokrollahi , and r. l. urbanke , `` design of capacity - approaching irregular low - density parity - check codes , '' _ ieee transactions on information theory _ , vol . 619 - 637 , 2001 .	the covariance evolution is a system of differential equations with respect to the covariance of the number of edges connecting to the nodes of each residual degree . solving the covariance evolution , we can derive distributions of the number of check nodes of residual degree 1 , which helps us to estimate the block error probability for finite - length ldpc code . amraoui et al . resorted to numerical computations to solve the covariance evolution . in this paper , we give the analytical solution of the covariance evolution .
nearly all of our knowledge about the masses and ages of low mass young stars comes from their location in the hr diagram relative to theoretical calculations of stellar evolution to the main sequence . despite considerable advances over the past 5 - 10 years in understanding the structure and atmospheres of stars of mass m@xmath1 , comparison of the currently available predicted evolutionary paths of young stars shows obvious differences . empirical tests of the calculations have not been possible until recently because astronomers have not had independent measurements of either the mass or age of a young star . this situation is changing rapidly . one way to test the calculated tracks is to investigate whether they yield the same ages for stars expected to be coeval on physical grounds . hartigan et al . ( 1994 ) , casey et al . ( 1998 ) , and white et al . ( 1999 ) have applied this test to wide binaries , the ty cra system , and the gg tau system , respectively , using tracks available at the time . over a mass range extending down to the brown dwarfs , they found significant differences in the extent to which the several theoretical calculations satisfy the coevality requirement . the capability of mm - wave interferometers to resolve the spectral line emission of the outer disks of pms stars offers the possibility to map their rotation . since the disk mass is usually very small compared to the mass of the star , this provides the means to measure the stellar mass dynamically ( e.g. dutrey et al . 1994 , dgs94 ; guilloteau and dutrey 1998 , gd98 ) . we use this technique to obtain new measurements of the mass of 5 single pms stars and one pms binary . combined with our previous measurements of the pms single gm aur and the binary gg tau ( dutrey et al . 1998 , guilloteau , dutrey and simon 1999 ) the measured stellar masses span the range @xmath2 to 0.5 . we use the results to test the theoretical calculations of pms evolution . 2 describes the sample of stars and the interferometric observations . 3 summarizes the analysis and presents the measured masses and related parameters . 4 compares the measured masses with those implied by location of the stars in the hr diagram relative to the tracks calculated by dantona and mazzitelli ( 1997 , dm97 ) , baraffe et al . ( 1998 , bcah ) , palla and stahler ( 1999 , ps99 ) , and siess et al . ( 2000 , sdf ) . all the stars studied are in taurus - auriga ( table 1 ) . we obtained new interferometric observations in the @xmath0co j= 2 - 1 line of the singles cy tau , dl tau , dm tau , lkca15 , and mwc480 ( hd 31648 ) , and the spectroscopic binary uz tau e ( mathieu et al . 1996 ) . they were selected because earlier interferometric observations ( table 1 , col . 5 ) indicated that they were , or were likely to be , associated with resolvable circumstellar disks . bp tau was observed because its _ hipparcos _ distance , 56@xmath3 pc ( favata et al . 1998 ) , is very different from the 140 pc average distance to the taurus star forming region ( kenyon et al . 1994 ) , suggesting that it is much older than previously thought ( see also bertout et al . mass measurements of the single gm aur and the binary gg tau aa by @xmath0co j= 2 - 1 line interferometric mapping have already been reported ( dutrey et al . 1998 and guilloteau et al . 1999 ) . we included dm tau in our program because new data at 3 times better angular resolution would provide an independent check on the method and earlier results ( gd98 ) . table 1 provides spectral types , luminosities , and effective temperatures for our sample . most of the spectral types and stellar luminosities , l@xmath4 , in cols . 2 and 3 of table 1 are taken from kenyon and hartmann ( 1995 ) . the luminosities are the l@xmath5 in their table a4 . all the luminosities in col . 3 are evaluated at 140 pc distance . the actual distances to the individual stars are important for our analysis but are presently unknown ; we will discuss this problem in sect . 4 . the effective temperatures in col . 4 , t@xmath6 , corresponding to the spectral types , for all the stars , are obtained using table a1 of kenyon and hartmann ( 1995 ) for main sequence stars ; we discuss the spectral - type to t@xmath6 conversion further in sect . our iram interferometer observations used 5 antennas operating in snapshot mode ( guilloteau et al . 1992 , dutrey et al . 1996 ) and were carried out in winter 1997/1998 . for bp tau , the full synthesis mapping was obtained in winter 1998/1999 . we observed all the sources simultaneously at 89.2 ( hco@xmath7 j= 1 - 0 ) ghz and 230.5 ghz ( @xmath0co j= 2 - 1 ) . the spectra were analyzed using a correlator with one band of 10 mhz centered on the hco@xmath7 j= 1 - 0 line , one band of 20 mhz centered on the @xmath0co j= 2 - 1 line , 2 bands of 160 mhz for the 1.3 mm and 3.4 mm continuum , respectively . the spectral resolution was 0.23 and 0.18 km / s in the narrow bands at 3.4 and 1.3 mm , respectively . the phase calibrators were 0415 + 379 ( 3c 111 ) and pks 0528 + 134 . the rms phase noise was 8@xmath8 to 25@xmath8 and 15@xmath8 to 50@xmath8 at 3.4 mm and 1.3 mm , respectively , which introduced position errors of @xmath9 . the seeing , estimated from observations of the calibrators , was @xmath10 . baselines up to 400 m provided @xmath11 resolution for the 1.3 mm continuum data and 1.8@xmath12 at 3.4 mm . the flux scale was checked by observation of mwc349 which has been used as calibrator at iram since 1996 . its flux is given by @xmath13 jy , following altenhoff et al . ( 1981 ) , and subsequent measurements of the planets made at iram ( iram flux report 13 ) . the gildas software package was used to reduce the data . at 1.3 mm , the continuum images are the results of summing the lower and the upper side - band data . at 3.4 mm , the tuning was purely in the lower side band . the continuum and line maps were produced using natural weighting of the visibilities . for clarity , tapering to about @xmath14 resolution was used on line maps presented in figs.1 and 2 , but the full angular resolution was used in the fitting of models by the @xmath15 technique . we did not subtract the continuum emission from the co map because its contribution is small , but included it in the analysis . the essential inputs are the angularly resolved maps of the co line emission at each velocity channel and the `` standard '' model of a rotating disk in hydrostatic equilibrium ( e.g. dgs94 ) . we use the procedure developed by gd98 to derive the parameters of the model by the method of @xmath15 minimization . in order to avoid nonlinear effects arising from deconvolution , the @xmath15 minimization is performed in the _ u , v _ plane . the derived parameters divide naturally into those relevant for the determination of the central mass and those that pertain mostly to the properties of the disk . here , with the exception of dm tau , for which we summarize the complete analysis , we will concentrate on the parameters required to determine the central mass - the disk inclination @xmath16 and position angle @xmath17 , @xmath18 sin@xmath16 at @xmath19 = 100 au , the exponent @xmath20 of the rotation curve @xmath21 and the outer disk radius @xmath22 . a paper describing the disk parameters ( e.g. the temperature radial distribution , the @xmath0co abundance and spectral line turbulence width ) is in preparation . gd98 demonstrated the @xmath15 parameter fitting procedure by applying it to @xmath0co j= 1 - 0 data for the single star dm tau . they used d=150 pc for the distance to dm tau . since we now adopt 140 pc as the reference distance , we scale gd98 s results to this value . on this basis , gd98 s value for dm tau s mass is @xmath23pc ) @xmath24 ( table 3 ) . 1 shows 7 of the velocity channel maps measured in the @xmath0co j=2 - 1 line for all the sources in our sample . 1 also presents velocity gradient maps which show on the red and blue shifted sides of the systemic velocity , the characteristic co emission pattern attributable to rotation of a tilted disk . 1a shows that dm tau , gm aur , lkca15 and mwc480 appear to have no co emission from the parent cloud . uz tau e and dl tau ( fig . 1b ) suffer some confusion with the co emission of the cloud . confusion is maximum at the cloud systemic velocity , which is usually ( but not always , see dl tau ) the stellar velocity . hence it preferentially affects the disk emission which is elongated along the projected disk minor axis . this results in an incorrect estimate of the disk inclination . in the case of cy tau , we can not be sure about the level of confusion because of the weakness of its co emission . for uz tau e and dl tau , the velocity channels hidden by the foreground co cloud have been removed from the @xmath15 analysis . in the case of uz tau e , 4 out of 36 significant channels were removed . in the case of dl tau , 5 out of 18 significant channels had to be removed because of confusion . this increased the uncertainty in the disk inclination quite significantly . the first step of the model fitting procedure is to measure the coordinates of the mm - wave continuum source in order to recenter the @xmath25 plane visibilities on it . table 2 lists these source coordinates for dm tau and for the other stars newly observed for this program . the uz tau e spectroscopic binary can not be resolved by our observations ; its continuum source coordinates must be effectively centered on the binary . table 3 compares the results of the full analysis of dm tau obtained by gd98 from their analysis of the @xmath0co 1 - 0 data at @xmath26 resolution with that derived from our new @xmath0co j=2 - 1 data . only the fitted parameters are shown ; the gas density and @xmath0co abundance are the same in the two analyses and do not affect the stellar mass determination because the @xmath0co j= 1 - 0 and j= 2 - 1 lines are optically thick . we scaled gd98 s values of the parameters that depend on physical size , and hence distance , from 150 pc distance to the 140 pc distance used here . the uncertainties are the formal 2@xmath27 values from the @xmath15 fit . the position angles in table 3 refer to the major axis of the disk ; the pa listed in gd98 is that for the disk rotation axis so differs by 90@xmath8 from the pa@xmath28 given here . the agreement between gd98 and the new results is excellent for all the parameters except for the inclination which now has the opposite sign . it is not surprising that the higher resolution data better define the sense of the inclination . its sign does not affect the mass determination . table 4 lists the parameters most relevant for determination of the central mass derived from our new observations and also from our previously published observations . columns 2 and 3 are described below . columns 4 and 5 list the disk pa and @xmath16 derived from the co line maps . the disk outer radius , r@xmath29 , in col . 5 is included here because its determination and that of @xmath16 are coupled ( see , for example , gd98 sect.4.4 ) . 6 lists the value of vsin@xmath16 at the reference radius 100 au and col . 7 the exponent of the radial power law dependence of the velocity . within the uncertainties , the rotation is keplerian . the stellar masses in col . 8 are derived using @xmath30 where 2.98 km / s is the circular velocity at 100 au radius from a 1 star . the derived masses scale as @xmath31 and hence depend on distance to the star as @xmath32 . the main source of uncertainty in the measured stellar mass is the inclination ; the two objects in our sample with the lowest inclinations , cy tau and bp tau , have the highest uncertainties in their derived masses . this approach ignores the mass of the disk because it is usually only a few per cent that of the star . even in gg tau a , which probably has the most massive disk of the stars in our sample , the mass of the circumbinary is only @xmath33 that of the stars within it ( guilloteau et al . it is interesting that uz tau e has a circumbinary ring like gg tau a and uy aur ( duvert et al . 1998 ) , although the central hole is as yet unresolved . the 1.3 mm continuum emission is detected by a separate back - end ( 2.2 ) and the detected continuum emission arises mostly from the innermost regions of the disk ( see below ) . the continuum images can therefore provide independent measurements of the disk orientation . cols.2 - 3 of table 4 give the position angle , pa@xmath34 of the major axis of the disk , and its inclination , @xmath35 , measured from the 1.3 mm wavelength continuum image . to derive the inclination from the ratio of apparent major and minor axes , we used an effective seeing of 0.3@xmath12 . we did not include uncertainty in the seeing in the error estimate . the agreement between the orientations and inclinations derived from 1.3 mm continuum data and @xmath0co j=2 - 1 data is excellent for most of the targets . the uncertainties of the values derived from the continuum images are larger , in most cases , than those from the co data because the measured sizes of the disks in the continuum are generally smaller than in co , specially when the inclinations are low . the sizes measured in the continuum are smaller because the dust continuum opacity is much smaller than in the co line . the detected continuum emission therefore arises mostly in the dense inner regions of the disks ( see , for example , d96 ) . the values for cytau suffer from low signal to noise . in the case dltau , significant confusion from the molecular cloud seems to have affected the determination of the inclination from the co emission ( 3 ) . the confusion affects the channels near the systemic velocity , which are the most important to determine the inclination . accordingly , pending better measurements ( e.g. @xmath36co ) , we prefer to use the inclination determined from the continuum emission , @xmath37 , to derive the mass for dltau . with this inclination , the stellar mass is @xmath38 , rather than @xmath39 when using the co derived inclination . the observational measure of the stellar photospheric temperature is its spectral type while the parameter provided by the models is usually @xmath40 . the surface gravities of the pms stars lie between those of the main sequence stars and giants so the conversion between spectral type and t@xmath6 determined for the giants and dwarfs may not apply . white et al . ( 1999 ) and luhman ( 1999 ) , in addition to others , have investigated this problem and demonstrate that the differences in temperature scales become apparent for the lowest mass stars , and below the hydrogen - burning limit . luhman ( 1999 ) proposed an `` intermediate '' temperature scale for the m stars . its differences from kenyon and hartmann s ( 1995 ) temperature scale become significant for stars cooler than about m6 . since our sample does not include stars as cool as this , we use the kenyon and hartmann scale in the present work without modification . the hr diagram in fig . 2a shows excerpts from pms tracks calculated by dm97 , bcah , ps99 and sdf . in this and subsequent figures , the bcah tracks for @xmath41 are for mixing length parameter ( mixing length / pressure scale height)= 1.0 and 1.9 for @xmath42 ( bcah98 and baraffe , priv . differences among the theoretical calculations for a star of given age and mass are obvious . for example , the dm97 tracks are hotter than those of bcah for @xmath43 . sdf s models at 0.2 and 0.1 appear to contract more slowly than those of dm97 , bcah , and ps99 . since the different treatments of stellar convection , the equation of state , opacities , and stellar atmospheres become important in specific regions of the hr diagram , tests of the theoretical tracks require accurate measurements of mass over as wide a range of mass and age as possible . the dependence of @xmath44 on distance ( 3 ) is important because while the average distance to the taurus star forming region ( sfr ) is reasonably well known , the actual distance to a given star in it is not . the taurus sfr extends for at least @xmath45 on the sky . if its depth is comparable to its width , the distance to a specific member may scatter by @xmath46 pc around the 140 pc mean , a relative error of @xmath47 . only cy tau and bp tau ( table 4 ) have a larger uncertainty of their mass measurement . for most of the stars in table 4 the uncertainty in their actual mass is therefore dominated by the uncertainty in their distance . the distance of course also determines the star s location in the hr diagram because @xmath48 . we will therefore present most of our results in 4 on modified hr diagrams in which the distance - independent parameter @xmath49 is plotted versus @xmath40 . for reference , fig . 2b compares the tracks on this basis . 3 plots the single t tauris in our sample on modified hr diagrams using pms tracks calculated by dm97 , bcah , ps99 , and sdf . the uncertainties displayed along the horizontal axis are @xmath50 spectral type sub - class . along the vertical axis , the uncertainties are the propagated internal uncertainties in stellar mass ( table 4 ) and an assumed @xmath51 uncertainty in the luminosity given in table 1 . we discuss the fit of each star to the theoretical tracks separately . [ [ lkca15 ] ] lkca15 : + + + + + + + the @xmath49 value for lkca15 lies near the 1.0 tracks calculated by bcah , ps99 , and sdf ( fig . 3 ) consistent with its mass @xmath52 at the nominal 140 pc distance ( table 4 ) . these tracks yield a consistent age estimate of 3 - 5 my . lkca15 s @xmath49 values lies on the 0.8 track calculated by dm97 . agreement with the dm97 track would require a distance of @xmath53 pc but would not affect the age estimate . [ [ dl - tau ] ] dl tau : + + + + + + + the position of dl tau s @xmath49 in the diagrams is within one spectral type uncertainty of the 0.8 track for each of the calculations . this is in good agreement with the mass derived using the continuum inclination , @xmath38 . the theoretical tracks yield a consistent age estimate of @xmath2 my . [ [ gm - aur ] ] gm aur : + + + + + + + the position of the @xmath49 value relative to the bcah , ps99 , and sdf tracks is consistent with its mass @xmath54 at 140 pc . the tracks provide an age estimate of @xmath55 my . the @xmath49 value lies close to dm97 s 0.6 track which would require a distance of 103 pc to force agreement with the dynamical mass . [ [ bp - tau ] ] bp tau : + + + + + + + bp tau s @xmath49 values lies between dm97 s 0.6 and 0.8 tracks , and close to bcah , ps99 , and sdf s 0.8 tracks , while our measured mass at 140 pc is in the range 0.92 to 1.49 . this suggests that the distance to bp tau may be closer than 140 pc but not as extreme as the _ hipparcos _ value , 53@xmath56 pc ( favata et al . the large uncertainty in the mass produces a large spread in the age estimate , 2 - 10 my . [ [ dm - tau ] ] dm tau : + + + + + + + its dynamical mass at 140 pc , @xmath57 , is consistent with its @xmath49 value relative to four sets of tracks . the tracks yield a consistent age estimate @xmath58 my . [ [ cy - tau ] ] cy tau : + + + + + + + the nominal value 0.55 is consistent with its @xmath49 relative to the four sets of tracks with an age in the range 2 - 5 my . however , the uncertainty in its mass is so large that this agreement is not significant , and we can not plot it meaningfully in fig . [ [ mwc480-hd-31648 ] ] mwc480 ( = hd 31648 ) : + + + + + + + + + + + + + + + + + + + fig . 4 provides a similar comparisons of the dynamical mass of the haebe star mwc480 . we calculated its stellar luminosity ( table 1 ) for a distance of 140 pc using the spectral energy distribution , corrected for extinction and circumstellar emission , derived by malfait et al . this value , 11.5 , is about half that calculated by mannings and sargent ( 1997 ) from photometry available at the time . table 1 also lists grady s ( 1999 ) spectral type estimate , a4 , from hst spectra , which is cooler than th et al s ( 1994 ) earlier value , a2-a3 . bcah pms models do not extend to @xmath59 , so we make the comparisons only with dm97 , ps99 , and sdf s tracks . s @xmath49 value lies close to the 2.0 tracks for the three calculations at an age @xmath60 my . the mass at 140 pc distance @xmath61 suggests that mwc480 lies at a somewhat greater distance . a distance of 170 pc would yield a dynamical mass of 2.0 and would be within @xmath62 of the _ hipparcos _ measurement of @xmath63 pc ( van den ancker et al 1998 ) . the results for the binaries can not be plotted on @xmath64 diagrams because the component luminosities of uz tau e or component masses of gg tau are not known . we discuss the binaries using conventional hr diagrams . [ [ uz - tau - e ] ] uz tau e : + + + + + + + + + mathieu et al ( 1996 ) discovered that uz tau e is a single - lined spectroscopic binary with period 19.1 days and projected semi - major axis of the primary @xmath65 au in which @xmath66 is the inclination of the orbit . our measurement of the total binary mass @xmath67 ( table 4 ) and the period determine @xmath68 au . we can solve for @xmath69 and @xmath70 , and hence @xmath71 and @xmath72 if we assume that the circumbinary disk and binary orbit are coplanar , @xmath73 . this assumption yields @xmath74 and @xmath75 . we plot the primary in the hr diagram ( fig . 5 ) assuming that the system spectral type m1 ( table 1 ) applies to the primary , assigning , as before , an uncertainty of @xmath50 one sub - class , and luminosity in the range 0.5 to 1 times the total system luminosity , 1.6 . 5 shows that , with these assumptions , the primary lies in the @xmath76 region of the hr diagram for all the tracks , clearly discrepant with the derived mass , @xmath77 . it seems very unlikely that the distance to uz tau e could be @xmath78 of the nominal 140 pc . rather , the assumption that @xmath79 is probably wrong ; the circumbinary disk and binary star orbits appear not to be coplanar . the spectroscopic binary orbit will be resolvable by the next generation space - borne astrometric missions so a precision measurement of the orbital parameters should be possible within a decade . [ [ gg - tau ] ] gg tau : + + + + + + + white et al . ( 1999 ) compared the ( l , t@xmath6 ) positions of gg tau aa and ab , the components of the @xmath80 separation binary at the center of the circumbinary ring we observed , with pms evolutionary tracks calculated by swenson et al . ( 1994 ) , dantona and mazzitelli ( 1994 ) , dm 97 , and bcah . they used the effective temperature scales for main sequence dwarfs drawn from the works of bessel ( 1991 ) , leggett ( 1996 ) , and luhman and rieke ( 1998 ) . they found that the tracks calculated by bcah were the most consistent with the total dynamical system we have measured and the expected coevality of the components . the bcah indicated an age between 1 and 2 my . 6 shows a similar comparison with the tracks of dm97 , bcah , ps99 , and sdf . the difference between the temperature scale used by white et al and kenyon and hartmann s ( 1995 ) temperature scale is unimportant at the effective temperatures of gg tau aa and ab . the bcah tracks suggest primary and secondary masses of @xmath81 and @xmath82 , and age 1 - 2 my , in agreement with white et al s finding . the ps99 and sdf tracks suggest similar masses and age . the dm97 tracks also indicate a young age for the components , in this case less than 1 my . the total mass indicated by dm97 s tracks , @xmath83 , is significantly lower than our measured value . to reconcile it with the measured mass would require that the gg tau system lie at distance @xmath84 pc which seems unlikely . our results indicate that : \1 ) stellar masses can be measured dynamically by the rotation of their circumstellar disks with high enough precision , @xmath85 , that meaningful tests of calculations of pre - main sequence evolution are possible . \2 ) the bcah , ps99 , and sdf models in general are in reasonable agreement with the measured dynamical masses at the average distance to the taurus sfr . to force agreement between the dm97 models and mass measurements for stars in the @xmath82 to 1 range ( lkca15 , gmaur , and the components of ggtau ) would require that these stars lie at distances 100 - 115 pc which seem unacceptably near . this is a consequence of the apparently warmer photospheres of dm97 s models for stars close to 1 . our tests of the pre - ms tracks are limited by our present ignorance of actual distances to individual stars . we anticipate that this limitation will be overcome by precision distances that will be measured by the astrometric observatories currently under construction ( e.g. fame , gaia , and sim ) . all our measurements at present pertain to stars with @xmath87 so we have been unable to test the models at the lowest end of the stellar mass spectrum . this is a serious limitation and results , at least in part , from the range of masses of stars formed in the taurus sfr . we look forward to the capabilities of the alma mm - wave interferometric array which will open the rich orion sfr to investigation . conversion of the observationally derived parameter , spectral type , to that provided by the models , @xmath40 , is a lingering source of uncertainty , particularly at the lowest masses . since stars contracting to the main sequence represent a range of surface gravities , the conversion may depend on the stars age . relief from this problem should become available soon as the model calculations which include stellar atmospheres provide diagnostics such as color indices and model spectra , for closer comparison with the observations . we look forward to a fruitful decade ahead . we thank i. baraffe , j. bouvier , g. chabrier , c. dougados , m. forestini , f. palla , s. stahler , and l. siess for helpful conversations and communicating results before publication . this work began as a collaboration with g. duvert and f. mnard ; we thank them for their advice and help in its early stages . we thank the referee for a prompt and helpful report . this paper was begun during m.s.s sabbatical visit at the observatoire de grenoble ; m.s . thanks g. duvert and c. perrier for making it possible and for their hospitality . the work of m.s . was supported in part by nsf grants 94 - 17191 and 98 - 19694 . _ in memoriam : _ the plateau de bure interferometer was affected by two transportation tragedies in 1999 . these caused the death of 25 individuals including six members of the iram staff and two from insu - cnrs . we dedicate this work to the memory of the victims , especially our colleagues b.aubeuf , f.gillet , h.gontard , d.lazaro , r.prayer and p.vibert . llllll singles : & & & & & + mwc480 & a4 & 11.5 & 8460 & 1 & 1 + lkca15 & k5 & 0.74 & 4350 & 2 & + dl tau & k7 & 0.68 & 4060 & 3 & 2 + gm aur & k7 & 0.74 & 4060 & 4 & 3 + bp tau & k7 & 0.93 & 4060 & & 4 + dm tau & m1 & 0.25 & 3720 & 5 & + cy tau & m1 & 0.47 & 3720 & 3 & + binaries : & & & & & + gg tau aa & k7 & 0.84 & 4060 & 6 & 5 + gg tau ab & m0.5 & 0.71 & 3800 & & + uz tau e & m1 & 1.60 & 3720 & 3,7 & + lllllll mwc480 & 04@xmath88 & 58@xmath89 & 46.264@xmath90 & 29@xmath8 & 50@xmath91 & 37.03@xmath12 + lkca15 & 04 & 39 & 17.784 & 22 & 21 & 03.46 + dl tau & 04 & 33 & 39.074 & 25 & 20 & 38.14 + dm tau & 04 & 33 & 48.731 & 18 & 10 & 10.00 + cy tau & 04 & 17 & 33.727 & 28 & 20 & 46.95 + uz tau e & 04 & 32 & 43.070 & 25 & 52 & 31.14 + bp tau & 04 & 19 & 15.835 & 29 & 06 & 26.90 + llcc systemic velocity & @xmath92 ( km / s & @xmath93 & @xmath94 + orientation & pa@xmath28 ( deg ) & @xmath95 & @xmath96 + inclination & @xmath16 ( deg ) & @xmath97 & @xmath98 + outer radius & @xmath22 ( au ) & @xmath99 & @xmath100 + turbulent linewidth & @xmath101 ( km / s ) & @xmath102 & @xmath103 + + + + temperature at 100 au & @xmath104 ( k ) & @xmath105 & @xmath106 + temperature exponent & @xmath107 & @xmath108 & @xmath109 + + + + v sin@xmath16 at 100 au & @xmath110sin@xmath16 ( km / s ) & @xmath111 & @xmath112 + velocity exponent & @xmath20 & @xmath113 & @xmath114 + stellar mass & m@xmath4 ( @xmath24 ) & @xmath115 & @xmath116 + lrlrllrcc + mwc480 & @xmath117 & @xmath118&@xmath119&@xmath120&@xmath121 & @xmath122 & @xmath123 & @xmath124 + lkca15 & @xmath125&@xmath126 & @xmath127&@xmath128&@xmath129 & @xmath130&@xmath131&@xmath52 + dl tau@xmath132 & @xmath133 & @xmath134 & @xmath135&@xmath136&@xmath137 & @xmath138&@xmath57 & @xmath139 + gm aur & @xmath140 & @xmath141 & @xmath142&@xmath143&@xmath144 & @xmath145&@xmath146&@xmath54 + dm tau & @xmath147 & @xmath148 & @xmath149 & @xmath150&@xmath151 & @xmath112&@xmath114&@xmath57 + cy tau & @xmath152&@xmath153 & @xmath154 & @xmath155&@xmath156 & @xmath157 & @xmath158 & @xmath159 + bp tau & - & @xmath160 & @xmath161 & @xmath162 & @xmath163 & @xmath164 & @xmath165 & @xmath166 + + gg tau a & @xmath167&@xmath168 & @xmath169 & @xmath170&@xmath171 & @xmath172 & @xmath173 & @xmath174 + uz tau e & @xmath175 & @xmath176 & @xmath177 & @xmath178&@xmath179 & @xmath180&@xmath181&@xmath182 +	we have used the high sensitivity and resolution of the iram interferometer to produce sub - arcsecond @xmath0co j=2 - 1 images of 9 protoplanetary disks surrounding ttauri stars in the taurus - auriga cloud ( 7 singles and 2 binaries ) . the images demonstrate the disks are in keplerian rotation around their central stars . using the least square fit method described in guilloteau & dutrey ( 1998 ) , we derive the disk properties , in particular its inclination angle and rotation velocity , hence the dynamical mass . since the disk mass is usually small , this is a _ direct _ measurement of the stellar mass . typically , we reach an internal precision of 10% in the determinations of stellar mass . the over - all accuracy is limited by the uncertainty in the distance to a specific star . in a distance independent way , we compare the derived masses with theoretical tracks of pre - main - sequence evolution . combined with the mean distance to the taurus region ( 140 pc ) , for stars with mass close to 1 , our results tend to favor the tracks with cooler photospheres ( higher masses for a given spectral type ) . we find that in uz tau e the disk and the spectroscopic binary orbit appear to have different inclinations .
quantum bits ( qubits ) based on polarization or spatial degrees of freedom of optical modes have several advantages : they are easily manipulated and measured ; they exist in a low noise environment and ; they are easily communicated over comparitively long distances . recently considerable progress has been made on implementing two qubit gates in optics using the measurement induced non - linearities proposed by knill , laflamme and milburn @xcite . non - deterministic experimental demonstrations have been made @xcite and theory has found significant ways to reduce the resource overheads @xcite . nevertheless , the number of photons and gate operations required to implement a near deterministic two qubit gate remains high . a possible solution to this problem is the optical quantum zeno gate suggested by franson et al @xcite , @xcite . this gate uses passive two - photon absorption to suppress gate failure events associated with photon bunching at the linear optical elements , using the quantum zeno effect @xcite . in principle a near deterministic , high fidelity control - sign ( cz ) gate can be implemented between a pair of photonic qubits in this way . however , the slow convergence of the zeno effect to the ideal result , with ensuing loss of fidelity , and the effect of single photon loss raises questions about the practicality of this approach . here we consider a model of the gate that includes the effects of finite two - photon absorption and non - negligible single photon absorption . we obtain analytic expressions for the fidelity of the gate and its probability of success in several scenarios and show how the inclusion of optical distilling elements @xcite can lead to high fidelity operation under non - ideal conditions for tasks such as cluster state construction @xcite . the paper is arranged in the following way . we begin in the next section by introducing our model in an idealized and then more realistic setting and obtain results for a free - standing cz gate . in section 3 we focus on using the gate as a fusion element @xcite for the construction of , for example , optical cluster states . we introduce a distillation protocol that significantly improves the operation of the gate in this scenario . in section 4 we summarize and conclude . franson et al @xcite suggested using a pair of optical fibres weakly evanescently coupled and doped with two - photon absorbing atoms to implement the gate . as the photons in the two fibre modes couple the occurence of two photon state components is suppressed by the presence of the two - photon absorbers via the zeno effect . after a length of fibre corresponding to a complete swap of the two modes a @xmath2 phase difference is produced between the @xmath3 term and the others . if the fibre modes are then swapped back by simply crossing them , a cz gate is achieved . we model this system as a succession of @xmath4 weak beamsplitters followed by 2-photon absorbers as shown in fig . [ fig : ourcsign ] . as @xmath5 the model tends to the continuous coupling limit envisaged for the physical realization . the gate operates on the single - rail encoding @xcite for which @xmath6 and @xmath7 with the kets representing photon fock states . [ fig : cz ] shows how the single rail cz can be converted into a dual rail cz with logical encoding @xmath8 and @xmath9 with @xmath10 a fock state with @xmath11 photons in the horizontal polarization mode and @xmath12 photons in the vertical . + the general symmetric beam splitter matrix has the form : @xmath13\ ] ] according to figure [ fig : ourcsign ] , after the first beam splitter , the four computational photon number states become : @xmath14 to illustrate the operation of the gate we first assume ideal two - photon absorbers , i.e. they completely block the two - photon state components but do not cause any single photon loss . propagation through the first pair of ideal two - photon absorbers gives the mixed state @xmath15 where @xmath16 is the evolved two - mode input state obtained for the case of no two - photon absorption event and @xmath17 is the vacuum state obtained in the case a two - photon absorption event occurs . the individual components of @xmath16 transform as @xmath18 notice that , because we are embedded in a dual rail circuit , we can distinguish between the @xmath19 state that corresponds to input state @xmath20 and the @xmath21 state that results from two - photon absorption of input state @xmath22 . equation ( [ eqn : complete ] ) describes the transformation of each unit , hence repeating the procedure @xmath4 times gives , latexmath:[\[\begin{aligned } describing the transformations giving the evolved input state after @xmath4 units , @xmath24 . there are three conditions to satisfy for building a cz gate . the first condition is @xmath25 " , so that @xmath26 and @xmath27 . the second condition is @xmath28 " ( equivalently , @xmath29 ) , where @xmath30 is any integer , so that @xmath31 and @xmath32 and @xmath33 . phase shifters are needed to correct the sign of the output state of @xmath34 and @xmath35 for odd @xmath30 , but here we simply set @xmath36 . the last condition is @xmath37 " ( i.e. @xmath38 ) , such that a minus sign is induced on @xmath39 . this condition is always true because we are using many weak beam splitters ( i.e. @xmath40 is small ) . let @xmath41 , then swapping the fibres gives the transformations @xmath42 ) on the probability amplitude of the @xmath39 state . if there is some way to herald failure , i.e. two - photon absorption events , then the fidelity of the gate will be @xmath43 , where @xmath44 is the target state , and the probability of success will be @xmath45 . on the other hand if two - photon absorption events are unheralded then the fidelity will be @xmath46 . for simplicity we consider the equally weighted superposition input state @xmath47 . the corresponding @xmath24 after the zeno - cz gate is @xmath48 , to be compared with the target state @xmath49 . the heralded fidelity and probability of success are then @xmath50 and @xmath51 respectively . as @xmath4 becomes very large and hence tends to the continuous limit , @xmath52 tends to one , and so both @xmath53 and @xmath54 approach one . the previous analysis is clearly unrealistic as it assumes infinitely strong two - photon absorption but negligible single photon absorption . we now include the effect of finite two - photon absorption and non - negligible single photon loss . let @xmath55 and @xmath56 be the probability of single photon and two - photon transmission respectively for one absorber . here the parameter @xmath57 , where @xmath58 is the length of the absorber and @xmath59 is the corresponding proportionality constant related to the absorption cross section . furthermore , @xmath0 specifies the relative strength of the two transmissions and relates them by @xmath60 . now each unit of weak beam splitter and absorbers does the following transformation on the computational states of @xmath61 @xmath62 repeating the procedure @xmath4 times with the aforementioned conditions on @xmath40 gives the following latexmath:[\[\begin{aligned } expression for @xmath52 is given by : @xmath64 and we have suppressed the explicit form of the @xmath65 state components as they lie outside the computational basis and so do not explicitly contribute to the fidelity . these expressions can be used to calculate the unheralded fidelity , the heralded fidelity and probability of success . our numerical evaluations are all carried out in the ( near ) continuous limit of large @xmath4 . for a free - standing gate , as depicted in fig.[fig : cz ] , gate failure events are not heralded , thus the unheralded fidelity is appropriate to consider . the fidelity is a function of @xmath66 . as the length of the interaction region is increased ( @xmath66 increased ) the effective strength of the two - photon absorption is increased leading to an improvement in the heralded fidelity , @xmath53 . however , at the same time , the level of single photon absorption is also increasing with the length , acting to decrease the probability of success , @xmath54 . as the unheralded fidelity is @xmath67 , there is a trade - off between these two effects leading to an optimum value for @xmath66 for sufficiently large @xmath0 . an example of the dependence is shown in fig.[fig : fidelity_vs_lambda ] . the fidelity is plotted as a function of @xmath0 with @xmath66 optimized for each point in fig.[fig : fidelity_vs_kappa ] . for large ratios of two - photon absorption to single - photon absorption , @xmath0 , we tend to the ideal case of unit fidelity . however , the conditions required are demanding with absorption ratios of a million to one required for @xmath68 and 100 million to one for @xmath69 . recent estimates suggest @xmath0 s of ten thousand to one may be achievable @xcite , well short of these numbers . in the following we will consider a different scenario in which the gate can be usefully employed with less stringent conditions on @xmath0 . we have seen that the requirements on high fidelity operation for the free - standing gate are quite extreme . we now consider an alternate scenario in which probability of success is traded - off against fidelity by heralding failure events through direct detection . in particular we consider using the zeno gate to implement the fusion technique @xcite . fusion can be used to efficiently construct cluster states @xcite , or re - encode parity states @xcite . we will specifically consider cluster state construction here . essentially , the gate is used to make a bell measurement on a pair of qubits , as depicted in fig.[fig : fusion ] . one of the qubits comes from the cluster we are constructing , whilst the other comes from a resource cluster state , in a known logical state . the bell measurement has the effect of fusing " the resource state onto the existing state . by careful choice of the resource state , large 2-dimensional cluster states , suitable for quantum computation , can be constructed @xcite . because the bell measurement ends with the direct detection of the qubits , the loss of one or both of the photons , or the bunching of two photons in a single qubit mode can immediately be identified in the detection record , and hence failure events will be heralded . effectively we will postselect the density operator @xmath70 , where @xmath71 is the component of the output state which remains in the computational basis and @xmath72 are all the components that do not . the measurement record then allows us to herald the first term of the density operator as successful operation , with fidelity @xmath73 and probability of success of @xmath54 , and the second term as failure . we now consider techniques for improving the heralded fidelity of the gate and then evaluate its performance as a fusion gate . from equation ( [ eqn : incomplete ] ) , we can see that @xmath74 lowers the probability amplitude of the four computational states unevenly as previously discussed by jacobs et al @xcite . by distilling the states with beam splitters and detectors @xcite ( see figure [ fig : cz_distill ] ) , where each beam splitter has a transmission coefficient equal to @xmath75 , the four computational states of @xmath71 become : @xmath76 and @xmath77 respectively . as shown previously , after the cz gate and single photon distillation , the input state @xmath47 becomes @xmath80 . now we require two - photon distillation to renormalise the input state by inducing @xmath52 on the other three computational states as shown in figure [ fig : two_photon_distill ] . to do so , we first apply a bit - flip on the control qubit and then apply a @xmath52-gate ( see figure [ fig : tau_gate ] ) and a single photon distiller on the control qubit with transmission coefficient @xmath81 and another single photon distiller on the target qubit with transmission coefficient @xmath82 and then undo the previous bit - flip by applying another bit - flip on the control qubit . the @xmath52-gate does the same operation as the aforementioned cz gate ( excluding the single photon distillation ) except that no minus sign is induced on the output of @xmath83 . the construction of a @xmath52-gate is described in the next subsection . in summary the two - photon distillation circuit does the following : @xmath84 becomes @xmath85 . now the state can be renormalised to achieve unit fidelity _ independent _ of @xmath66 . the explicit expression for the probability of success is @xmath86 figure [ fig : two_photon_distill_psuccess ] shows the probability of success of this gate for different values of @xmath0 . + we can construct a @xmath52-gate with two 50 - 50 beam splitters , a pair of two - photon absorbers , and some phase shifters , as shown in figure [ fig : tau_gate ] . the first beam splitter performs @xmath87 , @xmath88 and @xmath89 . the pair of two - photon absorbers then induce @xmath81 on both @xmath34 and @xmath35 due to single photon loss , and induce @xmath90 on @xmath91 due to both single photon and two - photon loss . the second beam splitter undoes the operation of the first beam splitter . then with some phase shifters to correct the relative phase between the terms and having @xmath92 , we have a @xmath52-gate that does the following operation : the fusion approach is important because it is the most efficient method known for performing quantum computation using only linear optics . linear optics allows a partial bell measurement to be made with a probability of success of 50% ( assuming ideal detectors ) . in addition the failure mode measures the qubits in the computational basis , which does not affect the state of the remaining qubits in the cluster or parity state . thus a failure event only sacrifices a single qubit from the cluster being constructed and the probability of destroying @xmath94 qubits in the process of achieving a successful fusion is @xmath95 . in contrast , many of the failure events for the zeno gate will simply erase the photon giving no knowledge about its state . for simplicity , and to be conservative , we will assume all events lead to complete erasure of the photon state . in order to recover from this situation the adjoining qubit in the cluster must be measured in the logical basis , thus removing the affect of the erasure @xcite . this means that every failure event sacrifices two qubits from the cluster being constructed and the probability of destroying @xmath94 qubits in the process of achieving a successful fusion is @xmath96 . requiring @xmath97 we estimate that the zeno gate must have @xmath98 to offer an advantage over linear optics . + we can make one final improvement to the set - up by relocating the distillation process for the resource qubit to offline ( see fig.5 ) , which boosts the probability of success . the probability of success is then given by @xmath99 . the plot for the probability of success versus @xmath0 and optimal @xmath66 versus @xmath0 are shown in figure [ fig : psuccessvsnoffline ] and figure [ fig : optimalvsnoffline ] respectively . + the break even point between linear optics and the zeno gate is when @xmath100 , such that the probability of success is about @xmath101 . when @xmath102 the probability of success is about @xmath103 . thus we conclude that an absorption ratio of ten thousand to one or more would produce a zeno gate with significant advantage over linear fusion techniques . + * conclusion * + in this paper , we have modelled franson et al s cz gate with a succession of @xmath4 weak beam - splitters followed by two - photon absorbers , in the ( near ) continuous limit of large @xmath4 . we analysed this cz gate for both the ideal two - photon absorption case and the imcomplete two - photon absorption with single photon loss case , giving analytical and numerical results for the fidelity and probability of success . the result shows that for a free - standing gate we need an absorption ratio @xmath0 of a million to one to achieve @xmath104 and 100 million to one to achieve @xmath105 , where recent estimate only suggests that @xmath106 may be achievable . we therefore employ this gate for qubit fusion , where the requirement for @xmath0 is less restrictive . with the help of partially offline one - photon and two - photon distillations , we can achieve a cz gate with unity fidelity and with probability of success is about 0.87 for @xmath102 . we conclude that when employed as a fusion gate , the zeno gate could offer significant advantages over linear techniques for reasonable parameters . + * acknowledgement * + we thank w.j.munro , a.gilchrist and c.myers for useful discussions . this work was supported by the australian research council and the dto - funded u.s . army research office contract no . w911nf-05 - 0397 . + 99 e. knill , r. laflamme , and g.j . milburn , _ nature _ * 409 * , 46 - 52 ( 2001 ) . j.l.obrien , g.j.pryde , a.g.white , t.c.ralph , d.branning , nature * 426 * , 264(2003 ) . pittman , m.j . fitch , b.c . jacobs , and j.d . franson , phys . a * 68 * , 032316 ( 2003 ) . s.gasparoni , j .- w.pan , p.walther , t.rudolph , and a.zeilinger phys . lett . * 93 * , 020504 ( 2004 ) . n.yoran and b.reznik , phys . lett . * 91 * , 037903 ( 2003 ) . nielsen , phys . rev . lett . * 93 * , 040503 ( 2004 ) . a.j.f.hayes , a.gilchrist , c.r.myers and t.c.ralph , j.opt.b * 6 * , 533 ( 2004 ) . d.e.browne and t.rudolph , phys . lett . * 95 * , 010501 ( 2005 ) . franson , b.c . jacobs , and t.b . pittman , _ pra _ * 70 * , 062302 ( 2004 ) b.c . jacobs , t.b . pittman , and j.d . franson , _ pra _ * 74 * , 010303(r ) ( 2006 ) w. kaiser and c. garrett , _ prl _ * 7 * 229 ( 1961 ) r.t . thew and w.j . munro , _ pra _ * 63 * , 030302(r ) ( 2001 ) a.p . lund , t.c . ralph , _ pra _ , * 66 * , 032307 ( 2002 ) . j.d . franson and s.m . quant - ph _ 0603044 ( 2006 ) r. raussendorf and h.j . briegel , _ prl _ * 86 * 5188 ( 2001 ) t.c . ralph , a.j.f . hayes and a. gilchrist , _ prl _ * 95 * 100501 ( 2005 ) c.m . dawson , h.l . haselgrove , and m.a . nielsen , _ prl _ 96 , 020501 ( 2006 ) l.m . duan and r. raussendorf , _ prl _ * 95 * , 080503 ( 2005 ) y.l . lim , s.d . barrett , a. beige , p. kok , l.c . kwek , _ pra _ * 73 * , 012304 ( 2006 )	we have modelled the zeno effect control - sign gate of franson et al ( pra 70 , 062302 , 2004 ) and shown that high two - photon to one - photon absorption ratios , @xmath0 , are needed for high fidelity free standing operation . hence we instead employ this gate for cluster state fusion , where the requirement for @xmath0 is less restrictive . with the help of partially offline one - photon and two - photon distillations , we can achieve a fusion gate with unity fidelity but non - unit probability of success . we conclude that for @xmath1 , the zeno fusion gate will out perform the equivalent linear optics gate .
the classical and quantum propagation of strings in curved spacetimes has attracted a great deal of interest in recent years . the main complication , as compared to the case of flat minkowski space , is related to the non - linearity of the equations of motion . it makes it possible to obtain the complete analytic solution only in a very few special cases like conical spacetime @xcite and plane wave / shock wave backgrounds @xcite . there are however also very general results concerning integrability and solvability for maximally symmetric spacetimes @xcite and gauged wzw models @xcite . these are the exceptional cases , generally the string equations of motion in curved spacetimes are not integrable and even if they are , it is usually an extremely difficult task to actually separate the equations , integrate them and finally write down the complete solution in closed form . fortunately there are different ways to `` attack '' a system of coupled non - linear partial differential equations . besides numerical methods , which will not be considered here , there are essentially two ways to proceed : an approximative method was first developed by de vega and snchez @xcite ( see also @xcite ) . the idea was to expand the target space coordinates around a special solution , usually taken to be the string center of mass , and then to ( try to ) solve the string equations of motion and constraints order by order . this can be done , at least up to first or second order in the expansion , for most of the known gravitational and cosmological spacetimes . for a review of the method and its applications , see for instance @xcite . another approximative method , which is however not un - related to the first one , was considered in refs.@xcite . it consists of a large scale - factor expansion in spatially flat frw - universes and an important result of the analysis was the discovery of `` extremely unstable strings '' of negative pressure in inflationary universes . as already mentioned there are a few examples where the equations of motion and constraints can actually be solved completely . if this is not possible one can try to find special , but exact , solutions by making an ansatz . if the ansatz is properly chosen , i.e. such that it exploits the symmetries of the spacetime , it may reduce the original system of coupled non - linear partial differential equations to something much simpler , for instance decoupled non - linear ordinary differential equations , and it may then be possible to find the complete solution of this reduced system . this method has been used for stationary open strings @xcite and for circular strings in a variety of curved spacetimes @xcite . when using this method , one somehow has to argue that the special solutions under consideration are representative in the sense that their physical properties are more general than the solutions themselves . this is done by comparing with the results obtained from the approximative methods , so it is very important to have , and to use , both kinds of methods . in this talk i will present material bases on both exact and approximative methods and i will try to take a physical point of view and address the question : what are the strings actually doing in the curved spacetimes ? the talk is organized as follows . in section 2 we consider circular strings in stationary axially symmetric backgrounds . the main result is the effective potential determining the circular string radius . the main part of the talk concerns circular strings in de sitter space , as indicated by the title . the material is based on refs.@xcite and is presented in section 3 . we first consider the energy and pressure of the different types of strings . we then turn to the mathematical solutions and their physical interpretation , following closely ref.@xcite . finally we consider briefly the propagation of small perturbations around the circular strings . in section 4 we consider circular strings in the background of the @xmath0 dimensional black hole anti de sitter spacetime , recently found by baados et . @xcite and in section 5 we compare with circular strings in ordinary black hole and anti de sitter spacetimes . in section 6 we first refine the approximative string perturbation approach of de vega and snchez @xcite by eliminating , from the beginning , the perturbations in the direction of the geodesic of the string center of mass . we then discuss some of the results obtained in ref.@xcite , using this method , and we compare with the results obtained for circular strings . the main conclusions of the talk are compactly summarized in tables i , ii , iii . in this section we consider circular strings embedded in stationary axially symmetric backgrounds . the analysis is carried out in @xmath0 dimensions , but the results will hold for the equatorial plane of higher dimensional backgrounds as well . to be more specific we consider the following line element : @xmath2 that will be general enough for our purposes here . the circular string ansatz , consistent with the symmetries of the background , is taken to be : @xmath3 where the three functions @xmath4 and @xmath5 are to be determined by the equations of motion and constraints . the equations of motion lead to : @xmath6 @xmath7 @xmath8 while the constraints become : @xmath9 @xmath10 this system of second order ordinary differential equations and constraints is most easily described as a hamiltonian system : @xmath11 supplemented by the constraints : @xmath12 the function @xmath5 introduced in eq . ( 2.2 ) does not represent a physical degree of freedom . it describes the `` longitudinal '' rotation of the circular string and is therefore a pure gauge artifact . this interpretation is consistent with eq . ( 2.6 ) saying that there is no angular momemtum @xmath13 . the hamilton equations of the two cyclic coordinates @xmath14 and @xmath15 are : @xmath16 as well as : @xmath17 where @xmath18 is an integration constant and we used eq . ( 2.6 ) . the two functions @xmath19 and @xmath5 are then determined by : @xmath20 @xmath21 that can be integrated provided @xmath22 is known . ( 2.6 ) and eqs . ( 2.9)-(2.10 ) , the hamilton equation of @xmath23 becomes after one integration : @xmath24 so that @xmath22 can be obtained by inversion of : @xmath25}}.\ ] ] for the cases that we will consider in the following , eq . ( 2.12 ) will be solved in terms of either elementary or elliptic functions . by definition of the potential @xmath26 eq . ( 2.11 ) , the dynamics takes place at the @xmath27axis in a @xmath28diagram . the first thing to do for a stationary axially symmetric background under consideration , is therefore to find the zeros of the potential . then one can try to solve the equations of motion and finally extract the physics of the problem we close this section by the following interesting observation : insertion of the ansatz , eq . ( 2.2 ) , using the results of eqs . ( 2.9)-(2.11 ) in the line element , eq . ( 2.1 ) , leads to : @xmath29 we can then identify the invariant string size as : @xmath30 @xmath0 dimensional de sitter space is a 3 dimensional hyperboloid embedded in 4 dimensional minkowski space : @xmath31,\;\;\;\;\;\;\;\ ; \eta_{\mu\nu}q^\mu q^\nu=1.\ ] ] the hyperboloid coordinates cover the whole manifold contrary to the comoving coordinates : @xmath32,\ ] ] and the static coordinates : @xmath33 for our purposes it is however most convenient to start out with the static coordinates since then we can use directly the results of section 2 for the circular strings . afterwards , we then have to transform the solutions back to the hyperboloid using the appropriate coordinate transformations inside and outside the horizon . in static coordinates we can now immediately write down the potential , eq . ( 2.11 ) , determining the dynamics of circular strings in de sitter space : @xmath34 see fig.1 . for @xmath35 the top of the potential is above the @xmath27axis and there will be oscillating solutions to the left of the barrier and bouncing solutions to the right . the bouncing solutions will re - expand towards infinity . for @xmath36 on the other hand , the top of the potential is below the @xmath27axis and therefore the potential does not act as a barrier . strings can expand from @xmath37 towards infinity and very large strings can collapse . finally if @xmath38 there is a ( unstable ) stationary string at the top of the potential as well as expanding and collapsing strings . it is instructive to compare with the circular string dynamics in minkowski space . in that case the potential is given by @xmath39 and only oscillating strings exist . physically the difference is not so difficult to understand . in flat minkowski space the dynamics of a circular string is determined by the string tension only , that will always try to contract the string . this leads to strings oscillating between @xmath37 and some maximal size depending on the energy . in de sitter space there is an opposite effect namely the expansion of the universe , that will try to expand the string . this gives rise to a region where the tension is strongest ( to the left of the barrier ) , a region where the expansion of the universe is strongest ( to the right of the barrier ) and there may even be an exact balance of the two opposite forces , implying the existence of a stationary solution . before coming to the mathematical solution of the equations of motion , we consider the energy and pressure of the strings . the spacetime string energy - momentum tensor is : @xmath40 after integration over a spatial volume that completely encloses the string , the energy - momentum tensor takes the form of a fluid : @xmath41 where , in the comoving coordinates introduced in eq . ( 3.2 ) : @xmath42 @xmath43 represent the energy and pressure , respectively . from these expressions we can actually get a lot of information about the energy and pressure , without using the explicit time evolution of the strings . from the coordinate transformations between comoving and static coordinates we get the explicit expressions @xcite : @xmath44 @xmath45 both the energy and the pressure now depend on the string radius @xmath23 and the velocity @xmath46 the latter can however be eliminated using eq . ( 2.11 ) and eq . ( 3.4 ) . let us first consider a string expanding from @xmath37 towards infinity . this corresponds to a string with @xmath47 and @xmath48 , see fig.1 . for @xmath37 we find @xmath49 and @xmath50 thus the equation of state is @xmath51 this is like ultra - relativistic matter . as the string expands , the energy soon starts to increase while the pressure starts to decrease and becomes negative , see fig.2 . for @xmath52 we find @xmath53 and @xmath54 thus not surprisingly , we have recovered the equation of state of extremely unstable strings @xmath55 @xcite . now consider an oscillating string , i.e. a string with @xmath56 in the region to the left of the potential barrier , fig.1 . the equation of state near @xmath37 is the same as for the expanding string , but now the string has a maximal radius : @xmath57 for @xmath58 we find : @xmath59 corresponding to a perfect fluid type equation of state : @xmath60 notice that @xmath61,$ ] with @xmath62 decribing a string at the bottom of the potential while @xmath63 decribes a string oscillating between @xmath37 and the top of the potential barrier ( in this case the string actually only makes one oscillation @xcite ) . for @xmath64 the equation of state , eq.(3.13 ) , near the maximal radius reduces to @xmath55 in the other limit @xmath65 we find , however , @xmath66 corresponding to cold matter . for the oscillating strings we can also calculate the average values of energy and pressure by integrating over a full period , see fig.3 . since a @xmath67-integral can be converted into a @xmath23-integral , using eq . ( 2.11 ) and eq . ( 3.4 ) , the average values can be obtained without using the exact @xmath67-dependence of the string radius . the average energy becomes : @xmath68 where @xmath69 is the period and @xmath70 is the complete elliptic integral of the third kind ( using the notation of gradshteyn @xcite ) . the average pressure is zero , thus in average the oscillating strings describe cold matter . in minkowski space , for comparison , it can easily be shown @xcite that the energy is constant while the pressure depends on the string radius . the average pressure is however zero , just as in de sitter space . ( this and the following subsection follow closely ref.@xcite . ) + we now come to the explicit mathematical solutions . in the general case equations ( 2.11 ) and ( 3.4 ) are solved by : @xmath71 where @xmath72 is the weierstrass elliptic @xmath72-function @xcite : @xmath73 with invariants : @xmath74 and discriminant : @xmath75 it is convenient to consider separately the 3 cases @xmath76 , @xmath77 and @xmath78 , see fig.1 . @xmath79 : in this case the weierstrass function reduces to a hyperbolic function and eq . ( 3.15 ) becomes : @xmath80 two real independent solutions are obtained by the choices @xmath81 and @xmath82 , respectively : @xmath83 @xmath84 notice that : @xmath85 these are the 2 solutions originally found by de vega , snchez and mikhailov @xcite , corresponding to @xmath86 and @xmath87 in their notation , respectively . the interpretation of these solutions as a function of the world - sheet time @xmath67 is clear from fig.1 : the solution , eq . ( 3.20 ) , expands from @xmath88 towards infinity and then contracts until it reaches its original size . the solution , eq . ( 3.21 ) , contracts from @xmath89 until it collapses . it then expands again until it reaches its original size . the physical interpretation , that is somewhat more involved , was described in ref.@xcite and will be shortly reviewed in subsection 3.2 . there is actually also a stationary solution for @xmath90 , i.e. a string sitting on the top of the potential , see fig.1 . this solution with constant string size @xmath91 was discussed in ref.@xcite and will not be considered here . @xmath92 : here two real independent solutions are obtained by the choices @xmath81 and @xmath93 , respectively : @xmath94 @xmath95 where @xmath96 is the imaginary semi - period of the weierstrass function . it is explicitly given by @xcite : @xmath97 notice that : @xmath98 where @xmath99 is the real semi - period of the weierstrass function @xcite : @xmath100 and @xmath101 and @xmath102 are the complete elliptic integrals of first kind . the interpretation of these solutions as a function of @xmath67 is clear from fig.1 : the solution , eq . ( 3.23 ) , oscillates between infinity and its minimal size @xmath103 at the boundary of the potential , while the solution , eq . ( 3.24 ) , oscillates between @xmath104 and its maximal size @xmath105 . the physical interpretation will be considered in subsection 3.2 . @xmath106 : in this last case two real independent solutions are obtained by the choices @xmath81 and @xmath107 , respectively : @xmath94 @xmath108 where @xmath109 takes the explicit form : @xmath110 notice that : @xmath111 where : @xmath112 it should be stressed that in this case the primitive semi - periods are @xmath113 and @xmath114 , i.e. @xmath115 spans a fundamental period parallelogram in the complex plane . the interpretation of the solutions , eqs . ( 3.28)-(3.29 ) , as a function of @xmath67 follows from fig.1 : both of them oscillates between zero size ( collapse ) and infinite size ( instability ) . the physical interpretations will follow in subsection 3.2 . @xmath79 : we first consider the @xmath116-solution , eq . the hyperboloid time is obtained by integrating eq . ( 2.10 ) and transforming back to the hyperboloid coordinates : @xmath117 when we plot this function ( fig.4 . ) we see that the string solution actually describes 2 strings ( i and ii ) @xcite , since @xmath67 is a two - valued function of @xmath118 . for both strings the invariant string size is : @xmath119 but string i corresponds to @xmath120-\infty,0[\;$ ] and string ii to @xmath1200,\infty[\;$ ] . therefore , @xmath121 corresponds to @xmath122 for string i , but to @xmath123 for string ii . more generally , when @xmath121 the invariant size grows indefinitely for string i , while it approaches a constant value for string ii . we conclude that string i is an unstable string for @xmath121 , while string ii is a stable string . more details about the connection between hyperboloid time and world - sheet time for these solutions can be found in ref.@xcite . let us consider now the comoving time of the solution @xmath116 in a little more detail : @xmath124 when we plot this function ( fig.4 . ) we find that @xmath67 is a three - valued function of @xmath125 . what happens is that the time interval @xmath1200,\infty[\;$ ] for string ii splits into two parts . these features are easily understood when returning to the effective potential , fig.1 . : string i starts at @xmath88 for @xmath126 , it then expands through the horizon @xmath127 at : @xmath128 and continues towards infinity for @xmath129 . string ii starts at infinity for @xmath130 and contracts through the horizon at : @xmath131 this behaviour , approaching the horizon from the outside , corresponds to the going backwards in comoving time - part of fig.4 . string ii then continues contracting from @xmath127 at : @xmath132 until it reaches @xmath133 at @xmath134 . we now consider briefly the @xmath135-solution , eq . ( 3.21 ) . in this case the hyperboloid time is given by : @xmath136 which is a monotonically increasing function of @xmath67 . the @xmath135-solution therefore describes only one string . the proper size is given by : @xmath137 we can also express this solution in terms of the comoving time : @xmath138 but since everything now takes place well inside the horizon , this will not really give us more insight . the string starts with @xmath89 for @xmath139 it then contracts until it collapses for @xmath140 and expands again and eventually reaches @xmath89 for @xmath141 @xmath92 : using the notation introduced in eqs . ( 3.11)-(3.12 ) the solutions , eqs . ( 3.23)-(3.24 ) , can be written as : @xmath142},\ ] ] @xmath143,\ ] ] where @xmath144 consider first the @xmath116-solution , eq . it is clear from eq . ( 3.26 ) and the periodicity in general that we have infinitely many branches @xmath145,\ ; [ 2\omega,4\omega],\; ... $ ] . we will see in a moment that each of these branches actually corresponds to one string , that is , the @xmath116-solution describes infinitely many strings . for that purpose we will need the hyperboloid time and the comoving time as a function of @xmath67 . both of them are expressed in terms of the static coordinate time @xmath14 , that is obtained by integrating eq . ( 2.10 ) : @xmath146 where @xmath147 and @xmath148 are the weierstrass @xmath147 and @xmath148-functions @xcite and @xmath149 is a real constant obeying @xmath150=\mu$ ] , i.e. @xmath149 is expressed as an incomplete elliptic integral of the first kind . the expression , eq.(3.44 ) , can be further rewritten in terms of theta - functions @xcite : @xmath151 and finally as : @xmath152 where @xmath153 . in the latter expression we have isolated all the real singularities in the first term . to be more specific we see that the static coordinate time is singular for @xmath154 , where @xmath155 is an integer , with the asymptotic behaviour : @xmath156 on the other hand @xmath157 is completely regular at the boundaries of the branches , i.e. for @xmath158 . these results can be easily translated to the hyperboloid time @xcite : @xmath159 where : @xmath160 notice that the singularities , eq . ( 3.47 ) , that originated from the zeros of @xmath161 have canceled in @xmath162 so that @xmath163 is finite . @xmath162 blows up for @xmath164 , like : @xmath165 where @xmath155 is again an integer . this demonstrates that the world - sheet time @xmath67 is actually an infinite valued function of @xmath118 , and that the solution @xmath116 therefore describes _ infinitely many strings _ ( see fig.5 ) . this should be compared with the @xmath90 case where we found a solution describing two strings . in that case the two strings were of completely different type and had completely different physical interpretations . in the present case we find infinitely many strings but they are all of the same type . in the branch @xmath166 $ ] ( say ) the string starts with infinite string size at @xmath167 . it then contracts to its minimal size @xmath103 and reexpands towards infinity at @xmath168 . this solution , and the infinitely many others of the same type , are unstable strings . the comoving time of the @xmath116-solution is given by @xcite : @xmath169 @xmath170 @xmath171 it can be shown that @xmath172 and @xmath173 . it follows that : @xmath174 the comoving time , eq . ( 3.51 ) , is singular at @xmath175 but regular at @xmath176 and similarly in the other branches , fig.5 . therefore , the interpretation of the string solution in the branch @xmath166 $ ] ( say ) , as seen in comoving coordinates , is as follows : the string starts with infinite size at @xmath177 . it then contracts and passes the horizon from the outside at @xmath178 . the string now continues contracting from the inside of the horizon at @xmath179 until it reaches the minimal size at : @xmath180 from now on the string expands again . it passes the horizon from the inside after finite comoving time and continues towards infinity for @xmath181 . it is an interesting observation that the comoving time is not periodic in @xmath67 , i.e. @xmath182 , although the string size is . explicitly we find : @xmath183 this means that @xmath184 really describes infinitely many strings with different invariant size at a given comoving time . to be more specific let us consider a fixed comoving time @xmath125 and the corresponding world - sheet times : @xmath185 where @xmath186 . taking for simplicity a comoving time @xmath187 we have ( see fig.5 . ) : @xmath188 to the lowest orders we find from eq . ( 3.51 ) : @xmath189 so that : @xmath190\ ] ] the invariant string sizes are then : @xmath191\ ] ] i.e. they are separated by a multiplicative factor . this expression , of course , is only valid as long as @xmath192 , so @xmath193 should not be too large . we now consider the @xmath135-solution , eq . ( 3.43 ) . in this case the dynamics takes place well inside the horizon . the possible singularities of the hyperboloid time @xmath194 and the comoving time @xmath195 therefore coincide with the singularities of the static coordinate time @xmath196 . the static coordinate time is again obtained from eq . ( 2.10 ) , which we first rewrite as : @xmath197 integration leads to : @xmath198 where @xmath199 is a complex constant obeying @xmath200 i.e. @xmath201=1/\nu$ ] . it follows that @xmath202 where @xmath149 is real and @xmath150=\mu$ ] . again we can express the static coordinate time in terms of theta - functions @xcite : @xmath203 or in terms of the jacobi zeta - function @xmath204 @xcite : @xmath205 where @xmath153 . in this form we see explicitly that @xmath196 consists of a linear term plus oscillating terms . the comoving time takes the form : @xmath206)+ht_+(\tau)&\nonumber\end{aligned}\ ] ] @xmath207 + \log\mid\frac{\omega\vartheta_4(\omega(\tau - y))\vartheta'_1(0 ) } { \pi\vartheta_1(\omega y)\vartheta_4(\omega\tau)}\mid . & \end{aligned}\ ] ] notice that the argument of the @xmath208 has no real zeros : @xmath209 + \log\mid\frac{\omega\vartheta'_1(0)}{\pi\vartheta_1(\omega y)}\mid & \nonumber\end{aligned}\ ] ] @xmath210 the static coordinate time and the cosmic time are therefore completely regular functions of @xmath67 , and it follows that the string solution @xmath135 , which is _ oscillating regularly _ as a function of world - sheet time @xmath67 , is also oscillating regularly when expressed in terms of hyperboloid time or comoving time . this solution represents one _ stable _ string . @xmath106 : the analysis here is very similar to the analysis of the @xmath211solution in the @xmath212case so we shall not go into it here . the results are summarized in table i , together with the results from the other cases . * table i * circular string evolution in de sitter spacetime . for each @xmath213 there exists two independent solutions @xmath116 and @xmath135 : [ cols= " < , < , < " , ] using the covariant approach of frolov and larsen @xcite , we have considered small perturbations propagating along the circular strings in black hole and cosmological spacetime backgrounds @xcite . in the case of the @xmath1 dimensional de sitter space introduce two normal vectors @xmath214 and @xmath215 perpendicular to the string world - sheet : @xmath216 and then consider only physical perturbations : @xmath217 where @xmath218 and @xmath219 are the perturbations as seen by an observer travelling with the circular string . after fourier expanding @xmath220 : @xmath221 it can be shown that @xcite : @xmath222 @xmath223 determining the comoving perturbations . these equations have been discussed in detail in refs.@xcite , so we shall just give one simple result here . for @xmath52 the brackets in eqs . ( 3.69)-(3.70 ) become negative . this means that the perturbations develop imaginary frequencies and grow indefinitely . however , by considering the detailed solutions it turns out that the perturbations grow with the same rate as the radius of the underlying circular string ( which by the way grows with the same rate as the universe ) so although the perturbations grow , the circular shape of the string is actually stable . more details can be found in refs.@xcite . we now consider the circular string dynamics in the 2 + 1 black hole anti de sitter ( bh - ads ) spacetime recently found by baados et . this spacetime background has arised much interest recently . it describes a two - parameter family ( mass @xmath224 and angular momentum @xmath225 ) of black holes in 2 + 1 dimensional general relativity with metric : @xmath226 it has two horizons @xmath227 and a static limit @xmath228 defining an ergosphere , as for ordinary kerr black holes . using the general formalism of section 2 , we can immediately read off the potential ( see fig.6 . ) : @xmath229 the potential , eq . ( 4.2 ) , has a global minimum between the two horizons : @xmath230 which is always negative , since we only consider the case when @xmath231 ( otherwise there are no horizons ) . for large values of @xmath23 the potential goes as @xmath232 and at @xmath37 we have : @xmath233 that can be either positive , negative or zero . notice also that the potential vanishes provided : @xmath234 there are therefore three fundamentally different types of solutions . : for @xmath235 there are two positive-@xmath23 zeros of the potential ( fig.6a ) . the smallest zero is located between the inner horizon and @xmath37 , while the other zero is between the outer horizon and the static limit . therefore , this string solution never comes outside the static limit . on the other hand it never falls into @xmath37 . the mathematical solution oscillating between these two positive zeros of the potential may be interpreted as a string travelling between the different universes described by the maximal analytic extension of the spacetime ( the penrose diagram of the @xmath0 dimensional bh - ads spacetime is discussed in refs.@xcite ) . such type of circular string solutions also exist in other stringy black hole backgrounds @xcite . : for @xmath236 there is only one positive-@xmath23 zero of the potential , which is always located outside the static limit ( fig.6b ) . the potential is negative for @xmath37 , so there is no barrier preventing the string from collapsing into @xmath37 . by suitably fixing the initial conditions the string starts with its maximal size outside the static limit at @xmath237 . it then contracts through the ergosphere and the two horizons and eventually falls into @xmath37 . if @xmath238 it may however still be possible to continue this solution into another universe as in the case * ( i). : @xmath239 is the limiting case where the maximal string radius equals the static limit . the potential is exactly zero for @xmath37 so also in this case the string contracts through the two horizons and eventually falls into @xmath37 . the exact and complete mathematical solution can be obtained in terms of elementary or elliptic functions , the details can be found in ref.@xcite . in all cases we find only bounded string size solutions and no multi - string solutions . see also table ii . we will now compare the circular strings in the @xmath0 dimensional bh - ads spacetime and in the equatorial plane of ordinary @xmath1 dimensional black holes . in the most general case it is natural to compare the spacetime metric , eq . ( 4.1 ) , with the ordinary @xmath1 dimensional kerr anti de sitter spacetime with metric components : @xmath240 @xmath241 where we have introduced the notation : @xmath242 here the mass is represented by @xmath224 while @xmath199 is the specific angular momentum , and a positive @xmath243 corresponds to de sitter while a negative @xmath243 corresponds to anti de sitter spacetime . in the equatorial plane @xmath244 the metric , eq . ( 5.1 ) , is in the general form of eq . ( 2.1 ) so that we can use the analysis of section 2 . in the most general case the potential is given by : @xmath245 i.e. the potential covers seven powers in @xmath23 . the general solution will therefore involve higher genus elliptic functions . it is furthermore very complicated to deduce the physical properties of the circular strings from the shape of the potential ( the zeros etc . ) since the invariant string size defined in eq . ( 2.14 ) is non - trivially connected to @xmath246 @xmath247 we have exactly solved the string dynamics in a number of spacetimes of the form , eq . ( 5.1 ) @xcite . in the cases of minkowski space , anti de sitter space , schwarzschild and schwarzschild anti de sitter space , fig.7 . , there are only bounded string size solutions and no multi - string solutions . in schwarzschild de sitter space , on the other hand , the dynamics outside the schwarzschild horizon is similar to the dynamics in `` pure '' de sitter space , so we find the complicated spectrum of oscillating , expanding and contracting strings and the multi - string solutions , see table ii . to obtain more insight about the string propagation in all these curved spacetimes we solved the string equations of motion and constraints by considering perturbations around the exact string center of mass solution , refining the approach originally developed by de vega and snchez @xcite . in an arbitrary curved spacetime of dimension @xmath248 , the string equations of motion and constraints , in the conformal gauge , take the form : @xmath249 @xmath250 for @xmath251 and prime and dot represent derivative with respect to @xmath148 and @xmath252 respectively . consider first the equations of motion , eq . . a particular solution is provided by the string center of mass @xmath253 : @xmath254 then a perturbative series around this solution is developed : @xmath255 after insertion of eq . ( 6.4 ) in eq . ( 6.1 ) , the equations of motion are to be solved order by order in the expansion . to zeroth order we just get eq . ( 6.3 ) . to first order we find : @xmath256 the first three terms can be written in covariant form @xcite , c.f . the ordinary geodesic deviation equation : @xmath257 however , we can go one step further . for a massive string , corresponding to the string center of mass fulfilling ( in units where @xmath258 ) : @xmath259 there are @xmath260 physical polarizations of string perturbations around the geodesic @xmath253 . we therefore introduce @xmath260 normal vectors @xmath261 : @xmath262 and consider only first order perturbations in the form : @xmath263 where @xmath220 are the comoving perturbations , i.e. the perturbations as seen by an observer travelling with the center of mass of the string . the normal vectors are not uniquely defined by eqs . in fact , there is a gauge invariance originating from the freedom to make local rotations of the @xmath264-bein spanned by the normal vectors . for our purposes it is convenient to fix the gauge taking the normal vectors to be covariantly constant : @xmath265 this is achieved by choosing the basis @xmath266 obeying the conditions given by eqs . ( 6.8 ) at a given point , and defining it along the geodesic by means of parallel transport . another useful formula is the completeness relation that takes the form : @xmath267 using eqs . ( 6.7)-(6.10 ) in eq . ( 6.6 ) , we find after multiplication by @xmath268 the spacetime invariant formula @xcite : @xmath269 since the last term depends on @xmath148 only through @xmath270 it is convenient to make a fourier expansion : @xmath271 then eq . ( 6.12 ) finally reduces to : @xmath272 which constitutes a matrix schrdinger equation with @xmath67 playing the role of the spatial coordinate . notice that in the case of constant curvature spacetimes , @xmath273 the `` potential '' in eq . ( 6.14 ) is obtained directly from the normalization equations ( 6.7 ) and ( 6.8 ) without any calculations at all . for the second order perturbations the picture is a little more complicated . since they couple to the first order perturbations we consider the full set of perturbations @xmath274 @xcite : @xmath275 where the source @xmath276 is bilinear in the first order perturbations , and explicitly given by : @xmath277 after solving eqs . ( 6.14)-(6.15 ) for the first and second order perturbations , the constraints , eq . ( 6.2 ) , have to be imposed . in world - sheet light cone coordinates @xmath278 the constraints take the form : @xmath279 where @xmath280 . the world - sheet energy - momentum tensor @xmath281 is conserved , as can be easily verified using eq . ( 6.1 ) , and therefore can be written : @xmath282 at the classical level under consideration here , the constraints are then simply : @xmath283 up to second order in the expansion around the string center of mass we find : @xmath284 and the conditions @xmath285 eq . ( 6.19 ) , then give a formula for the string mass . in the case of ordinary @xmath286 de sitter space , the first order perturbations , eq . ( 6.14 ) , are determined by @xcite : @xmath287 for @xmath288 the frequencies become imaginary and classical instabilities develop . after solving the second order perturbation equations , eqs . ( 6.15)-(6.16 ) , the mass formula is found @xcite : @xmath289 after quantization one finds that real mass states can only be defined up to some maximal mass @xcite . this is reminiscent of the classical string instabilities in de sitter space . in ordinary anti de sitter space and in the @xmath0 black hole ads ( which is locally , but not globally , ads ) the first order perturbation equations and mass formula take the form of eqs . ( 6.21)-(6.22 ) , but with @xmath290 replaced by @xmath291 ( or @xmath292 depending on notation ) . it follows that in these cases there are no instabilities neither classically nor quantum mechanically . the perturbation series approach is perfectly well - defined in these cases . it is interesting to compare the @xmath0 black hole ads with ordinary @xmath286 black hole ads . in secs . 4,5 we found that the circular string motion is very similar in these backgrounds ( see table ii ) . this was actually somewhat surprising since the backgrounds are really very different : the ordinary black holes have a strong curvature singularity at @xmath293 the @xmath0 black hole ads has not . the circular strings are however very special configurations and that could be the reason why we did not really see any qualitative differences . for generic strings we would expect some differences , however , and that is indeed what we find when using the string perturbation series approach . in the ordinary black hole ads spacetime the first order perturbation equations become @xcite : @xmath294 @xmath295 for the transverse and longitudinal perturbations , respectively . for the transverse perturbations , eq . ( 6.23 ) , the bracket is always positive , thus the frequencies are real and no instabilities arise , not even for @xmath296 for the longitudinal perturbations , on the other hand , the bracket can be negative . in that case imaginary frequencies develop . the ( @xmath297)-instability sets in at : @xmath298 the higher modes develop instabilities for smaller @xmath299 i.e. closer to the singularity . similar results are obtained in the black string background @xcite , see table iii . we have studied the string propagation in de sitter and black hole backgrounds . the main part of the talk concerned the dynamics of circular strings in de sitter space , with special interest in the physical interpretation of the results of the mathematical analysis . we then compared with results obtained in various black hole backgrounds ( @xmath0 bh - ads , schwarzschild - anti de sitter , schwarzschild - de sitter ) . finally , more insight about the strings in curved spacetimes was obtained using the string perturbation series approach . the main results and conclusions are summarized in tables i , ii , iii . + i would like to thank h.j . de vega and n. snchez for a very fruitful collaboration on the material presented in this talk . 11 h.j . de vega and n. snchez , _ phys . rev . _ d42 * ( 1990 ) 3969 . de vega , m.r . medrano and n. snchez , _ nucl . phys . _ * b374 * ( 1992 ) 405 . de vega and n. snchez , _ phys . _ * b244 * ( 1990 ) 215 , _ phys . lett . _ * 65c * ( 1990 ) 1517 , _ ijmp _ * a7 * ( 1992 ) 3043 , _ nucl . * b317 * ( 1989 ) 706 and ibid 731 . + d. amati and c. klimcik , _ phys . _ * b210 * ( 1988 ) 92 . + m. costa and h.j . de vega , _ ann . _ * 211 * ( 1991 ) 223 and ibid 235 . + c. loust and n. snchez , _ phys . * d46 * ( 1992 ) 4520 . v.e . zakharov and a.v . mikhailov , _ jetp _ * 47 * ( 1979 ) 1017 . + h. eichenherr , in _ integrable quantum field theories _ , ed . j. hietarinta and c. montonen ( springer , berlin , 1982 ) . i. bars and k. sfetsos , _ mod . lett . _ * a7 * ( 1992 ) 1091 . de vega , j.r . mittelbrunn , m.r . medrano and n. snchez , `` the two - dimensional stringy black hole : a new approach and a pathology '' , _ par - lpthe _ * 93/14 * and `` the general solution of the 2-d sigma model stringy black hole and the complex sine - gordon equation '' , _ par - lpthe _ * 93/53. h.j . de vega and n. snchez , _ phys . _ b197 * ( 1987 ) 320 . mende , in _ string quantum gravity and the physics at the planck scale _ , ed . n. snchez ( world scientific , singapore , 1993 ) . a.l . larsen and v.p . frolov , _ nucl . _ * b414 * ( 1994 ) 129 . a.l . larsen and n. snchez , `` strings propagating in the 2 + 1 dimensional black hole anti de sitter spacetime '' , _ obs . de paris , demirm _ * 94013. the contributions by h.j . de vega and n. snchez in _ string quantum gravity and the physics at the planck scale _ , ed . n. snchez ( world scientific , singapore , 1993 ) . n. snchez and g. veneziano , _ nucl . phys . _ b333 * ( 1990 ) 253 . + m. gasperini , n. snchez and g. veneziano , _ nucl . phys . _ * b364 * ( 1991 ) 365 frolov , v.d . skarzhinsky , a.i . zelnikov and o. heinrich , _ phys . lett . _ * b224 * ( 1989 ) 255 . de vega , a.v . mikhailov and n. snchez , _ teor . _ * 94 * ( 1993 ) 232 . . de vega , a.l . larsen and n. snchez , `` infinitely many strings in de sitter spacetime : expanding and oscillating elliptic function solutions '' , _ obs . de paris , demirm _ * 93055. a.l . larsen , `` circular string - instabilities in curved spacetime '' , _ obs . de paris , demirm _ 93052 * , to appear in phys . rev . d. a. davidson , n.k . nielsen and y. verbin , _ nucl . _ * b412 * ( 1994 ) 391 . . de vega , a.l . larsen and n. snchez , `` semi - classical quantization of circular strings in de sitter and anti de sitter spacetime '' , in preparation . m. baados , c. teitelboim and j. zanelli , _ phys . * 69 * ( 1992 ) 1849 . gradshteyn and i.m . ryznik , _ table of integrals , series and products _ ( academic press inc , london , 1980 ) . m. abramowitz and i.a . stegun , _ handbook of mathematical functions _ ( dover publications inc , new york , 1970 ) . larsen , `` stable and unstable circular strings in inflationary universes '' , _ nordita _ * 94/14 p. m. baados , m. henneaux , c. teitelboim and j. zanelli , _ phys . rev . _ d48 * ( 1993 ) 1506 . horowitz and d.l . welch , _ phys . * 71 * ( 1993 ) 328 . . de vega and i.l . egusquiza , _ phys . * d49 * ( 1994 ) 763 .	string propagation is investigated in de sitter and black hole backgrounds using both exact and approximative methods . the circular string evolution in de sitter space is discussed in detail with respect to energy and pressure , mathematical solution and physical interpretation , multi - string solutions etc . we compare with the circular string evolution in the @xmath0 dimensional black hole anti de sitter spacetime and in the equatorial plane of ordinary @xmath1 dimensional stationary axially symmetric spacetime solutions of einstein general relativity . using an approximative string perturbation approach we consider also generic string evolution and propagation in all these curved spacetimes . + arne l. larsen +
as common constructions in the theory of data types and programming language semantics become better understood , there is a natural tendency toward generality . one desires to isolate common underlying principles , to unify related notions in a common framework , and to provide powerful abstract tools for reasoning and understanding . a good example of one successful such enterprise is the use of monads in functional and logic programming @xcite . monads provide a clean way to combine modules or extend functionality of programming languages or data structures with new features such as continuations , state , and concurrency @xcite . they have been applied to parsing and type checking @xcite , semantics of nondeterministic and probabilistic computation @xcite , and unification in logic programming @xcite . unfortunately , greater abstraction is often accompanied by reduced accessibility . many abstract constructions , although well motivated by applications , may at times be difficult to navigate when presented in more abstract form . in particular , reasoning about the basic properties of monads such as _ monad composition , the construction that underlies many of the applications above relies on the combinatorial manipulation of functors and natural transformations . verification often requires a complicated process of arrow chasing in large diagrams . specialized verification tasks such as the following example from @xcite are not uncommon . \(00 ) @xmath2 ; ( 01 ) [ right of=00 , node distance=35 mm ] @xmath3 ; ( 02 ) [ right of=01 , node distance=25 mm ] @xmath4 ; ( 10 ) [ below of=00 , node distance=1.3 cm ] @xmath5 ; ( 11 ) [ right of=10 , node distance=35 mm ] @xmath6 ; ( 12 ) [ right of=11 , node distance=25 mm ] @xmath7 ; ( 20 ) [ below of=10 , node distance=1.3 cm ] @xmath4 ; ( 21 ) [ right of=20 , node distance=35 mm ] @xmath7 ; ( 22 ) [ right of=21 , node distance=25 mm ] @xmath8 ; ( 00 ) edge node @xmath9 ( 01 ) edge node[swap ] @xmath10 ( 10 ) ; ( 01 ) edge node @xmath11 ( 02 ) edge node[swap ] @xmath12 ( 11 ) ; ( 02 ) edge node @xmath13 ( 12 ) ; ( 10 ) edge node[swap ] @xmath14 ( 11 ) edge node[swap ] @xmath15 ( 20 ) ; ( 11 ) edge node[swap ] @xmath16 ( 12 ) edge node[swap ] @xmath17 ( 21 ) ; ( 12 ) edge node @xmath18 ( 22 ) ; ( 20 ) edge node[swap ] @xmath13 ( 21 ) ; ( 21 ) edge node[swap ] @xmath18 ( 22 ) ; there are several general accounts of monad composition in the literature @xcite . in particular , eklund et al . @xcite make note of the difficulty of monad composition proofs and present a graphical technique aimed at simplifying the process . their technique provides a pictorial representation of various constructions . in this note we take a different approach involving string rewriting . we show that a given tuple @xmath0 is a monad if and only if @xmath1 is a terminal object in a certain category whose objects are strings and whose morphisms are rewrite rules , and the fact that @xmath1 is terminal can be established by proving confluence of the string rewriting system . we illustrate the technique on the monad composition problem studied by eklund et al . we also give a characterization of adjoint functors in terms of rewrite categories . let @xmath19 be categories . recall that a _ natural transformation _ @xmath20 between functors @xmath21 is a collection of morphisms @xmath22 of @xmath23 , one for each object @xmath24 of @xmath25 , such that for any morphism @xmath26 of @xmath25 , the following diagram commutes : @xmath27 \scriptsize \node ( 00 ) { $ fa$ } ; \node ( 01 ) [ right of=00 ] { $ fb$ } ; \node ( 10 ) [ below of=00 , node distance=12 mm ] { $ ga$ } ; \node ( 11 ) [ right of=10 ] { $ gb$ } ; \path ( 00 ) edge node { $ fh$ } ( 01 ) edge node[swap ] { $ { \varphi}_a$ } ( 10 ) ; \path ( 10 ) edge node[swap ] { $ gh$ } ( 11 ) ; \path ( 01 ) edge node { $ { \varphi}_b$ } ( 11 ) ; \end{tikzpicture } \end{array } \label{dia : nattrans}\ ] ] categories , functors , and natural transformations form a 2-category @xmath28 in which the 0-cells ( objects ) are the categories , the 1-cells ( morphisms ) are the functors , and the 2-cells ( morphisms of morphisms ) are the natural transformations . in a general 2-category , composition of 1-cells and composition of 2-cells are called _ horizontal _ and _ vertical composition _ , respectively , and are often denoted @xmath29 and @xmath30 , respectively ; but we will usually write @xmath31 for @xmath30 and omit the @xmath29 altogether . horizontal composition also acts on 2-cells and satisfies the property : if @xmath32 and @xmath33 are 1-cells and @xmath34 , @xmath35 , @xmath36 , and @xmath37 are 2-cells , then @xmath38 a special case of the action of horizontal composition on 2-cells is the following . if @xmath39 , @xmath21 , and @xmath40 are 1-cells and @xmath20 is a 2-cell , there is a 2-cell @xmath41 obtained from the horizontal composition of @xmath42 with the identities @xmath43 on the left and @xmath44 on the right . it is helpful to think of this operation as a kind of scalar multiplication . it satisfies the following properties : @xmath45 thus we can write @xmath46 and @xmath47 without ambiguity . in the 2-category @xmath28 , if @xmath48 and @xmath1 are functors and @xmath42 is a natural transformation with components @xmath49 , then @xmath50 is the natural transformation with components @xmath51 . it follows from that the commutativity of diagrams is preserved under scalar multiplication ; that is , if the left - hand diagram in commutes , then so does the right : @xmath27 \scriptsize \node ( 00 ) { $ f$ } ; \node ( 01 ) [ right of=00 ] { $ g$ } ; \node ( 10 ) [ below of=00 , node distance=12 mm ] { $ x$ } ; \node ( 11 ) [ right of=10 ] { $ y$ } ; \path ( 00 ) edge node { $ { \varphi}$ } ( 01 ) edge node[swap ] { $ \psi$ } ( 10 ) ; \path ( 10 ) edge node[swap ] { $ \sigma$ } ( 11 ) ; \path ( 01 ) edge node { $ \tau$ } ( 11 ) ; \node ( 00 ) [ right of=01 , node distance=25 mm ] { $ sft$ } ; \node ( 01 ) [ right of=00 ] { $ sgt$ } ; \node ( 10 ) [ below of=00 , node distance=12 mm ] { $ sxt$ } ; \node ( 11 ) [ right of=10 ] { $ syt$ } ; \path ( 00 ) edge node { $ s{\varphi}t$ } ( 01 ) edge node[swap ] { $ s\psi t$ } ( 10 ) ; \path ( 10 ) edge node[swap ] { $ s\sigma t$ } ( 11 ) ; \path ( 01 ) edge node { $ s\tau t$ } ( 11 ) ; \end{tikzpicture } \end{array } \label{dia : extend}\ ] ] from these observations , one can begin to see the motivation for viewing natural transformations as rewrite rules : a rewrite rule @xmath20 can be applied in the context of a string @xmath41 . for this to make sense , though , it had better be the case that the diamond property holds for non - overlapping redexes ; that is , it must be possible to apply two rewrite rules with non - overlapping redexes in either order . we will see later ( lemma [ lem : disjoint ] ) that this is indeed the case , and in fact holds in all 2-categories as a consequence of . a _ rewrite category _ is like a string rewrite system , except the semantics of the derivations are taken into account in the definitions of confluence and local confluence . formally , a _ rewrite category _ is defined in terms of a finitely presented 2-category @xmath52 , where @xmath53 , @xmath54 , and @xmath55 are finite sets of 0-cells , 1-cells , and 2-cells , respectively , and @xmath56 is a finite set of well - typed equations between 2-cell - valued expressions over @xmath54 and @xmath55 generating an equational theory on 2-cells . the 1-cells do not satisfy any equations except those imposed by the axioms of category theory , thus form the free category ( free typed monoid ) freely generated by @xmath54 . each 1-cell corresponds to a well - typed string @xmath57 , where _ well - typed _ means @xmath58 , @xmath59 . we denote by @xmath60 the set of all well - typed strings over @xmath54 . the elements of @xmath55 are called _ rewrite rules _ and generate the set of all 2-cells under horizontal ( @xmath29 ) and vertical ( @xmath30 ) composition . the rewrite category itself is a substructure of the category of 1- and 2-cells of this finitely presented 2-category . it is specified by a tuple @xmath61 , where @xmath62 and @xmath63 is any subset of @xmath60 closed under the action of @xmath64 . rewrite categories provide a syntax that can be interpreted in the 2-category @xmath28 of categories , functors , and natural transformations . a well - typed string @xmath57 is interpreted as a functor @xmath65 , where concatenation of strings is interpreted as composition of functors . the rewrite rules describe the action of natural transformations on these strings . the equations @xmath56 generating the equational theory of the 2-category typically represent an equational axiomatization of the construct under study . however , from the perspective of string rewriting , they serve a different purpose : they represent local confluence conditions that can be used to establish global confluence in the rewrite category , leading to the existence of terminal objects . [ ex : monad ] recall that a _ monad _ on a category @xmath25 is a triple @xmath0 , where @xmath1 is an endofunctor on @xmath25 and @xmath66 and @xmath67 are natural transformations such that @xmath68 and @xmath69 ; that is , the diagrams @xmath70 \scriptsize \node ( 00 ) { $ t^3 $ } ; \node ( 01 ) [ right of=00 ] { $ t^2 $ } ; \node ( 10 ) [ below of=00 , node distance=12 mm ] { $ t^2 $ } ; \node ( 11 ) [ right of=10 ] { $ t$ } ; \path ( 00 ) edge node { $ t\mu$ } ( 01 ) edge node[swap ] { $ \mu t$ } ( 10 ) ; \path ( 10 ) edge node[swap ] { $ \mu$ } ( 11 ) ; \path ( 01 ) edge node { $ \mu$ } ( 11 ) ; \node ( 00 ) [ right of=01 , node distance=25 mm ] { $ t$ } ; \node ( 01 ) [ right of=00 ] { $ t^2 $ } ; \node ( 10 ) [ below of=00 , node distance=12 mm ] { $ t^2 $ } ; \node ( 11 ) [ right of=10 ] { $ t$ } ; \path ( 00 ) edge node { $ t\eta$ } ( 01 ) edge node[swap ] { $ \eta t$ } ( 10 ) edge node { $ { \mathsf{id}}$ } ( 11 ) ; \path ( 10 ) edge node[swap ] { $ \mu$ } ( 11 ) ; \path ( 01 ) edge node { $ \mu$ } ( 11 ) ; \end{tikzpicture } \end{array } \label{dia : monad}\ ] ] commute . a typical example is the _ powerset monad _ @xmath71 on @xmath72 , where @xmath73 is the powerset of @xmath24 , @xmath74 , and @xmath75 . monads are characterized by a rewrite category @xmath76 , where @xmath77 is the free monoid on one generator @xmath1 with rewrite rules @xmath67 and @xmath66 . the underlying 2-category has one 0-cell . the equational theory on 2-cells is given by the defining equations for monads . the properties completely axiomatize the equational theory of monads , but more importantly from the from the point of view of rewriting , they specify local confluence properties in the case of overlapping redexes . for example , the string @xmath78 contains two overlapping redexes for the rule @xmath18 . the left - hand diagram of specifies that this configuration satisfies the diamond property . we will show later ( theorem [ thm : monad ] ) that @xmath1 is a terminal object in the rewrite category , and that this fact is equivalent to the axiomatization . [ ex : monads ] the composition of monads is characterized by a rewrite category @xmath79 where @xmath80 is the free monoid on two generators @xmath81 . as in example [ ex : monad ] , the underlying 2-category has one 0-cell . the equational theory is given by the axiomatization of two monads on endofunctors @xmath82 and @xmath1 , respectively , and a distributive law @xmath83 connecting them . we will give more details in [ sec : monads ] . again , the axioms can be viewed as local confluence properties for overlapping redexes in the rewrite category . [ ex : adjunction ] adjunctions are characterized by a rewrite category @xmath84 . in this example , the underlying 2-category has two 0-cells @xmath19 representing two categories , 1-cells @xmath85 and @xmath86 representing left and right adjoint functors , respectively , and 2-cells @xmath87 and @xmath88 representing the unit and counit of the adjunction , respectively . given a 1-cell @xmath89 and a rule @xmath90 of a rewrite category , the scalar multiple @xmath91 can be viewed as rewriting the indicated occurrence of @xmath92 to @xmath93 in the context of the string @xmath89 . this is called a _ reduction_. the _ redex _ of the reduction @xmath50 is the substring @xmath92 of @xmath89 . a sequence of reductions from @xmath93 to @xmath94 is called a _ derivation_. we write @xmath95 if @xmath96 is such a derivation . a string is in _ normal form _ if no rules apply . the rules @xmath55 of the rewrite category each have a type @xmath97 for various @xmath98 , where @xmath99 and @xmath100 . applying these rules to strings in @xmath60 , one can generate diagrams in the rewrite category . we say that the system @xmath55 * is _ confluent _ if for any two derivations @xmath101 and @xmath102 , there is a word @xmath103 and derivations @xmath104 and @xmath105 such that the resulting diagram commutes ; * is _ locally confluent _ if for any two single reductions @xmath106 and @xmath107 , there is a word @xmath103 and derivations @xmath104 and @xmath105 such that the resulting diagram commutes ; and * has the _ diamond property _ if for any two reductions @xmath106 and @xmath107 , there is a word @xmath103 and reductions @xmath108 and @xmath109 such that the resulting diagram commutes . this might be a pushout in the rewrite category , but not necessarily . local confluence does not imply confluence , but the diamond property does . see @xcite for a thorough treatment of these concepts . note , however , that our reductions have semantic content as well as syntactic . in our definitions of confluence , local confluence , and the diamond property , it is not enough that two derivations derive the same word ; the resulting diagrams must also commute in the rewrite category . we say that two derivations @xmath110 are _ equivalent _ if the composition of the 2-cells along the two paths are equal . one of the defining properties for monads , namely the left - hand diagram of , says that @xmath18 as a reduction rule can be applied to the string @xmath78 in two ways to obtain @xmath111 : one way as @xmath112 to the leftmost two occurrences of @xmath1 ( the left arrow of the diagram ) and the other as @xmath113 to the rightmost ( the top arrow ) , in both cases giving @xmath111 . by applying @xmath18 again to the two occurrences of @xmath111 , we obtain a commutative diamond . by , we can compose on the left and right with any strings in @xmath77 : @xmath114 \scriptsize \node ( 00 ) { $ t^{m+n+3}$ } ; \node ( 01 ) [ right of=00 ] { $ t^{m+n+2}$ } ; \node ( 10 ) [ below of=00 , node distance=12 mm ] { $ t^{m+n+2}$ } ; \node ( 11 ) [ right of=10 ] { $ t^{m+n+1}$ } ; \path ( 00 ) edge node { $ t^{m+1}\mu t^n$ } ( 01 ) edge node[swap ] { $ t^m\mu t^{n+1}$ } ( 10 ) ; \path ( 10 ) edge node[swap ] { $ t^m\mu t^n$ } ( 11 ) ; \path ( 01 ) edge node { $ t^m\mu t^n$ } ( 11 ) ; \end{tikzpicture } \end{array } \label{dia : muext}\ ] ] this says that any two reductions involving the rewrite rule @xmath18 with overlapping redexes , applied anywhere in a string of length at least three , can be completed to a commutative diamond . for nonoverlapping redexes , the diamond property always holds . this is a consequence of property of 2-categories . [ lem : disjoint ] any two applications of rewrite rules with disjoint redexes can be applied in either order , and the resulting diagram commutes . that is , any well - typed diamond of the form @xmath115 \scriptsize \node ( 00 ) { $ pqrst$ } ; \node ( 01 ) [ right of=00 ] { $ pqryt$ } ; \node ( 10 ) [ below of=00 , node distance=12 mm ] { $ pxrst$ } ; \node ( 11 ) [ right of=10 ] { $ pxryt$ } ; \path ( 00 ) edge node { $ pqr\tau t$ } ( 01 ) edge node[swap ] { $ p\sigma rst$ } ( 10 ) ; \path ( 10 ) edge node[swap ] { $ pxr\tau t$ } ( 11 ) ; \path ( 01 ) edge node { $ p\sigma ryt$ } ( 11 ) ; \end{tikzpicture } \end{array } \label{dia : disjoint}\ ] ] commutes , where @xmath116 are 1-cells and @xmath117 and @xmath118 are 2-cells . it follows from property of 2-categories that if @xmath20 and @xmath119 are 2-cells and the following diagram is well - typed , then it commutes . @xmath27 \scriptsize \node ( 00 ) { $ fm$ } ; \node ( 01 ) [ right of=00 ] { $ fn$ } ; \node ( 10 ) [ below of=00 , node distance=12 mm ] { $ gm$ } ; \node ( 11 ) [ right of=10 ] { $ gn$ } ; \path ( 00 ) edge node { $ f\psi$ } ( 01 ) edge node[swap ] { $ { \varphi}m$ } ( 10 ) ; \path ( 10 ) edge node[swap ] { $ g\psi$ } ( 11 ) ; \path ( 01 ) edge node { $ { \varphi}n$ } ( 11 ) ; \path ( 00 ) edge node { $ { \varphi}\psi$ } ( 11 ) ; \end{tikzpicture } \end{array } \label{dia : disjoint2}\ ] ] ( in fact , is equivalent to and . ) the diagram is a special case of with the following substitutions : @xmath120 for @xmath121 , @xmath122 for @xmath123 , @xmath124 for @xmath125 , @xmath126 for @xmath127 , @xmath128 for @xmath42 , and @xmath129 for @xmath130 . the property of natural transformations is a stronger form of . in this section we demonstrate the use rewrite categories in the verification of monad composition as presented by eklund et al . we also give a characterization of monads and adjunctions in terms of rewrite categories . the following lemma and its proof introduce our approach at a basic level . [ lem : basic ] let @xmath131 denotes the set of nonnull strings of @xmath1 s and @xmath18 the rewrite rule @xmath132 . the following are equivalent : 1 . @xmath18 satisfies the left - hand diagram of ; 2 . @xmath1 is a terminal object in the rewrite category @xmath133 . \(1 ) @xmath134 ( 2 ) : combining lemma [ lem : disjoint ] with the observation , we have that the rewrite system consisting of the single rule @xmath18 on strings @xmath135 for @xmath136 satisfies the diamond property and is therefore confluent . it follows that any diagram starting from @xmath135 , @xmath136 , and ending with the normal form @xmath1 commutes . moreover , there exists a reduction sequence from any such @xmath135 to @xmath1 . thus there is a unique morphism @xmath137 for @xmath136 , so @xmath1 is a terminal object . \(2 ) @xmath134 ( 1 ) : conversely , if @xmath1 is a terminal object , then the left - hand diagram of must commute , since there is a unique morphism @xmath138 . now we add the rewrite rule @xmath139 to the mix . this rule can be used to introduce a new occurrence of @xmath1 anywhere in the string . in the presence of @xmath139 , @xmath1 is no longer a normal form , and in fact normal forms in the strict sense of string rewriting no longer exist . nevertheless , @xmath1 is still a terminal object . moreover , @xmath140 can now be included , since there is an arrow @xmath141 . [ thm : monad ] consider a rewrite category on 1-cells @xmath77 and rules @xmath66 and @xmath67 . the following are equivalent : 1 . @xmath18 and @xmath139 satisfy ; 2 . @xmath1 is a terminal object in the rewrite category @xmath142 . \(1 ) @xmath134 ( 2 ) : suppose we have a rewrite system in which the rules can be classified as either _ bad rules _ or _ good rules _ ( in our application , a rule is _ bad _ if it increases the length of the string , e.g. @xmath139 ) . call a derivation _ good _ if it uses only good rules . suppose further that 1 . every pair of good reductions @xmath106 and @xmath107 can be completed to a diamond using only good reductions @xmath108 and @xmath109 ; and 2 . every pair of reductions @xmath106 and @xmath107 , good or bad , are confluent using only good derivations @xmath104 and @xmath105 . these conditions hold for @xmath142 under the assumption ( 1 ) . we have already argued condition 1 for @xmath18 in lemma [ lem : basic ] . the right - hand diagram in implies condition 2 by immediately inverting any application of @xmath139 whenever it is applied to a nonnull string . for example , consider applications of @xmath139 and @xmath18 to a substring @xmath111 , where @xmath139 is applied between the two occurrences of @xmath1 . the two redexes thus overlap . @xmath143 \scriptsize \node ( 00 ) { $ t^2 $ } ; \node ( 01 ) [ right of=00 ] { $ t$ } ; \node ( 10 ) [ below of=00 , node distance=12 mm ] { $ t^3 $ } ; \node ( 11 ) [ right of=10 ] { } ; \path ( 00 ) edge node { $ \mu$ } ( 01 ) edge node[swap ] { $ t\eta t$ } ( 10 ) ; \end{tikzpicture } \end{array}\end{aligned}\ ] ] the top arrow is good , but the left arrow is bad . however , using the right - hand diagram of , the diagram can be completed using only good arrows : @xmath143 \scriptsize \node ( 00 ) { $ t^2 $ } ; \node ( 01 ) [ right of=00 ] { $ t$ } ; \node ( 10 ) [ below of=00 , node distance=12 mm ] { $ t^3 $ } ; \node ( 11 ) [ right of=10 ] { $ t^2 $ } ; \node ( 12 ) [ right of=11 ] { $ t$ } ; \path ( 00 ) edge node { $ \mu$ } ( 01 ) edge node[swap ] { $ t\eta t$ } ( 10 ) edge node { $ { \mathsf{id}}$ } ( 11 ) ; \path ( 10 ) edge node[swap ] { $ t\mu$ } ( 11 ) ; \path ( 11 ) edge node[swap ] { $ \mu$ } ( 12 ) ; \path ( 01 ) edge node { $ { \mathsf{id}}$ } ( 12 ) ; \end{tikzpicture } \end{array}\end{aligned}\ ] ] now we argue that any system satisfying 1 and 2 is confluent . let @xmath101 and @xmath102 be any two derivations involving good or bad rules . starting from the apex @xmath92 , move down the two derivations , adding good diamonds to the diagram in the case 1 and good confluent derivations in case 2 . all new transitions added to the diagram are good . when done , there are no more exposed bad rules , and the diagram can be completed by filling in with good diamonds . thus any two derivations @xmath144 are confluent via good derivations , which must be of the form @xmath145 . but the only good derivation @xmath145 is the identity . it follows that any two derivations @xmath146 are equivalent ; in other words , @xmath1 is a terminal object . \(2 ) @xmath134 ( 1 ) : conversely , if @xmath1 is a terminal object of the rewrite category , then all diagrams starting with any @xmath135 and ending with @xmath1 must commute , in particular those of , the defining conditions for monads . [ cor : monad ] let @xmath147 be a 2-functor from the free 2-category on one 0-cell , one 1-cell @xmath1 , and two 2-cells @xmath148 to @xmath149 , the 2-category of categories , functors , and natural transformations . let @xmath56 be the equational theory on 2-cells induced by @xmath150 . the following are equivalent : * the equations are a logical consequence of @xmath56 . * @xmath1 is terminal in the rewrite category @xmath151 modulo @xmath56 . * the image @xmath152 is a monad . let @xmath71 and @xmath153 be monads on a category @xmath25 connected by a _ distributive law _ ( or _ swapper _ in the terminology of @xcite ) , which is a natural transformation @xmath83 satisfying the following properties : @xmath154 \scriptsize \node ( 00 ) { $ tp^2 $ } ; \node ( 01 ) [ right of=00 ] { $ ptp$ } ; \node ( 02 ) [ right of=01 ] { $ p^2t$ } ; \node ( 10 ) [ below of=00 , node distance=12 mm ] { $ tp$ } ; \node ( 12 ) [ below of=02 , node distance=12 mm ] { $ pt$ } ; \path ( 00 ) edge node { $ \theta p$ } ( 01 ) edge node[swap ] { $ t{\ensuremath{\mu^p}}$ } ( 10 ) ; \path ( 10 ) edge node { $ \theta$ } ( 12 ) ; \path ( 01 ) edge node { $ p\theta$ } ( 02 ) ; \path ( 02 ) edge node { $ { \ensuremath{\mu^p}}t$ } ( 12 ) ; \node ( 20 ) [ below of=10 , node distance=12 mm ] { $ t^2p$ } ; \node ( 21 ) [ right of=20 ] { $ tpt$ } ; \node ( 22 ) [ right of=21 ] { $ pt^2 $ } ; \path ( 20 ) edge node[swap ] { $ t\theta$ } ( 21 ) edge node { $ { \ensuremath{\mu^t}}p$ } ( 10 ) ; \path ( 21 ) edge node[swap ] { $ \theta t$ } ( 22 ) ; \path ( 22 ) edge node[swap ] { $ p{\ensuremath{\mu^t}}$ } ( 12 ) ; \node ( 00 ) [ right of=02 , node distance=25 mm , yshift=-6 mm ] { $ tp$ } ; \node ( 10 ) [ below of=00 , node distance=12 mm ] { $ t$ } ; \node ( 11 ) [ right of=10 ] { $ pt$ } ; \path ( 00 ) edge node { $ \theta$ } ( 11 ) ; \path ( 10 ) edge node { $ t{\ensuremath{\eta^p}}$ } ( 00 ) edge node[swap ] { $ { \ensuremath{\eta^p}}t$ } ( 11 ) ; \node ( 01 ) [ right of=00 ] { $ p$ } ; \path ( 01 ) edge node[swap ] { $ { \ensuremath{\eta^t}}p$ } ( 00 ) edge node { $ p{\ensuremath{\eta^t}}$ } ( 11 ) ; \end{tikzpicture } \end{array } \label{dia : distrib}\end{aligned}\ ] ] distributive laws are discussed in depth in @xcite . a typical application is the construction of the _ complex algebra _ of an algebra , whose elements are sets of elements of the original algebra . here @xmath82 would be the powerset monad and @xmath1 the term monad of some variety of algebras , and @xmath155 takes a term of sets @xmath156 and turns it into a set of terms @xmath157 . these constructions are discussed in @xcite . another example would be the combination of the additive and multiplicative monoid structures in semirings . here @xmath82 would be the finite powerset monad and @xmath1 the free monoid construction . in light of the theme of this note , it should be clear that the conditions in are nothing more than a way to establish local confluence in the case of overlapping redexes between @xmath155 and the monad operations . for example , the top rectangle of the left - hand diagram of handles overlapping redexes involving @xmath155 and @xmath158 . the two monads @xmath82 and @xmath1 can be combined as follows . define @xmath159 then @xmath160 is again a monad @xcite . we will verify this using theorem [ thm : monad ] . most of the work is contained in the following lemma . [ lem : monadcomp ] consider a rewrite category on 1-cells @xmath80 in a 2-category with a single 0-cell and rules @xmath161 , @xmath162 , @xmath163 , @xmath164 , and @xmath83 , such that @xmath165 and @xmath166 both satisfy the monad axioms with appropriate substitutions ) . the following statements are equivalent : 1 . @xmath167 , and @xmath155 satisfy the distributive laws . @xmath168 is a terminal object in the rewrite category @xmath169 . \(1 ) @xmath134 ( 2 ) : first we show that the rewrite system is confluent . recall that a rule is _ bad _ if it increases length , _ good _ otherwise . the good rules are @xmath158 , @xmath170 , and @xmath155 , and the bad rules are @xmath171 and @xmath172 . every pair of good reductions can be completed to a good confluent diagram . if the redexes do not overlap , this follows from lemma [ lem : disjoint ] . for redexes that overlap , all cases are covered by the left - hand diagram of and the left - hand diagram of . given any derivation @xmath173 , possibly containing bad reductions , produce a new derivation @xmath174 as follows : 1 . extend the derivation to derive @xmath168 . 2 . rearrange the resulting derivation @xmath175 to obtain an equivalent derivation @xmath176 in which all the bad rules are applied after all the good rules . step 1 can be accomplished by first introducing an occurrence of @xmath82 and/or @xmath1 using @xmath171 and @xmath172 if necessary , then moving all occurrences of @xmath82 to the left of all occurrences of @xmath1 using @xmath155 , then collapsing the @xmath82 s using @xmath158 and the @xmath1 s using @xmath170 . step 2 can be done without increasing the length of the derivation . for any bad reduction followed immediately by a good reduction , if the symbol introduced by the bad reduction is not part of the redex of the good reduction , then the two reductions can be switched by lemma [ lem : disjoint ] . otherwise , the symbol introduced by the bad reduction is part of the redex of the good reduction . there are only six ways this can happen : @xmath177 and symmetrically for @xmath82 . the first two are equivalent to the identity by the right - hand diagram of and can be deleted . the last is equivalent to @xmath178 by the right - hand diagram of . we can continue this process until there are no more bad reductions occurring before good reductions in the derivation . if @xmath92 contains at least one occurrence each of @xmath82 and @xmath1 , then this must also be true of any string derived from @xmath92 , since all rules preserve this property . but then @xmath174 can contain no bad reductions at all . if it did , then the last reduction would be bad . but it is impossible to derive @xmath168 from such a reduction , since it would have to come from a string of length one , and no such string can be derived from @xmath92 . by a similar argument , if @xmath179 , then @xmath174 contains exactly one bad reduction to introduce @xmath1 , and it occurs last in the derivation . this last reduction must be of the form @xmath180 . similarly , if @xmath181 , the last reduction of @xmath174 is of the form @xmath182 , and this is the only bad reduction in the derivation . if @xmath183 , a similar argument shows that @xmath174 must be either @xmath184 now suppose we are given derivations @xmath185 and @xmath186 , possibly using both good and bad rules . to show confluence of @xmath96 and @xmath187 , it suffices to show that @xmath176 and @xmath188 are confluent . as argued above , if @xmath92 contains at least one occurrence each of @xmath82 and @xmath1 , then @xmath174 and @xmath189 are good . thus we can complete them to a good commutative diagram by filling in with good commutative diamonds and . this process must terminate , since there is a fixed upper bound , quadratic in the length of @xmath92 , on the length of any good derivation from @xmath92 , since each good reduction strictly decreases the string in length or lexicographic order relative to @xmath190 ( * ? ? ? * lemma 2.7.2 ) . if @xmath179 , then as argued above , @xmath174 and @xmath189 are both of the form @xmath191 , where the prefixes @xmath192 contain no occurrence of @xmath1 . by theorem [ thm : monad ] , @xmath82 is terminal in @xmath193 , therefore the two prefixes @xmath192 are equivalent , and consequently so are @xmath174 and @xmath189 . the argument is similar for @xmath181 . when @xmath183 , we need only observe that the two derivations form a commutative diamond by lemma [ lem : disjoint ] . finally we show that @xmath168 is a terminal object . we have already argued that there is at least one derivation of @xmath168 from every @xmath194 , so there is at least one arrow @xmath195 in the rewrite category . to show that there is at most one , let @xmath196 be any two derivations . by confluence , we can complete to a pair of equivalent derivations @xmath197 . rearranging the final portions @xmath198 of these two derivations by step 2 above , we obtain good derivations . but the only good derivation @xmath198 is the identity , therefore @xmath96 and @xmath187 were already equivalent . \(2 ) @xmath134 ( 1 ) : if @xmath168 is a terminal object , then holds , as all maximal paths in all diagrams of lead to @xmath168 . [ thm : monadcomp ] let @xmath153 and @xmath71 be monads connected by a distributive law @xmath83 satisfying . let @xmath199 and @xmath200 be defined by . then @xmath160 is a monad . consider the rewrite category , @xmath158 , etc . to refer to the components of the rewrite category as well as their images in @xmath28 . ] @xmath169 . by theorem [ thm : monad ] , it suffices to show that @xmath168 is a terminal object in the rewrite category @xmath201 . by lemma [ lem : monadcomp ] , we know that it is terminal in the category @xmath169 , thus for any @xmath202 , there is exactly one arrow @xmath203 in that category . it follows that there is at most one arrow @xmath203 in the subcategory @xmath201 . but there is also at least one , since we can derive @xmath168 from @xmath204 by the rule @xmath200 and from @xmath205 for @xmath136 by @xmath206 applications of the rule @xmath199 . as a final application , we give a characterization of adjoint functors in terms of rewrite categories . recall that a functor @xmath85 is a left adjoint of another functor @xmath86 if there are natural transformations @xmath87 and @xmath88 , the _ unit _ and _ counit _ of the adjunction , respectively , satisfying @xmath143 \scriptsize \node ( 00 ) { $ g$ } ; \node ( 10 ) [ below of=00 , node distance=12 mm ] { $ gfg$ } ; \node ( 11 ) [ right of=10 ] { $ g$ } ; \path ( 00 ) edge node { $ { \mathsf{id}}$ } ( 11 ) edge node[swap ] { $ \eta g$ } ( 10 ) ; \path ( 10 ) edge node[swap ] { $ g{\varepsilon}$ } ( 11 ) ; \node ( 00 ) [ right of=00 , node distance=25 mm ] { $ f$ } ; \node ( 01 ) [ right of=00 ] { $ fgf$ } ; \node ( 11 ) [ below of=01 , node distance=12 mm ] { $ f$ } ; \path ( 00 ) edge node[swap ] { $ { \mathsf{id}}$ } ( 11 ) edge node { $ f\eta$ } ( 01 ) ; \path ( 01 ) edge node { $ { \varepsilon}f$ } ( 11 ) ; \end{tikzpicture } \end{array } \label{dia : adj}\end{aligned}\ ] ] [ thm : adj ] consider a 2-category on two 0-cells @xmath25 , @xmath23 generated by 1-cells @xmath85 and @xmath86 and 2-cells @xmath87 and @xmath88 . the following statements are equivalent : 1 . @xmath139 and @xmath207 satisfy . 2 . @xmath121 is a terminal object in the rewrite category @xmath208 and @xmath125 is a terminal object in the rewrite category @xmath209 . \(1 ) @xmath134 ( 2 ) : by arguments similar to the proof of theorem [ thm : monad ] and lemma [ lem : monadcomp ] , in any derivation @xmath173 , all applications of @xmath139 ( the bad rule ) can be moved after all applications of @xmath207 ( the good rule ) , by either lemma [ lem : disjoint ] in the case of disjoint redexes or in the case of overlapping redexes . the resulting equivalent derivation @xmath210 is no longer than @xmath96 . now if @xmath174 is of the form @xmath211 , there can be no application of @xmath139 , because the last @xmath139 would produce a string of length at least 2 . but then all redexes are redexes of @xmath207 , therefore are disjoint and can be done in any order by lemma [ lem : disjoint ] . thus all derivations of the form @xmath211 are equivalent . moreover , there exists a derivation @xmath211 consisting of @xmath206 applications of @xmath207 . thus @xmath121 is terminal in @xmath208 . a symmetric argument shows that @xmath125 is terminal in @xmath209 . \(2 ) @xmath134 ( 1 ) : in , there is a unique morphism to @xmath125 in the left - hand diagram and a unique morphism to @xmath121 in the right - hand diagram , thus the diagrams commute . [ cor : adj ] let @xmath147 be a 2-functor from the free 2-category on two 0-cells @xmath19 , two 1-cells @xmath85 , @xmath86 , and two 2-cells @xmath87 , @xmath88 to @xmath149 . let @xmath56 be the equational theory on 2-cells induced by @xmath150 . the following are equivalent : * the equations are a logical consequence of @xmath56 . * @xmath121 is terminal in the rewrite category @xmath208 modulo @xmath56 and @xmath125 is terminal in the rewrite category @xmath209 modulo @xmath56 . * @xmath212 and @xmath213 are adjoint functors with @xmath212 the left adjoint and @xmath213 the right adjoint with unit @xmath214 and counit @xmath215 . thanks to riccardo pucella for valuable comments . peter jipsen . a note on complex algebras of semigroups . in r. berghammer , b. mller , and g. struth , editors , _ relational and kleene - algebraic methods in computer science : proc . 7th int . sem . relational methods in computer science and 2nd int . workshop applications of kleene algebra _ , volume 3051 of _ lecture notes in computer science _ , pages 171177 . springer - verlag , may 2003 . prakash panangaden . the category of markov kernels . in christel baier , michael huth , marta kwiatkowska , and mark ryan , editors , _ electronic notes in theoretical computer science _ , volume 22 elsevier , 2000 . norman ramsey and avi pfeffer . stochastic lambda calculus and monads of probability distributions . in _ proceedings of the 29th acm sigplan - sigact symposium on principles of programming languages _ , pages 154165 . acm press , 2002 . philip wadler . monads for functional programming . in johan jeuring and erik meijer , editors , _ advanced functional programming : 1st int . school on advanced functional programming techniques _ , volume 925 of _ lecture notes in computer science _ , pages 2452 . springer - verlag , 1995 .	eklund et al . ( 2002 ) present a graphical technique aimed at simplifying the verification of various category - theoretic constructions , notably the composition of monads . in this note we take a different approach involving string rewriting . we show that a given tuple @xmath0 is a monad if and only if @xmath1 is a terminal object in a certain category of strings and rewrite rules , and that this fact can be established by proving confluence of the rewrite system . we illustrate the technique on the monad composition problem . we also give a characterization of adjunctions in terms of rewrite categories . _ in honor of ji admek on the occasion of his seventieth birthday _
the galaxy distribution in large spectroscopic surveys , such as sloan digital sky survey ( sdss ; * ? ? ? * ) , have revealed a complex hierarchical network of structure , called the cosmic web , composed of clusters , filaments , sheets , and voids ( e.g. * ? ? ? * ) . in the current paradigm of galaxy formation , these structures arise from the linear growth of gaussian density fluctuations in a nearly homogeneous early universe . dark matter tends to flow out of the voids , accretes onto the sheets , collapses to the filaments , and finally accumulates onto the clusters at the intersections of the filaments . due to the accretion history of the universe , the angular momentum of dark matter halos and galaxies are generated and affected by their large - scale environment . according to the tidal torque theory ( ttt ) , protogalaxies acquire their angular momentum by the asymmetric interaction between their inertia tensor and the local tidal field @xcite . the galaxy and halo spin directions are expected to be correlated with the principal axes of the local tidal tensor . over the past decades , both numerical simulations and real redshift surveys have revealed such correlation between halos / galaxies and their local large - scale structure @xcite . using numerical simulations , @xcite found that halo spin vectors are preferentially aligned with the intermediate axes of the tidal field , which is consistent with the theoretical prediction using the linear tidal torque theory . similar trends are also claimed from observational data . based on the real - space tidal field reconstructed from the 2mass redshift survey , @xcite claimed that there is a clear intrinsic alignment between the galaxy spins and the intermediate principal axes of the tidal tensor . for halos in filaments from n - body simulations , recent studies reported a mass - dependence alignment signal between halo spins and filaments . the spins of low - mass halos tend to be parallel to filaments , while high - mass halos have orthogonal alignment @xcite . for halos in sheets , the spin vectors tend to lie in the plane of the sheets , independent of halo mass @xcite . observationally , the picture is not as clear because of limited statistics and because of the difficulty of properly defining cosmic filaments from observational data . several studies have endeavored to solve this problem but give partly contradictory results @xcite . for instance , while @xcite found alignment signals between the spin vectors of galaxies and the surrounding large - scale structures ( e.g. filament and sheet ) , @xcite found no statistical evidence for departure from random orientations . using 2dfgrs and sdss , @xcite found that spiral galaxies located on the shells of cosmic voids have their spins aligned preferentially in the void surface , whereas , @xcite , who used the sdss dr7 and galaxy zoo samples , found that the spins of spiral galaxies around voids tend to be perpendicular to the void surface . based on sdss dr7 , a recent study by @xcite showed that there is a different alignment between spiral and elliptical galaxies , arguing that the spin of spiral galaxies tends to lie parallel to their host filaments , whereas the spin of elliptical galaxies are aligned preferentially perpendicular to their filaments . in short , observationally the spin alignment with the surrounding large - scale structure is still somewhat contradictory and inconclusive . in our previous paper @xcite , we examined the alignment between the major axes of galaxies and their surrounding large - scale structure . we found that red galaxies are very significantly aligned with their filaments , while blue galaxies are only marginally aligned with respect to their filaments . in this paper , we investigate the spin vectors of spiral galaxies with respect to the surrounding large - scale environments . here again we use the galaxy group catalog constructed from the sdss dr7 @xcite . the cosmic tidal field we used is directly reconstructed from the group catalog using the method described in @xcite . based on the signs of the eigenvalues of this tidal field , we classify the environments of groups into four categories : clusters , filaments , sheets , and voids . the morphologies of galaxies used here are those obtained from the galaxy zoo 2 catalog ( gz2 ; * ? ? ? we measure the spin vectors of spiral galaxies with respect to the principal axes of the local tidal tensor and study how these alignments depend on galaxy properties . the paper is organized as follows . in section [ sec_data ] , we present the observational data and the methodology used to measure the various alignment signals . the alignments of spiral galaxies with respect to cosmic web are present in section [ sec_result ] . we compare the observational results with dark matter halos from n - body simulations in section [ sec_sim ] . finally we summarize our results in section [ sec_summary ] . in what follows we describe the observational data we adopt to characterize the cosmic web and measure the alignment between the spin axes of galaxies relative to the large - scale structures . the galaxy sample we use here is the new york university value - added galaxy catalog ( nyu - vagc ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , which is based on the multi - band imaging and spectroscopic survey sdss dr7 @xcite . as the completion of the survey phase sdss - ii , dr7 features an imaging survey in five bands over a continuous @xmath0 region in the northern galactic cap . the continuity over this large area is essential and critical to reconstruct the real - space tidal field from the sdss survey . from the nyu - vagc , we select @xmath1 galaxies in the main galaxy sample with an extinction - corrected apparent magnitude brighter than @xmath2 , with redshifts in the range @xmath3 and with redshift completeness @xmath4 . a small subset galaxies have redshifts adopted from the korea institute from advanced study ( kias ) value - added galaxy catalog @xcite . using the adaptive halo - based group finder developed by @xcite , we construct a catalog of @xmath5 galaxy groups ( about @xmath6 groups with single galaxies ) . for each group , we assign a halo mass according to the ranking of its characteristic group luminosity , defined as the total luminosity of all group members with @xmath7 , where @xmath8 is the @xmath9-band absolute magnitude , @xmath10-corrected and evolution corrected to @xmath11 , using the method described by @xcite . in order to ensure sample completeness , in what follows we focus our analysis on the volume covering the redshift range @xmath12 , and within the largest continuous region in the northern galactic cap . using the complete galaxy groups with mass @xmath13 , @xcite constructed the mass , tidal and velocity ( sdss - mtv ) fields in this volume . we use the method developed by @xcite to correct for the peculiar motion of groups as well as the finger - of - god effect . we have tested that our alignment signals are insensitive to the boundary effect . the smoothing length scale we used to reconstruct the tidal field is @xmath14 , as this value provides the best agreement with the visual classification of the large - scale structure @xcite . the smoothing scale @xmath15 is related to the mass @xmath16 contained in the gaussian filter at mean density @xmath17 via @xmath18 , thus a smoothing scale @xmath19 corresponds to the mass @xmath20 . following @xcite , we calculate the tidal tensor using @xmath21 where @xmath22 and @xmath23 are indices with values of @xmath24 , @xmath25 , or @xmath26 , and @xmath27 is the peculiar gravitational potential , which can be calculated from the distribution of dark matter halos with mass @xmath28 above some threshold value @xmath29 through the poisson equation . at the position of each group , the tidal tensor is diagonalized to obtain the eigenvalues @xmath30 and @xmath31 ( @xmath32 ) , as well as the corresponding eigenvectors @xmath33 , @xmath34 , and @xmath35 . according to the zeldovich theory @xcite , the sign of the eigenvalues of local tidal tensor can be used to classify the group s environment into one of four categories @xcite : * : a point where all three eigenvalues are positive ; * : a point where @xmath36 has one negative and two positive eigenvalues ; * : a point where @xmath36 has two negative and one positive eigenvalues ; * : a point where all three eigenvalues are negative . the direction of a filament at the location of a group can be identified with the eigenvector corresponding to the single negative eigenvalue . likewise , the normal vector of a sheet for a group is the eigenvector corresponding to the single positive eigenvalue of the tidal tensor . central galaxies ; the dashed line shows the distribution for @xmath37 satellite galaxies.,scaledwidth=48.0% ] we adopt the morphological classifications of galaxies from the galaxy zoo @xmath25 ( gz2 ) catalog @xcite , which provides robust classifications into elliptical and spiral for @xmath38 galaxies . in the following analysis , we only use spiral galaxies with @xmath39 strings beginning with s. this results in a sample of @xmath40 spirals in our survey volume . based on group catalog , the sample contains @xmath41 centrals and @xmath42 satellite galaxies , among which @xmath43 centrals and @xmath37 satellites have reliable halo masses . to compare with the results from n - body simulations , we focus on the @xmath43 central galaxies with reliable halo masses . based on local tidal fields described in the last subsection , @xmath44 galaxies ( @xmath45 ) are located in environments , @xmath46 ( @xmath47 ) in , @xmath48 ( @xmath49 ) in , and @xmath50 ( @xmath51 ) in . in order to obtain a more reliable detection of the spin vectors @xmath52 of galaxies , in this paper we only use a sample of spiral galaxies . if a spiral galaxy were a thin circular disk , the inclination angle @xmath53 , between the plane of the disk and the line of sight , can be determined from the projected minor - to - major axis ratio @xmath54 . many studies used a simple model for the finite disk thickness to relate the apparent axis ratio @xmath54 and an intrinsic flatness parameter @xmath55 to obtain the inclination angles , is defined as the angle between the line of sight and the normal to the galaxy disk , so that @xmath56 . ] @xmath53 , of galaxies @xcite . here we follow the same practice and obtain @xmath53 through @xmath57 for @xmath58 and set @xmath59 for @xmath60 . according to @xcite , the intrinsic flatness @xmath55 depends on the morphological type of galaxies . however , we will use an average value of @xmath61 @xcite , since the variation of the flatness @xmath55 does not affect the calculation of the inclination angle @xmath53 significantly . we have repeated our entire calculation of the alignment assuming @xmath62 , and found that the results are very similar to those using @xmath63 . figure [ fig : ia ] shows the probability distribution of the inclination angles of galaxies . we find that the distribution differs significantly from a uniform distribution . this is mostly likely due to the fact that spiral galaxies have bulges which distort the projected axis ratio . if the @xmath54 axis ratio is measured around an isophote where the bulge contributes significantly , then the result can be affected . this effect becomes more and more significant for systems that are close to edge - on . because of this , our following analysis will focus on galaxies with @xmath64 , which results in a sample of @xmath65 centrals and @xmath66 satellites . from a two - dimensional image of a galaxy , projected on the sky , we can not decide which side of the galaxy is closer to us , which means that in the determination of @xmath53 using equation ( [ incang ] ) , a single value of @xmath54 corresponds to two values of the inclination angle : @xmath67 . to avoid this problem , many studies constrain their analysis to edge - on ( @xmath68 ) and face - on ( @xmath69 ) galaxies , which usually results in small samples of about a few hundreds of galaxies ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) . in this paper , we use all the spirals selected from gz2 catalog . throughout we apply the positive sign of @xmath53 for all the spirals . we have tested and found no difference in the alignment signal if using the negative sign for @xmath53 . however , the sign ambiguity is expected to reduce the strength of the alignment signal , as we quantify in the following section . following @xcite , we compute the components of the spin vector @xmath52 as @xmath70 where @xmath71 and @xmath72 are the right ascension and declination of a galaxy , respectively , @xmath53 is the inclination angle calculated from equation ( [ incang ] ) . the position angle @xmath27 increases counterclockwisely from north to east in the plane of the sky , and is specified by the @xmath73 magnitudes per square arc - second isophote in the @xmath9-band . to estimate the correlation between spin vectors of galaxies and the large - scale structure , we compute the cosine of the angle , @xmath74 , defined as the angle between the spin axis of the galaxy , @xmath52 , and the three principal axes @xmath33 , @xmath34 , and @xmath35 of the local tidal tensor . to quantify the alignments and their significance , we generate @xmath75 random samples in which the orientations of the principal axes of the tidal tensors are kept fixed , but the spin axes of the galaxies are randomized . the alignment strength is then expressed by the normalized pair count ( e.g. yang et al . 2006 ) : @xmath76 where @xmath77 is the number of pairs for each bin in @xmath78 , and @xmath79 is the average number of such pairs obtained from the @xmath75 random samples . we use @xmath80 , where @xmath81 is the s.d . of @xmath82 obtained from the 100 random samples , to assess the significance of alignment in each bin . since the significance is quantified with respect to the null hypothesis , we plot the error bars on top of the @xmath83 line . in addition to the normalized pair count , we also calculate the mean angle @xmath84 and @xmath85 , which is the standard deviation of @xmath86 for the @xmath75 random samples . in this paper , @xmath78 is restricted to the range @xmath87 $ ] . in the absence of any alignment , @xmath88 and @xmath89 . @xmath90 implies that the galaxy spin is parallel to the vector in question , while @xmath91 indicates it is perpendicular to it . but for central galaxies in halos ( groups ) of different masses . here results are shown only for galaxies with @xmath64 . , scaledwidth=50.0% ] but for galaxies in clusters ( first row ) , filaments ( second row ) , sheets ( third row ) , and voids ( fourth row ) . here results are shown only for galaxies with @xmath64 . , scaledwidth=50.0% ] figure [ fig : tf ] shows the normalized pair count @xmath92 of the cosine of the angle between the galaxy spin vector and the eigenvectors @xmath33 , @xmath34 , and @xmath35 of the local tidal tensor . in the left panel , the calculation of the alignment is based on all @xmath43 centrals and @xmath93 satellite spiral galaxies selected from gz2 . in the right panel , we repeat the measurements for galaxies with @xmath64 . as expected , galaxies with @xmath64 have stronger alignment signals because of the reduced uncertainty in the values of @xmath53 . the average cosines and their errors are displayed for reference . figure [ fig : tf ] shows that the galaxy spin vector has a weak tendency to be aligned with the intermediate axis @xmath34 , of the local tidal tensor , which is consistent with the theoretical prediction using the linear tidal torque theory @xcite . similar trends have been also found in simulation data . based on the tidal field traced by halos from n - body simulation , @xcite found that the spin vector of halos with masses larger than @xmath94 shows significant alignment with the intermediate axis of the tidal field . using a cosmological hydrodynamic simulation of galaxy formation in a cosmic filament , @xcite found that the spin vector has a weak alignment with the intermediate axis of the tidal field for their medium mass sample at @xmath95 ( see their figure @xmath96 ) . we can also see that the galaxy spin tends to be perpendicular to the minor axis @xmath35 , which is consistent with the result shown in the right panel of figure @xmath25 in @xcite . the alignment with the major axis @xmath33 is almost random . it is interesting to see that the alignment signal for satellites is as strong as that for centrals . this may suggest that satellites can preserve their alignment with the large scale tidal field after they have fallen into their host halos . since we have the halo mass for each central galaxy , we can investigate how the spin - tidal field alignment signal depends on the halo mass of groups . to that extent we split the central galaxies into four sub - samples of halo masses . figure [ fig : tf_mass ] shows the alignment signals as a function of halo mass . as one can see , the alignment signals in middle panels appear to be slightly stronger for centrals in more massive halos . it s also interesting to see that there is a transition of the alignment signals in the lower left panel , although the error bars are very large due to the small number galaxies with halo mass larger than @xmath97 . figure [ fig : tf1 ] shows the alignment between the three eigenvectors of the local tidal tensor and the spin vector of central galaxies which are located in cluster ( red ) , filament ( orange ) , sheet ( green ) , and void ( blue ) environments . here we see that for spiral galaxies in environments ( first row ) , the spin vectors have a significant alignment with the intermediate axis of the local tidal tensor . the upper right panel indicates that the spin vectors of galaxies in environments have a strong anti - alignment with the minor axis of the local tidal tensor . for galaxies in and environments , the alignments between spin vectors and the intermediate eigenvectors are rather weak . due to the small number of galaxies in environments , we can not obtain any useful information for these galaxies . according to the definition of , the right panel in the second row in figure [ fig : tf1 ] can be considered as the normalized pair count for the cosine of the angle between the spin vector and the direction of the filament in which the galaxy resides . there is a weak tendency that galaxies tend to have their spins preferentially perpendicular to their filaments , consistent with the prediction from the tidal torque theory . however , some of previous investigations found opposite alignments . using @xmath98 edge - on galaxies ( @xmath99 ) in filaments detected from sdss dr5 , @xcite found that only @xmath100 galaxies have their spin axes aligned perpendicular to their host filaments . this indicates that the spin of the edge - on galaxies tends to be parallel to their hosted filaments . however , as we have tested , if one only use edge - on or face - on galaxies , one needs to have a fairly large sample of filament directions to get statistically meaningful results . otherwise the signals may not be reliable . for example , if we have only one filament which lies along the line of sight direction , the edge - on or face - on galaxies will all have perfect perpendicular or alignment signals by selection . based on a galaxy sample constructed from sdss dr8 , @xcite found that the spin axes of spiral galaxies tend to align with the host filaments , which is opposite to our alignment signal . this discrepancy may be due to the different methods used to identify filamentary structures . in their investigation , filaments are traced by thin cylinders placed on the galaxy number density field , while the filaments in our samples are obtained from the continuous potential field constructed with the use of the method of @xcite . for spiral galaxies in environments , the left panel in the third row in figure [ fig : tf1 ] may be considered as the normalized pair count for the cosine of the angle between the spin and the normal of the sheet in which the galaxy resides . the results indicate that the spin vectors of galaxies have a very weak tendency to align along the normal of the sheet . using a sample of disk galaxies lying in the surfaces of voids obtained from sdss dr7 , @xcite found a significant tendency of galaxies around very large voids to have their spins aligned with the radial direction of the voids . this trend is consistent with our results . using @xmath101 edge - on and @xmath102 face - on galaxies from the 2dfgrs and sdss dr3 , @xcite found that spiral galaxies located on the walls of the largest voids have spins lying preferentially on the void surface . the disagreement may again be due to the small number of galaxies @xcite used . ( left ) , @xmath34 ( middle ) , and @xmath35 ( right ) of the tidal tensor field for fof halos . the dashed and solid line correspond to the original and control halo catalog , respectively . in the control sample , the spin vectors of @xmath103 halos have been changed ( see text for a detailed description ) . , scaledwidth=50.0% ] , but for so halos with @xmath104 . , scaledwidth=50.0% ] , but here spin vectors are estimated for inner halos within which the over density is @xmath105 . , scaledwidth=50.0% ] to facilitate the interpretation of our results , we adopt two different n - body simulations using the massively parallel treesph gadget-2 code @xcite . the simulations evolved @xmath106 dark matter particles in @xmath107 and @xmath108 periodic boxes . the ` glass - like ' initial conditions @xcite were generate using the zeldovich approximation at redshift @xmath109 . the simulations adopted the cosmological parameters with @xmath110 , @xmath111 , @xmath112 , @xmath113 , @xmath114 and @xmath115 . for the @xmath107 box simulation , the mass resolution and softening lengths are , respectively , @xmath116 and @xmath117 . this small volume simulation can give better statistics for halos with mass less than @xmath94 . the @xmath108 box simulation ( mass resolution of @xmath118 ) is used for computing the alignment signals for halos with mass larger than @xmath94 . dark matter halos are identified at redshift @xmath95 using the standard fof algorithm @xcite with a linking length of @xmath119 times the mean inter - particle separation . to obtain a reliable measurement , only halos with more than @xmath120 dark matter particles are included for further analysis , resulting in a catalog including @xmath121 ( @xmath122 ) halos with mass larger than @xmath123 ( @xmath124 ) for @xmath107 ( @xmath108 ) simulation . as comparison , we also use the so ( spherical over - density ) algorithm to identify halos . beginning with the most bound particles from the fof halos , a sphere is grown around the particle until the mean interior density is equal to the input value @xmath125 , where @xmath125 is the over - density within a sphere with respect to the critical density @xmath126 of the universe . here , we adopt @xmath104 and @xmath105 , representing the entire halo and the inner region , respectively . to obtain a spin vector of a halo containing @xmath127 particles , we compute the total angular momentum @xmath128 using @xmath129 where @xmath130 is the mass of a dark matter particle , @xmath131 denotes the position vector of particle @xmath22 relative to the center of mass , and @xmath132 is the velocity relative to the bulk velocity of the halo . as mentioned before , the spin vectors of galaxies in observation are not uniquely defined because we do not know which side of the galaxy is closer to us . in order to check this projection effect , we construct a control halo catalog in which we have changed the spin vectors of @xmath103 halos randomly selected from the original halo catalog . the changed spin unit vector @xmath133 is computed by @xmath134 where @xmath135 , @xmath136 is the original spin unit vector of a halo , and @xmath137 is the position unit vector of the halo , where we place a virtual observer at the corner of the simulation box . note that the original and changed spin vector of a halo are symmetric with respect to the line - of - sight direction @xmath138 . in the simulation , we have classified the cosmic web environment of a dark matter halo using the density field hessian matrix method @xcite , similar to the method used above based on the tidal tensor . here , in order to compare with the alignment signals detected in the sdss data , we compute the probability distribution function @xmath139 , where @xmath140 is the angle between the halo spin and the eigenvectors of the tidal tensor . figure [ fig : fof ] shows the normalized pair count @xmath141 of the cosine of the angle between the halo spin and the eigenvectors of the tidal tensor . the solid line corresponds to the control sample , in which the spin vectors of @xmath103 halos have been changed to model the projection effect in observation . the dashed line denotes the alignment signal from the original halo catalog in the n - body simulation , which is consistent with the results described in @xcite . from figure [ fig : fof ] , we find that the alignment strength for control sample decreases about @xmath103 , which means the indetermination of the spin vector in observation can reduce the alignment strength significantly . the middle panels in figure [ fig : fof ] show that the halo spin vector is aligned with the intermediate axis of the tidal tensor , and the strength of the alignment is stronger for higher mass halos . the results plotted in the right panels show that halo spin vector is aligned with the minor axis of the tidal field for low mass halos but tends to be perpendicular to it for massive halos . this mass dependence is consistent with recent results obtained for halo spin alignments with respect to filaments @xcite . as comparison , figure [ fig : so200 ] shows the alignment signals for so halos with the over - density @xmath104 . the strength of the alignment signals is only slightly weaker than that for fof halos . the observational alignment signals shown in figure [ fig : tf ] are in general agreement with n - body simulation results , in that only galaxies in halos with masses larger than @xmath142 show significant alignments with the intermediate axis . figure [ fig : so2500 ] shows the alignment signals for so halos where spin vectors are calculated only by using the inner regions within which the over - density @xmath105 . here we see that the alignment signals are weaker than those for the entire halos . in particular the tendency that spin vectors are perpendicular to the major axis seen for the total fof and so halos is now replaced by a weak alignment . there is a stronger tendency for spin axes of the inner halos to be perpendicular to @xmath35 for halos with mass larger than @xmath94 . these results match better the observational results for spiral galaxies , suggesting that disk material may follow better the angular momentum of the halo material in the inner parts . using spiral galaxies selected from the sdss dr7 and the gz2 , we have investigated the alignment between the spin of galaxies and their surrounding large - scale structure , which is characterized by the local tidal tensor . based on the eigenvalues of the tidal tensor at the location of each galaxy group , we have classified the galaxy population into four different environments : , , and . we have found that the spin axes of central spiral galaxies selected from the gz2 have a weak tendency to be aligned with the intermediate direction of the local tidal tensor . this is consistent with the theoretical prediction using the linear tidal toque theory @xcite . there is almost no significant alignment between the spin axes of spiral galaxies and the major axis @xmath33 of the local tidal tensor , while the spins of spiral galaxies have a weak tendency to be perpendicular to the minor axis @xmath35 vector ( defined as the direction of the eigenvector of the tidal tensor associated with the single negative eigenvalue @xmath31 ) . the galaxies in environment show the most prominent alignment signals . for all spiral galaxies in environment , the spin axes show a weak tendency to be perpendicular to their filaments . for galaxies in environment , there is a weak tendency for their spins to align with the norm of the sheet . using large n - body simulations , we check the projection effect in observation by constructing a control halo catalog , in which the spin vectors of @xmath103 halos have been changed to mimic the fact that it is impossible to determine which side of the galaxy is closer to us from the observed axis ratio . this indetermination can reduce the alignment strength significantly . the simulations also show that the spin vectors of the inner regions of halos have weaker alignments with the local tidal field than the spin vectors obtained for the whole halos . the observational results are consistent with the alignments obtained for the inner regions of halos , suggesting that the disk material traces the angular momentum of dark matter halos in the inner region . this work is supported by the 973 program ( no . 2015cb857002 ) , national science foundation of china ( grant nos . 11203054 , 11128306 , 11121062 , 11233005 , 11073017 ) , ncet-11 - 0879 , the strategic priority research program the emergence of cosmological structures " of the chinese academy of sciences , grant no . xdb09000000 and the shanghai committee of science and technology , china ( grant no . 12zr1452800 ) .	using a sample of spiral galaxies selected from the sloan digital sky survey data release 7 ( sdss dr7 ) and galaxy zoo 2 ( gz2 ) , we investigate the alignment of spin axes of spiral galaxies with their surrounding large scale structure , which is characterized by the large - scale tidal field reconstructed from the data using galaxy groups above a certain mass threshold . we find that the spin axes only have weak tendency to be aligned with ( or perpendicular to ) the intermediate ( or minor ) axis of the local tidal tensor . the signal is the strongest in a environment where all the three eigenvalues of the local tidal tensor are positive . compared to the alignments between halo spins and local tidal field obtained in n - body simulations , the above observational results are in best agreement with those for the spins of inner regions of halos , suggesting that the disk material traces the angular momentum of dark matter halos in the inner regions .
the aim of this study is to draw a connection between topological surgery in dimensions 1 , 2 and 3 and many natural phenomena . for this we introduce new theoretical concepts which allow to explain the topology of such phenomena via surgery and also to connect topological surgeries in different dimensions . the new concepts are the introduction of forces , attracting or repelling , in the process of surgery , the notion of solid 1- and 2-dimensional surgery and the notion of truncated 1- , 2- and 3-dimensional surgery . topological surgery is a technique used for changing the homeomorphism type of a topological manifold , thus for creating new manifolds out of known ones . a homeomorphism between two @xmath0-manifolds is a continuous bijection , such that the inverse map is also continuous . further , manifolds with homeomorphic boundary may be attached together and a homeomorphism between their boundaries can be used as ` glue ' . @xmath0-dimensional topological surgery _ on an @xmath0-manifold @xmath1 is , roughly , the topological procedure whereby an appropriate @xmath0-manifold with boundary is removed from @xmath1 and is replaced by another @xmath0-manifold with the same boundary , using a ` gluing ' homeomorphism , thus creating a new @xmath0-manifold @xmath2 ( not necessarily different from the starting one ) . for details see , for example , @xcite . apart from just being a formal topological procedure , topological surgery appears in nature in numerous , diverse processes of various scales for ensuring new results . such processes are initiated by attracting or repelling forces between two points , or _ ` poles ' _ , which seem to be joined by some _ invisible ` thread'_. to list some examples , 1-dimensional surgery happens in dna recombination and in the reconnection of cosmic magnetic lines . 2-dimensional surgery is exhibited in the formation of whirls , in blowing bubbles , in the falaco solitons and in the cell mitosis . 3-dimensional surgery can be observed , for example , in the formation of tornadoes , or the magnetic field excited by a current loop . surgery in nature is usually performed on basic manifolds with boundary . in each dimension the basic closed ( compact without boundary ) , connected , oriented ( c.c.o . ) @xmath0-manifold , on which surgery is usually performed , is the @xmath0-sphere , @xmath3 , which may be viewed as @xmath4 with all points at infinity compactified to one single point . we also need to recall that the basic connected , oriented @xmath0-manifold with boundary is the solid @xmath0-ball , @xmath5 . in particular for @xmath6 , other 3-manifolds with boundary that we will be using are : the solid torus , which can be described as the product set @xmath7 , and the handlebodies , which generalize the solid torus , having higher genus . we are particularly interested in situations related to 2- and 3-dimensional topological surgery of the attracting type . here , a _ ` drilling ' process _ along the invisible thread seems to be initiated , resulting in passage from spherical to toroidal shape . ` drilling ' with coiling seems to be a natural choice in various physical processes , probably for being the most effective way for opening up a hole . from the above , topological surgery is not just a mathematical technique used for changing the homeomorphism type of a manifold . it can also serve as a mathematical tool for explaining the change of topology in many natural phenomena . for this reason we introduce dynamics in the process of surgery . in sections 1 , 2 and 3 we recall first the mathematical definitions of topological surgery in dimensions 1 , 2 and 3 respectively . then , we introduce dynamics in topological surgery distinguishing two different types : via _ attracting forces _ and via _ repelling forces _ between two selected points , the ` poles ' . each one of these two types of dynamics can be eventually viewed as the reverse of the other . we also introduce the notions of _ solid 1- and 2-dimensional surgery _ , whereby the interior space is now filled in . also , the notions of _ truncated 1- , 2- and 3-dimensional surgery _ , whereby surgery is being localized . all these notions are better adapted to natural or physical processes exhibiting topological surgery and we address briefly in each dimension such examples . moreover , truncated 3-dimensional surgery allows to visualize 3-dimensional surgery , for which the fourth dimension is otherwise needed . this is presented in section 3 . finally , in section 4 we use all the new notions introduced in the previous sections in order to pinn down the relation among topological surgeries in dimensions 1 , 2 and 3 . the first author was always fascinated by 3-dimensional surgery and was trying to find ways to visualize it . so figure [ 3dtruncated ] dates back several years ago . further , our work is inspired by our connection of 3-dimensional topological surgery with a dynamical system @xcite . then , on one hand we will have a mathematical model for 3-dimensional surgery . on the other hand , through our connection many natural phenomena can be modelled through our dynamical system . we hope that our observations , new definitions and ideas will serve in turn as inspiration for many more interesting connections . [ [ section ] ] starting with @xmath8 , _ 1-dimensional surgery _ means that : two segments @xmath9 are removed from @xmath8 and they are replaced ( in the closure of the remaining manifold ) by two different segments @xmath10 by reconnecting the four boundary points @xmath11 in a different way . in the end we obtain two circles @xmath12 or one , depending on the type of reconnection , see figure [ formal1d ] . recall that @xmath13 consists in two points . the above definition of 1-dimensional surgery gives only the initial and the final stage . in order to adress natural phenomena exhibiting 1-dimensional surgery or to understand how 1-dimensional surgery happens we need a non - static description . we will describe the process by introducing dynamics . the process starts with two points specified on the circle , on which attracting forces are applied . then , the two segments @xmath9 , which are neighbourhoods of the two points , get close to one another . when the two segments touch , recoupling takes place giving rise to the two final segments @xmath10 , which split apart . see figure [ 1dattract ] . this type of 1-dimensional surgery shall be called _ attracting 1-dimensional surgery_. we also have the _ repelling 1-dimensional surgery _ , whereby repelling forces are applied on the two points , as illustrated in figure [ 1drepel ] . note here that the recoupling does not take place between the neighbourhoods of the two repelling points but between ` complementary ' segments , which get closer by passive reaction . [ [ section-1 ] ] in practice , 1-dimensional surgery happens on arcs or lines . that is , the initial space is the closure of @xmath14 , and we remove from it a smaller @xmath15 . we shall call this type of surgery _ truncated 1-dimensional surgery_. see figure [ 1dattrtrunc ] for the case of attracting forces . truncated 1-dimensional surgery happens , for example , on the double helix and recombines dna , thus changing the genetic sequence . see figure [ dnarecomb ] . also , in magnetic reconnection the phenomenon whereby cosmic magnetic field lines from different magnetic domains are spliced to one another changing the patterns of connectivity with respect to the sources . see figure [ magneticrecon ] ( cf . @xcite ) . : wikipedia [ dnarecomb ] : r.b . dahlburg , s.k . antiochos , _ reconnection of antiparallel magnetic flux tubes _ , j. geophysical research * 100 * , no . a9 ( 1995 ) 1699116998 . [ magneticrecon ] [ [ section-2 ] ] there are phenomena which seem to undergo the process of 1-dimensional surgery but happen on surfaces , such as tension on membranes or soap films . in order to model topologically such phenomena we introduce the notion of solid 1-dimensional surgery . _ solid 1-dimensional surgery _ on the @xmath16-disc , @xmath17 , is the topological procedure whereby a ribbon @xmath18 is being removed , such that the closure of the remaining manifold comprises two discs @xmath19 . see figure [ formal1d ] where the interior is now supposed to be filled in . this process is equivalent to performing 1-dimensional surgeries on the whole continuum of concentric circles included in @xmath17 . more precisely , and introducing at the same time dynamics , we define : we start with the @xmath16-disc of radius 1 with polar layering : @xmath20 where @xmath21 the radius of a circle and @xmath22 the limit point of the circles , that is , the center of the disc . we specify colinear pairs of antipodal points , with neighbourhoods of analogous lengths , on which the same colinear forces act , attracting or repelling , see figure [ solid1d ] . then we perform 1-dimensional surgery , attracting or repelling , on the whole continuum of concentric circles . we also define 1-dimensional surgery on the limit point @xmath22 to be the two limit points of the resulting surgeries . that is , the effect of _ 1-dimensional surgery on a point is the creation of two new points_. the above process is the same as first removing the center @xmath22 from @xmath17 , doing the 1-dimensional surgeries and then taking the closure of the resulting space , see figure [ solid1d ] . the resulting manifold is @xmath23 which comprises two copies of @xmath17 . _ attracting solid 1-dimensional surgery _ on @xmath17 is the above topological procedure whereby attracting forces act on the circles @xmath24 , see figure [ solid1d ] . _ repelling solid 1-dimensional surgery _ on @xmath17 is the above topological procedure whereby repelling forces act on the circles @xmath24 , see figure [ solid1d ] . [ [ section-3 ] ] starting with @xmath25 , there are two types of _ 2-dimensional topological surgery_. in the first type two discs @xmath26 are removed from @xmath25 and are replaced in the closure of the remaining manifold by a cylinder @xmath27 , which gets attached along the common boundary @xmath28 , comprising two copies of @xmath8 , via a homeomorphism . the gluing homeomorphism of the common boundary is a number of full twists of each copy of @xmath8 . the above operation changes the homeomorphism type from the 2-sphere to that of the torus ( see figure [ formal2da ] ) . in fact , every c.c.o . surface arises from the 2-sphere by repeated surgeries and each time the above process is performed the genus of the surface is increased by one . note that , if the cylinder were attached on @xmath25 externally , the result would still be a torus . physical examples reminiscent of 2-dimensional surgery comprise the formation of whirls and the falaco solitons @xcite ( see figure [ falaco ] ) . the other possibility of 2-dimensional surgery on the 2-sphere is the following : an annulus @xmath29 ( perhaps twisted a number of times ) is removed from @xmath25 and is replaced in the closure of the remaining manifold by two discs @xmath30 attached along the common boundary @xmath12 , resulting in two copies of @xmath25 . see figure [ formal2dr ] . phenomena exemplifying this type of surgery comprise soap bubble blowing and , similarly , glass blowing , see figure [ bubbles ] . it is worth noting that this type of surgery applied on a torus is the reverse process of the attracting type . namely , if a cylinder were removed from a torus and were replaced by two discs the result would be a 2-sphere . [ [ section-4 ] ] in order to model topologically phenomena exhibiting 2-dimensional surgery or to understand 2-dimensional surgery through continuity we need , also here , to introduce dynamics . the _ attracting 2-dimensional surgery _ starts with two poles specified on @xmath25 with attracting forces applied on them . then two discs @xmath26 , neighbourhoods of the two poles , approach each other , with a possible number of full twists . when the two discs touch , _ recoupling _ takes place and the discs get transformed into the final cylinder . see figure [ 2dattract ] . the twisting propagates along the cylinder , reminding the process of hole drilling . in the _ repelling 2-dimensional surgery _ two poles are specified on @xmath25 with repelling forces pulling them to opposite directions . this creates , by passive reaction , a cylindrical ` necking ' in the middle , which eventually tears apart and new material , two discs , gets attached along the boundary @xmath12 . see figure [ 2drepel ] . it is worth observing that the process of repelling 2-dimensional surgery in reverse time would mean that the initial surface comprises two copies of @xmath25 and there are two discs to be removed , one on each sphere , replaced by a cylinder , thus merging the two spheres into one . similarly , the process of attracting 2-dimensional surgery in reverse time would mean that the initial surface is the torus and there is a cylinder to be removed and replaced by two discs , thus yielding back the 2-sphere . in other words , the reverse process of repelling surgery ( where repelling forces are applied on the boundary circles ) can be viewed as attracting surgery ( where the attracting forces are now applied on the centers of the two discs ) and vice versa . [ [ section-5 ] ] in some natural phenomena the object undergoing surgery is not a surface but three - dimensional . for this reason we introduce also here the notion of _ solid 2-dimensional surgery_. there are two types of solid 2-dimensional surgery on the @xmath31-ball , @xmath32 , analogous to the two types of 2-dimensional surgery . the first one is the topological procedure of removing a solid cylinder homeomorphic to the product set @xmath33 , @xmath34 ( such that the part @xmath35 of its boundary lies in the boundary of @xmath32 ) and taking the closure of the remaining manifold @xmath36 , which is a twisted solid torus . see figure [ formal2da ] where the interior is supposed to be filled in . the second type is the topological procedure of removing a solid cylinder homeomorphic to the product set @xmath37 , @xmath38 , ( such that the part @xmath29 of its boundary lies in the boundary of @xmath32 ) and taking the closure of the remaining manifold @xmath39 , which is two copies of @xmath32 . see figure [ formal2dr ] where the interior is supposed to be filled in . [ [ section-6 ] ] in order to model better natural phenomena exemplifying solid 2-dimensional surgery we shall introduce dynamics : [ continuum2d ] start with the @xmath31-ball of radius 1 with polar layering : @xmath40 where @xmath21 the radius of a 2-sphere and @xmath22 the limit point of the spheres , that is , the center of the ball . _ attracting solid 2-dimensional surgery _ on @xmath32 is the topological procedure where : on all spheres @xmath41 colinear pairs of antipodal points are specified , on which the same colinear attracting forces act . the poles have disc neighbourhoods of analogous areas , see figure [ solid2dattract ] . then attracting 2-dimensional surgeries are performed on the whole continuum of the concentric spheres using the same homeomorphism @xmath42 . moreover , attracting 2-dimensional surgery on the limit point @xmath22 is defined to be the limit circle of the nested tori resulting from the continuum of 2-dimensional surgeries , see figure [ solid2dattract ] . that is , the effect _ of attracting 2-dimensional surgery on a point is the creation of a circle_. the process is characterized by the 1-dimensional core @xmath43 of the removed solid cylinder , joining the antipodal points on the outer shell and intersecting each spherical layer in the two antipodal points ; also , by the homeomorphism @xmath42 , resulting in the whole continuum of layered tori , and it can be viewed as drilling out a tunnel along @xmath43 according to @xmath42 . for @xmath42 non - trivial , this agrees with our intuition that , for opening a hole , drilling with twisting seems to be the easiest way . on @xmath32 is the topological procedure where : on all spheres @xmath41 nested annular peels of the solid annulus of analogous areas are specified and the same colinear repelling forces act on all spheres , see figure [ solid2drepel ] . then repelling 2-dimensional surgeries are performed on the whole continuum of the concentric spheres using the same homeomorphism @xmath42 , see figure [ solid2drepel ] . moreover , repelling 2-dimensional surgery on the limit point @xmath22 is defined to be the two limit points of the nested pairs of 2-spheres resulting from the continuum of 2-dimensional surgeries , see figure [ solid2drepel ] . that is , the effect of _ repelling 2-dimensional surgery on a point is the creation of two new points_. the process is characterized by the 2-dimensional central disc of the solid annulus and the homeomorphism @xmath42 , and it can be viewed as pulling apart along the central disc , after a number of twists according to @xmath42 . for @xmath42 non - trivial , this operation agrees with our intuition that for cutting a solid object apart , pulling with twisting seems to be the easiest way . in either case the above process is the same as first removing the center @xmath22 from @xmath32 , performing the 2-dimensional surgeries and then taking the closure of the resulting space . namely we obtain : @xmath44 which is a solid torus in the case of attracting solid 2-dimensional surgery and two copies of @xmath32 in the case of repelling solid 2-dimensional surgery . see figures [ solid2dattract ] and [ solid2drepel ] . the notions of 2-dimensional ( resp . solid 2-dimensional ) surgery , attracting or repelling , can be generalized from @xmath25 ( resp . @xmath32 ) to a surface ( resp . handlebody of genus @xmath45 ) creating a surface ( resp . handlebody of genus @xmath46 ) . a good natural model reminiscent of attracting solid 2-dimensional surgery is the formation of an apple ( formed from a seed ) . here no twisting occurs , so @xmath42 is trivial . repelling solid 2-dimensional surgery can be exemplified by the biological process of mitosis , where a cell splits into two new cells . see figure [ mitosis ] ( for description and instructive illustrations see for example @xcite , p. 395 ) . further , it is worth noting that the reverse process of repelling solid 2-dimensional surgery can be found in the mechanism of gene transfer in bacteria . see figure [ genetransfer ] ( for description and instructive illustrations see , for example , @xcite ) . here donor dna is transferred directly to recipient through a connecting tube " and two copies of @xmath32 merge in one . : w.t . keeton , c.h . mcfadden , _ elements of biological science _ , w.w . norton & company inc . , 3rd edition ( 1983 ) , p. 395 . [ mitosis ] : @xcite , p. 486 . [ genetransfer ] [ [ section-7 ] ] attracting solid 2-dimensional surgery can be also observed in the formation of the falaco solitons @xcite ( figure [ falaco ] ) and in the formation of whirls . the falaco solitons are pairs of singular surfaces ( poles ) connected by means of a stabilizing invisible thread . as shown in figure [ falaco ] , starting by the two poles , and by drilling along the joining line , surgery seems to be performed . based on the experimental creation of falaco solitons in a swimming pool , it has been conjectured that m31 and the milky way galaxies could be connected by a ` topological thread ' . kiehn , _ non - equilibrium systems and irreversible processes _ , adventures in applied topology * 1 * , non equilibrium thermodynamics , university of houston copyright csdc . inc , ( 2004 ) , pp . 147 , 150 . [ falaco ] in such phenomena we do not see the whole space @xmath32 ; the process can be viewed as taking place between the two attracting discs of the solid cylinder , so the initial space can be considered to be @xmath47 . this type of surgery shall be called _ truncated attracting solid 2-dimensional surgery_. in the above examples @xmath42 is non - trivial . one could also define theoretically the non - solid analogue , the _ truncated attracting 2-dimensional surgery _ as attracting 2-dimensional surgery taking place just between the two attracting discs , which are neighbourhoods of the two specified points on @xmath25 . so , the initial manifold can be considered to be just these two discs , that is , @xmath48 . another phenomenon falling topologically in the scheme of repelling solid 2-dimensional surgery is tension on metal speciments and the ` necking effect ' . more precisely , in experiments in mechanics tensile forces ( or loading ) are applied on a cylindrical speciment made of dactyle material ( steel , aluminium , etc . ) . up to some critical value of the force the deformation is homogenuous ( the cross - sections have the same area ) . at the critical value the deformation is localized within a very small area where the cross - section is reduced drastically , while the sections of the remaining portions increase slightly . this is the ` necking phenomenon ' . shortly after the speciment is fractured . view figure [ necking ] . : http://www.ara.com/projects/svo/weld.htm . [ necking ] in such phenomena we do not see the whole space @xmath32 ; the process can be seen as being localized just in the region of the solid annulus , so the initial space can be considered to be @xmath37 . this type of surgery shall be called _ truncated repelling solid 2-dimensional surgery_. one could also define theoretically the non - solid analogue , the _ truncated repelling 2-dimensional surgery _ as repelling 2-dimensional surgery taking place just in the region of the annulus @xmath29 which is complementary to the two repelling discs . so , the initial manifold can be considered to be just this annulus , that is , @xmath49 . a cross - section of 2-dimensional surgery of attracting or repelling type , truncated or solid , passing through the specified points is precisely the corresponding type of 1-dimensional surgery . [ [ section-8 ] ] in dimension 3 , the simplest c.c.o . 3-manifolds are : the 3-sphere @xmath50 and the lens spaces @xmath51 . we start with @xmath50 and we recall its three most common descriptions . firstly , @xmath50 can be viewed as @xmath52 with all points at infinity compactified to one single point : @xmath53 . see figure [ layeredspheres](b ) . @xmath52 can be viewed as an unbounded continuum of nested 2-spheres centered at the origin , together with the point at the origin , see figure [ layeredspheres](a ) , and also as the de - compactification of @xmath50 . so , @xmath50 minus the point at the origin and the point at infinity can be viewed as a continuous nesting of 2-spheres . is the compactification of @xmath52.,width=453 ] secondly , @xmath50 can be viewed as the union of two @xmath31-balls : @xmath54 , see figure [ twoballs](a ) . the two descriptions of @xmath50 are clearly related , since a ( closed ) neighbourhood of the point at infinity can stand for one of the two @xmath31-balls . note that , when removing the point at infinity in figure [ twoballs](a ) we can see the concentric spheres of the 3-ball @xmath55 ( in red ) wrapping around the concentric spheres of the 3-ball @xmath32 , see figure [ twoballs](b ) . this is another way of viewing @xmath52 as the de - compactification of @xmath50 . this picture is the analogue of the stereographic projection of @xmath25 on the plane @xmath56 , whereby the projections of the concentric circles of the south hemisphere together with the projections of the concentric circles of the north hemisphere form the well - known polar description of @xmath56 with the unbounded continuum of concentric circles . is the result of gluing two 3-balls.,width=453 ] the third well - known representation of @xmath50 is as the union of two solid tori , @xmath57 , via the torus homeomorphism @xmath58 along the common boundary . @xmath58 maps a meridian of @xmath59 to a longitude of @xmath60 which has linking number zero with the core curve @xmath61 of @xmath60 . the illustration in figure [ splittingofs3 ] gives an idea of this splitting of @xmath50 . in the figure , the core curve of @xmath60 is in dashed red . so , the complement of a solid torus @xmath60 in @xmath50 is another solid torus @xmath59 whose core curve @xmath62 ( the dashed red curve in the figure ) may be assumed to pass by the point at infinity . note that , @xmath50 minus the core curves @xmath61 and @xmath62 of @xmath60 and @xmath59 ( the red curves in figure [ splittingofs3 ] ) can be viewed as a continuum of nested tori . as a union of two solid tori.,width=415 ] when removing the point at infinity in the representation of @xmath50 as a union of two solid tori , the core of the solid torus @xmath59 becomes an infinite line @xmath62 and the nested tori of @xmath59 can now be seen wrapping around the nested tori of @xmath60 . see figure [ decomps3 ] . therefore , @xmath52 can be viewed as an unbounded continuum of nested tori , together with the core curve @xmath61 of @xmath60 and the infinite line @xmath62 . this line @xmath62 joins pairs of antipodal points of all concentric spheres of the first description . note that in the nested spheres description ( figure [ layeredspheres ] ) the line @xmath62 pierces all spheres while in the nested tori description the line @xmath62 is the ` untouched ' limit circle of all tori . viewed as two tori.,width=340 ] [ [ section-9 ] ] the third description of @xmath50 is a bit harder to connect with the first two . we shall do this here . a way to see this connection is the following . consider the description of @xmath50 as the union of two 3-balls , @xmath55 and @xmath32 ( figure [ layeredspheres](b ) ) . combining with the third description of @xmath50 ( figure [ splittingofs3 ] ) we notice that both 3-balls are pierced by the core curve @xmath62 of the solid torus @xmath59 . therefore , @xmath32 can be viewed as the solid torus @xmath60 to which a solid cylinder @xmath63 is attached via the homeomorphism @xmath58 : @xmath64 this solid cylinder is part of the solid torus @xmath59 , a ` cork ' filling the hole of @xmath60 . its core curve is an arc @xmath43 , part of the core curve @xmath62 of @xmath59 . view figure [ s3ballstotori ] . the second ball @xmath55 ( figure [ layeredspheres](b ) ) can be viewed as the remaining of @xmath59 after removing the cork @xmath63 : @xmath65 in other words the solid torus @xmath59 is cut into two solid cylinders , one comprising the ` cork ' of @xmath60 and the other comprising the 3-ball @xmath55 . as two tori to ( b ) @xmath50 as two balls.,width=491 ] [ truncate ] if we remove a whole neighbourhood @xmath55 of the point at infinity and focus on the remaining 3-ball @xmath32 , the line @xmath62 of the previous picture is truncated to the arc @xmath43 and the solid cylinder @xmath59 is truncated to the cork of @xmath32 . another way to see the connection among the different descriptions of @xmath50 is by combining the above with definition [ continuum2d ] . indeed , one can pass from the second description of @xmath50 to the third by performing attracting solid 2-dimensional surgery ( with trivial homenomorphism ) on the 3-ball @xmath32 along the arc @xmath43 . note that , by definition [ continuum2d ] , the point at the origin turns into the core curve of @xmath60 . [ [ section-10 ] ] starting with @xmath50 and its description as the splitting of two solid tori , _ 3-dimensional topological surgery _ means that a solid torus @xmath66 is removed from @xmath50 and in the closure of the remaining manifold is replaced by another solid torus @xmath67 ( with the factors reversed ) , which gets attached via a homeomorphism @xmath68 along the boundary @xmath69 of @xmath59 . this boundary ( which is a torus ) is the common boundary of @xmath59 with the complement solid torus @xmath60 . surgery starts and ends with two 3-manifolds and it may change the homeomorphism type of the initial 3-manifold . from the description above we obtain : @xmath70 the core of @xmath59 is called the _ surgery curve_. before surgery the meridians of @xmath59 bound discs , so they cut through the surgery curve ( red line @xmath62 in figure [ 3dinr3 ] ) . so , before surgery @xmath59 is layered via the indicated meridional discs . the action of the gluing homeomorphism @xmath68 is determined by specifying a @xmath71-torus knot on the boundary of @xmath59 , which is a _ parallel curve _ to the surgery curve in @xmath59 . figure [ 3dtruncatedinitial](a ) illustrates a @xmath72-torus knot on the boundary of @xmath60 . the solid torus @xmath59 is represented by the red surgery curve , which is assumed to pass by the point at infinity . note that , from the point of view of @xmath59 the above curve is a @xmath73-torus knot on the boundary of @xmath59 and it is illustrated in figure [ 3dtruncatedinitial](b ) . this @xmath71-torus knot is the image of the meridian via @xmath68 , so it becomes a meridian in the new 3-manifold and therefore it now bounds a disc ; while the meridians of @xmath59 that were bounding discs before they do not any more . see figure [ 3dinr3 ] . this exchange of roles can be clearly seen in the blue parallel curve ( left hand illustration ) turning into a blue meridional disc ( right hand illustration ) . so , after surgery , the layering of @xmath59 is via the discs bounded by the @xmath71-torus knots . this is where we need the fourth dimension to visualize 3-dimensional surgery . practically , before one could slide through a meridional disc in @xmath59 ( and could also cross the surgery curve ) , while after surgery , the only way to come closer to the surgery curve is by following the parallel @xmath71-torus knot . note that the new meridians can never reach the surgery curve . note that the appearance of the surgery line changes instantly the layering of the space from spheres to tori and initiates the instant creation of a simple closed curve @xmath61 , which is the core of the solid torus @xmath60 . [ duality ] there is an apparent duality and a natural exchange of roles of the two solid tori . therefore , the core curve of @xmath60 could be equally considered as the surgery curve . the above topological process is called _ @xmath74-rational surgery along the unknot _ and starting from @xmath50 it results in the lens space @xmath51 . in fact , by a fundamental theorem of topology , every c.c.o . 3-manifold can be created from @xmath50 by performing surgery along a knot or link ( see @xcite ) . [ [ section-11 ] ] 3-dimensional surgery is much harder to visualize than lower - dimensional surgeries . a first step in this direction is to use the de - compactification of @xmath50 . so , we define _ topological surgery in _ @xmath52 . the only difference from the definition of surgery in @xmath50 is that the surgery curve is now an infinite line @xmath62 . figure [ 3dinr3 ] illustrates surgery in @xmath52 . note that this figure resembles very much the electromagnetic field excited by a current loop which is located in the innermost torus in the drawing . here there is no apparent drilling , that is , no obvious specified longitude , but by remark [ duality ] the surgery curve is the core of the solid torus @xmath60 . .,width=529 ] [ [ section-12 ] ] a second step toward visualizing 3-dimensional surgery is achieved by removing a whole neighbourhood @xmath55 of the point at infinity . by remark [ truncate ] we are then left with a 3-ball @xmath32 , which is viewed as the solid torus @xmath60 corked by a ( bounded ) solid cylinder whose core is the arc @xmath43 , which is part of the surgery curve . a surgery in @xmath50 along an unknotted curve passing by the point at infinity would correspond to surgery in @xmath32 along the arc @xmath43 , the core of a solid cylinder . this solid cylinder is complemented by the outer ball @xmath55 , which is homeomorphic to another solid cylinder , to form together the solid torus @xmath59 . the above lead to the following ` localized ' definition of 3-dimensional surgery . a _ truncated 3-dimensional surgery _ in a 3-manifold @xmath1 is a 3-dimensional surgery , such that the surgery curve passes through the point at infinity , and such that a neighbourhood of the point at infinity is removed . this definition can help us visualize step - by - step 3-dimensional surgery along the unknot in @xmath50 , especially the formation of the new meridian in the resulting lens space . for this we shall consider for simplicity a @xmath75-torus knot as the specified parallel curve . view figure [ 3dtruncated ] . we start with a solid cylinder , which is a part of the solid torus @xmath59 . on its boundary a @xmath75-curve ( blue ) is specified which is parallel to the core curve ( red ) . then the solid cylinder gets thicker and it is transformed into a 3-ball . then opposite parts of the cylinder move inwardly and at the same time a twisting takes place that results in ` straightening ' of the parallel curve . then merging and recupling takes place resulting in a hole ; thus the solid cylinder is turned into a solid torus on which the blue curve bounds now a disc . note that the solid torus @xmath60 surrounding the initial solid cylinder is omitted in the figure . [ [ section-13 ] ] considering now life - like situations , we will proceed with inserting dynamics in truncated 3-dimensional surgery . this can help understand the topological mechanism behind some natural or physical phenomena . we start with the attracting type . consider two points in 3-space , surrounded by spherical neighbourhoods , say @xmath76 and @xmath77and assume that on these points strong attracting forces act . view figure [ 3dattract ] . as a result , a ` joining thread ' , say @xmath43 , is created between the two points and ` drilling ' along @xmath43 is initiated . the joining arc @xmath43 is seen as part of a simple closed curve @xmath62 passing by the point at infinity . this is the surgery curve . further , the two 3-balls @xmath76 and @xmath77 together with the space in between make up a solid cylinder , the ` cork ' ( cf . figure [ s3ballstotori ] ) . let @xmath60 be a solid torus , which filled by the cork gives rise to a 3-ball @xmath32 , such that the centers of the two balls @xmath76 and @xmath77 lie on its boundary ( compare with figure [ s3ballstotori ] ) . the process of attracting 3-dimensional surgery restricted in @xmath32 shall be called _ attracting truncated 3-dimensional surgery_. note that the cork in the above definition is complemented by a solid cylinder , a tubular neighbourhood @xmath55 of the arc @xmath78 , to the solid torus @xmath59 , the complement of @xmath60 in @xmath50 . this completes our familiar picture . we shall then define _ repelling truncated 3-dimensional surgery _ to be the dual phenomenon to the above , whereby strong repelling forces are applied on the two points , so strong as to initiate attracting surgery in the complementary 3-ball @xmath55 , along the segment @xmath78 with central point the point at infinity . [ [ section-14 ] ] structural similarities exhibited on vastly different scales of the universe can be visualized and explained with 3-dimensional surgery . a natural phenomenon resembling strongly the process of truncated 3-dimensional surgery is the formation of tornadoes , see figure [ tornado ] . namely , if certain meteorological conditions are met , funnel - shaped clouds start descending toward the ground . once they reach it , they become tornadoes . drawing the analogy to 3-dimensional surgery , first the poles are chosen , one on the tip of the cloud and the other on the ground , and they seem to be joined through an invisible line . then , starting from the first point , the wind revolves in a helicoidal motion toward the second point , resembling ` hole drilling ' along the line until the hole is drilled . topologically speaking , in this case seems to be undergoing rational surgery along the unknot . : http://www.smartsuburbansurvival.com/category/natural-disasters/tornadoes.dpbs and noaa ( http://www.photolib.noaa.gov/htmls/wea00308.htm ) [ tornado ] there are other examples exhibiting topological behaviour of 3-dimensional surgery . figure [ babystar ] , for example , illustrates a dusty disc closely encircling a massive baby star " . : http://www.spitzer.caltech.edu/news/1153-feature10-11-unravelling-the-mystery-of-star-birth-dust-disk-discovered-around-massive-star [ babystar ] note that solid 2-dimensional surgery can be almost viewed as the intermediate stage of 3-dimensional surgery . indeed , there is a great resemblance between solid 2-dimensional surgery and truncated 3-dimensional surgery . they both begin with a solid ball and there is ` drilling ' occuring along a ` cork ' , a solid cylinder passing through the center . in fact , by definition [ continuum2d ] , the solid 2-dimensional surgery is responsible for the creation of the curve @xmath61 in truncated 3-dimensional surgery . yet , there is a crucial difference : in solid 2-dimensional surgery the cylindrical cork is removed afterwards and we left with just the solid torus @xmath60 with its core curve @xmath61 . while , in truncated 3-dimensional surgery , matter is still there ( surrounding the arc @xmath43 ) but it is completely altered . the above descriptions explain the connection between 2-dimensional and 3-dimensional topological surgery , up to now not so explicitly presented . the meeting ground is the three - space with solid 2-dimensional surgery on the one end and truncated 3-dimensional surgery on the other end . we shall now go a bit further than that and explain the connection of attracting surgeries in all three dimensions . view figure [ crossections ] . on the left - hand top and bottom pictures we see truncated 3-dimensional surgery . taking on the top picture the intersection with the boundary of the 3-ball @xmath32 we pass to the initial picture of attracting 2-dimensional surgery , where two points with surrounding discs are specified . restricting truncated 3-dimensional surgery only to this sphere results in the final stage of attracting 2-dimensional surgery ( middle bottom illustration ) . taking finally the intersection with a meridional plane gives rise to 1-dimensional surgery ( rightmost illustrations ) . topological surgery occurs in numerous natural phenomena of varying scales where two points ( poles ) are selected and attracting or reppeling forces are applied . examples of such phenomena comprise : dna recombination , magnetic reconnection , mitosis , gene transfer , the creation of falaco solitons , the formation of whirls and tornadoes and magnetic fields . in this paper we tried to pinn down the connection of such phenomena with topological surgery . in order to do this we first enhanced the static description of topological sugery of dimensions 1 , 2 and 3 by introducing dynamics by means of attracting or repelling forces between two ` poles ' . we then filled in the interior space in 1- and 2-dimensional surgery , introducing the notion of solid 1- and 2-dimensional surgery . this way more natural phenomena can be accommodated in the connection . finally we fitted many more natural phenomena in the connection by intoducing the notion of truncated 1- , 2- , and 3-dimensional topological surgery , whereby surgery is more localized . thus , instead of considering surgery as an abstract topological process , it can now be viewed as a property of many natural phenomena . on the other hand , all these new notions enabled us understand and visualize 3-dimensional surgey and reveal the relation between topological surgeries in all three lower dimensions . in @xcite these notions are used for connecting 3-dimensional topological surgery with a dynamical system . then , phenomena related to 3-dimensional surgery could be modelled by this dynamical system . we hope that through this study , topology and dynamics of natural phenomena as well as topological surgery may now be better understood and that our connections will serve as ground for many more insightful observations .	topological surgery occurs in natural phenomena where two points are selected and attracting or repelling forces are applied . the two points are connected via an invisible ` thread ' . in order to model topologically such phenomena we introduce dynamics in 1- , 2- and 3-dimensional topological surgery , by means of attracting or repelling forces between two selected points in the manifold , and we address examples . we also introduce the notions of solid 1- and 2-dimensional topological surgery , and of truncated 1- , 2- and 3-dimensional topological surgery , which are more appropriate for modelling natural processes . on the theoretical level , these new notions allow to visualize 3-dimensional surgery and to connect surgeries in different dimensions . we hope that through this study , topology and dynamics of many natural phenomena as well as topological surgery may now be better understood . `` = 1 ''
this work is devoted to analytical study of ideal incompressible emhd flows ( see , e.g. , @xcite and references therein about emhd and its applications ) . our primary goal here is to consider a simplified 1d problem that has many similar qualitative properties with the problem about attractive interaction of two distributed currents in 2d ideal emhd , that was numerically simulated recently with a high resolution @xcite . more precisely , we introduce and partially analyse local approximations for particular class of the 2d ideal emhd flows , that are reduced in mathematical sense to dynamics of a single or few 1d objects , the vortex lines . the most interesting result of present work is the demonstration of exact solvability by the known hodograph method of long - scale dynamics in the unstable vortex structure constituted by two nearly anti - parallel sheets of the generalized vorticity in 2d ideal emhd . as known , the emhd model approximately describes dynamics of the low - inertial electron component of plasma in situations when the heavy ion component is almost motionless and serves just to provide a neutralizing background for electrically charged electron fluid and to keep a constant concentration @xmath0 of the electrons . the ( divergence - free in this case ) electric current @xmath1 creates the quasi - stationary magnetic field , @xmath2 which contributes to the generalized electron vorticity , @xmath3 the most simple way how to derive the ideal emhd equation of motion is just to use the well known fact that the generalized vorticity in an ideal homogeneous fluid is frozen - in , @xmath4.\ ] ] as the result , the corresponding equation of motion can be represented in the remarkable form @xmath5,\ ] ] where the hamiltonian functional of ideal incompressible emhd in the fourier representation is given by the expression @xmath6 here @xmath7 or @xmath8 depending on dimensionality of the problem and the electron inertial length @xmath9 is introduced . below we normalize all length scales to this quantity . our analytical approach is based on the representation of ideal emhd in terms of frozen - in lines of the generalized vorticity @xmath10 , as described , for instance , in @xcite . the general form ( [ omega_hamiltonian ] ) of equation of motion allows one to represent the field @xmath10 through the shapes of frozen - in vortex lines ( the so called formalism of vortex lines ) , @xmath11 where @xmath12 is the 3d delta - function , @xmath13 is some fixed 2d manifold depending on the particular problem , @xmath14 is a label of vortex line , @xmath15 is an arbitrary longitudinal parameter along the line . dynamics of the line shape @xmath16 is determined by the variational principle @xmath17 , with the lagrangian of the form @xmath18\cdot{{\bf { d}}}({{\bf { r}}}))d\xi \,\ , -{\cal h}\{{{\bf { \omega}}}\{{{\bf { r}}}\}\},\ ] ] where the vector function @xmath19 must satisfy the only relation @xmath20 below we take @xmath21 . now we apply this formalism to the 2d case , when the three - component field @xmath22 does not depend on the @xmath23-coordinate . the field @xmath24 can be parameterized by two scalar functions , @xmath25 and @xmath26 , @xmath27 because of the freezing - in property , the @xmath28-function is just transported by the @xmath29-component of the velocity field , that results in conservation of the integrals @xmath30 with arbitrary function @xmath31 . if initially @xmath32 was piecewise constant , then at any time we have a flow with cylindrical sheets of frozen - in generalized vorticity . each such cylinder is numbered by a number @xmath33 , has a constant in time value @xmath34 of the jump of @xmath35 , and consists of a family of closed ( if @xmath36 ) vortex lines with identical shape but with different shift along @xmath23-axis , @xmath37 , where @xmath15 is a longitudinal parameter along a line , @xmath38 is the shift . obviously , the number @xmath39 together with the sift @xmath38 serve in this case as the 2d label @xmath40 . for 2d ideal emhd in the physical space we have from eq.([emhd_hamiltonian ] ) the double integral @xmath41 where @xmath42 is the modified bessel function of the second kind . we do not write the exact coefficient in front of this expression since it only influences on a time scale and thus is not very interesting for us . as follows from equations written above , dynamics of this set of contours in 2d ideal incompressible emhd is determined by the lagrangian @xmath43 where the new constants @xmath44 are proportional to the corresponding jumps of @xmath28 function , @xmath45 and so on . for a given contour number , locally , a cartesian coordinate can be used as the longitudinal parameter , for instance , the @xmath46-coordinate . in this case the function @xmath47 plays the role of the canonical coordinate , while @xmath48 plays the role of the canonical momentum . thus , we have a `` natural '' system with the hamiltonian being the sum of a quadratic on the generalized momentum `` kinetic energy '' and a `` potential energy '' @xmath49 depending on the shape of the contours in @xmath29-plane , or , in other words , on the @xmath28 function . in emhd the `` potential energy '' describes the interaction between parallel electric currents . at this point it is interesting to compare the 2d emhd with the usual eulerian 2d hydrodynamics , which differs from ( [ h_2d_emhd ] ) by the @xmath50-function instead of the @xmath51-function . in that case @xmath28 function is just the @xmath23-component of the velocity field , and the `` potential energy '' @xmath49 is an integral of motion for eulerian 2d flows , as follows from the expression @xmath52 and from eq.([f_integrals ] ) with @xmath53 . equations of motion , that follow from the variational principle with the lagrangian like ( [ exact_lagrangian ] ) , but with `` @xmath54 '' instead of `` @xmath51 '' , are such that this term does not have influence on the contour dynamics in @xmath29-plane , only it adds a linear function of the time to @xmath55-coordinate of a vortex line . this property corresponds to conservation of the @xmath23-component of the velocity in 2d eulerian flows for each moving element of the fluid . obviously , in 2d emhd such conservation does not take place . for practical analytical calculations the system ( [ exact_lagrangian ] ) is not very convenient because of the non - locality . however , since the @xmath51-function is exponentially small at large values of its argument , it is possible to introduce local approximations for long - scale dynamics . let us first have a single contour of a large size @xmath56 . then for smooth configurations approximate local equations of motion ( with the time appropriately rescaled ) can be obtained by variating the expression @xmath57 which naturally arises after we perform one integration in the double integral in eq.([exact_lagrangian ] ) with ( almost ) straight shape of the adjacent piece ( a few units of @xmath58 ) of the contour . although for us this system seems to be very interesting and deserving much attention , now we concentrate on another case and consider unstable vortex structure constituted by two close contours . let us now consider the case of two contours with equal jumps @xmath59 , symmetric with respect to the line @xmath60 , and parameterize ( locally ) their shapes as @xmath61 and @xmath62 , as shown in fig.[sketch ] in the small frame . if a typical wave - length @xmath63 along the @xmath46-direction is large comparatively to both @xmath64 and @xmath65 , then in the long - scale localized approximation , with additional condition @xmath66 , we have from eq.([exact_lagrangian ] ) @xmath67[1-\exp(-2y)]\right\}dx,\ ] ] after an appropriate time rescaling we can put the `` const '' in the above expression equal to @xmath68 for convenience . then , after introducing new quantities @xmath69 and @xmath70 , as well as the function @xmath71 , @xmath72 it is possible to write down the corresponding equations of motion in the following remarkable general form , @xmath73 as known , any nonlinear system of such form can be locally reduced to a linear equation after taking as the new independent variables @xmath74 and @xmath75 ( the so called hodograph method ; see , e.g. , @xcite for a particular case ) . indeed , since from ( [ mu_definition ] ) and ( [ bernoulli_general ] ) we see @xmath76 it is useful to introduce an auxiliary function @xmath77 as @xmath78 in order to obtain @xmath79 from here we easily derive @xmath80 after that we rewrite eq.([contin_general ] ) as @xmath81 and multiply it by the jacobian @xmath82 : @xmath83 thus , now we have @xmath84 differentiating this equation over @xmath74 with taking into account eqs.([t]-[x ] ) and subsequent simplifying give us the linear partial differential equation for the function @xmath85 : @xmath86 it is also useful to write down here the general equation for the function @xmath77 : @xmath87 thus , the localized approximation ( [ approx_lagrangian ] ) appears to be integrable in the sense that it is reduced to solution of a _ linear _ equation . however , the functions @xmath85 and @xmath88 are multi - valued in general case . therefore statement of the cauchy problem for the time evolution of the system [ originally the cauchy problem was formulated in @xmath89-representation in terms of initial functions @xmath90 and @xmath91 at @xmath92 now becomes much more complicated , since in @xmath93-plane initial data are placed on the parametrically given curve @xmath94 , @xmath95 which can have self - intersections . it should be noted here that for @xmath77 initial data are determined directly by eqs.([t]-[x ] ) , while for @xmath85 their determination needs additional differentiation of eq.([x ] ) over @xmath74 . besides this , the particular function @xmath71 given by eq.([h_rm ] ) results in the _ elliptic _ partial differential equation for the function @xmath85 , @xmath96_{\rho\rho}+e^{-\rho}[(1+\mu^2)t]_{\mu\mu}=0,\ ] ] in contrast to the usual 1d gas - dynamic case described in @xcite , where the corresponding equation is _ hyperbolic_. generally speaking , the ellipticity makes the cauchy problem ill - posed in the mathematical sense , if the initial data are not very smooth . however , for sufficiently smooth initial data the problem remains correctly formulated though still difficult for complete solution . nevertheless , the linear equation seems to have an advantage , and we hope with its help to investigate more easily the problem of classification of possible singularities in this system . in the future work we will discuss how the quantity @xmath74 can tend to zero at some point @xmath46 . it should be also noted that an analogous approach is useful in studying another unstable vortex structure , the pair of anti - parallel vortex filaments in the usual hydrodynamics and in other hydrodynamic - type models @xcite ( the corresponding instability in eulerian hydrodynamics is known as the crow instability ) . for instance , if we consider nonlinear development of the crow instability in long - scale limit , then the localized approximation for symmetric ( respectively to the plane @xmath60 ) dynamics of the vortex pair gives us the hamiltonian @xmath97 where @xmath98 is the ( small ) width of the filaments . taking the @xmath46-coordinate as a longitudinal parameter @xmath15 , we have the system like ( [ mu_definition]-[bernoulli_general ] ) , but with @xmath99 , @xmath100 investigation of the corresponding linear equation for the function @xmath85 is now in progress . these investigations were supported by the intas ( grant no . 00 - 00292 ) , by rfbr ( grant no . 00 - 01 - 00929 ) , by the russian state program of support of the leading scientific schools ( grant no . 00 - 15 - 96007 ) , and by the science support foundation , russia .	the formalism of frozen - in vortex lines for two - dimensional ( 2d ) flows in ideal incompressible electron magnetohydrodynamics ( emhd ) is formulated . a localized approximation for nonlinear dynamics of two close sheets of the generalized vorticity is suggested and its integrability by the hodograph method is demonstrated .
be stars are early main - sequence stars which have shown h@xmath0 emission at least once ( e.g. * ? ? ? they are rapidly rotating close to their break up velocity @xcite . the stars are surrounded by decretion disks @xcite , formed from ejected material that settles into a keplerian disk and spreads outwards through viscous diffusion @xcite . the mechanism for ejection is uncertain ( rotation alone is insufficient ) , with magnetic fields ( e.g. * ? ? ? * ; * ? ? ? * ) or stellar pulsation @xcite being candidates . the disk and its associated emission appears and disappears on a timescale of a few months to a few years ( e.g. * ? ? ? gas from the _ decretion _ disk of the be star can be captured and _ accreted _ by a companion in be / x - ray binaries . these systems contain a be star in a relatively wide orbit ( orbital period of tens to hundreds of days ) of significant eccentricity ( @xmath1 ) with a compact object , often a neutron star @xcite or less commonly a black hole @xcite . typically the orbital angular momentum of the binary is misaligned to the spin of the be star , with a relative inclination of @xmath2 ( see table 1 in @xcite and references therein ) . both the misalignment and the orbital eccentricity are plausibly the result of kicks imparted during the supernova that formed the compact object ( e.g. * ? ? ? * ; * ? ? ? * ) . in this paper , we investigate how the interaction of the misaligned disk and eccentric binary may give rise to the observed x - ray phenomenology of be / x - ray binaries . two types of x - ray outbursts are produced in such systems @xcite . _ type i _ x - ray outbursts are periodic and coincide with periastron passage with a typical luminosity of @xmath3 that lasts for a few days . _ type ii _ x - ray outbursts occur much less frequently but have a larger luminosity of @xmath4 and last weeks to months , showing no orbital modulation . the majority of giant outbursts are finished in less than an orbital period , but some may last up to five orbital periods @xcite . several suggestions as to the origin of these outbursts have been made @xcite , all of which involve as a central element the tilted and warped disks that have been observed before and during the outbursts @xcite . most recently , @xcite , presented analytic arguments suggesting that simultaneous growth of eccentric and tilt modes could occur on time - scales relevant to the giant outbursts . exactly how a warped ( and potentially eccentric ) disk causes a long duration accretion event onto the neutron star remains , however , unclear . the outline of the paper is as follows . in 2 we present results from three dimensional numerical simulations of accretion in be / x - ray binaries . type i outbursts can be produced when the neutron star captures a small fraction of low angular momentum gas from the disk at closest approach . we show that type ii outbursts can be explained as a consequence of eccentricity growth within the be star disc . capture of larger amounts of gas from the eccentric disk generates a type ii outburst , and in 3 we show that the long duration of these events can be due to the increased viscous time when the material circularizes around the compact object . we discuss our results in 4 and 5 . .be / x - ray binary model properties [ cols="<,<,<,<,<",options="header " , ] [ tab ] previous work on misaligned be star systems mainly focused on coplanar systems or those with a small misalignment angle , @xmath5 ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? they found that the resonant truncation radius for the be star disk is the same for the inclined case unless @xmath6 ( @xmath7 is the angle between the spin of the be star and the binary orbital axis ) . in this work we use the smoothed particle hydrodynamics ( sph ) code . phantom @xcite to model highly misaligned systems . phantom has been used to model misaligned disks in other binary systems @xcite and we use the same method here . the simulations have two sink particles , one representing a be star with mass @xmath8 and the other a neutron star of mass @xmath9 . the binary has an orbital period of @xmath10 ( thus the orbit has semi - major axis @xmath11 ) and eccentricity @xmath12 , to target the be / x - ray binary 4u 0115 + 63 , although our results are more generally applicable . the parameters of the simulation are summarized in table [ tab ] . [ plot1]-[plot3 ] show the overall evolution of the calculation . the accretion radius of the be star is equal to its stellar radius of @xmath13 . for the neutron star we take the accretion radius to be @xmath14 ( the bondi - hoyle accretion radius at periastron is about @xmath15 ) . thus , we do not model the inner parts of the accretion disk around the neutron star ( if it forms ) . fig . [ plot1 ] shows the initial flat , but tilted disk with an inclination of @xmath16 to the @xmath17-axis . the disk has a mass of @xmath18 distributed with a surface density power law @xmath19 between @xmath20 and @xmath21 with @xmath22 sph particles . varying the outer disk radius , we find little difference in the disk behaviour for smaller disk radii once the disk has expanded to the tidal truncation radius . the disk is isothermal with an aspect ratio @xmath23 at the inner edge , where @xmath24 , and the @xcite @xmath0 parameter varies in the range @xmath25 ( e.g. * ? ? ? * ) over the disk ( we implement the disk viscosity in the usual manner by adapting the sph artificial viscosity according to the procedure described in @xcite , using @xmath26 and @xmath27 ) . the be star disc is resolved with shell - averaged smoothing length per scale height @xmath28 . the left hand panel of fig . [ outburst ] shows the mass flow rate through the neutron star accretion radius as a function of time . initially small ( type i ) accretion outbursts occur every orbital period . [ plot2 ] show the disk structure during a type i outburst at a time of @xmath29 . these outbursts occur when the disk is close to circular and the neutron star passes close to the spiral arm of the be star disk . a small amount of material is pulled from the spiral arm on to the neutron star . fig [ outburst ] shows that at a time of around @xmath30 the accretion changes phase . the peak accretion rate is about an order of magnitude higher than during the type i phase and the accretion does not subside between the periastron passages . we suggest that this is a giant type ii outburst . this can be seen clearly in the right panel of fig . [ outburst ] where we show the total mass accreted on to the neutron star as a function of time . the mass increases slowly during the type i phase and then dramatically during the type ii phase before returning to the type i phase . the top row of panels in fig . [ plot3 ] show the disk structure during the type ii outburst at a time of @xmath31 . the be star disk is highly eccentric and it is this that causes the type ii outburst where a large amount of material is transferred to the neutron star . the bottom row of panels show the end of the outburst and the neutron star disk . the precession of the eccentric be star disk can also be seen by comparing the disk structures at different times . we bin the particles in ( instantaneous ) radius from the be star and calculate properties of the be star disk . initially the disk is misaligned by @xmath16 . just before the type ii outburst , the disk has aligned to @xmath32 . the disk remains approximately flat during this process with the outer parts aligning only a few degrees more than the inner parts . this is in agreement with the analytic estimates for the alignment timescale for the be star disk with the binary orbit of the order of a few years @xcite . furthermore , we find that the eccentricity grows first at large radius in the disk . just before the type ii outburst , the eccentricity of the disk is approximately constant with radius at @xmath33 . during and after the giant outburst , the eccentricity decays . there are several mechanisms that may drive eccentricity in a circumprimary disk that we discuss here . first , eccentricity may be directly excited due to the eccentric companion . a perturber on an eccentric orbit forces an eccentricity on an initially circular particle orbit . the particle eccentricity grows nearly linearly in time ( see section iii.c of @xcite ) . we have performed test particle simulations of a coplanar eccentric binary system including the eccentric forcing of the companion ( see equation ( 9 ) in * ? ? ? . the initially constant rate of eccentricity growth of the particle increases strongly with distance from the be star , @xmath34 , as @xmath35 ( for @xmath36 ) . for example , a particle with an initial distance of @xmath37 , the eccentricity grows linearly at a rate of @xmath38 per binary orbit . thus , the disk becomes eccentric on a timescale of around 10 orbital periods . the eccentricity of the particle orbit grows perpendicularly to the eccentricity of the binary orbit . this is observed in our simulations and thus we suggest that gravitational forcing from the companion is the most likely mechanism for eccentricity growth here . hence , the disk must expand radially in order for eccentricity to be driven . we will investigate this process in the inclined disk in more detail in a future publication . secondly , instability at the 3:1 lindblad resonance can lead to eccentricity growth even in an initially circular disk @xcite . the radius of the 3:1 resonance is @xmath39 ( e.g. * ? ? ? * ) , which for our typical parameters is @xmath40 . finally , it could be possible for eccentricity to be driven by the stream from the outside of the be star disk falling back towards and hitting the be star disk @xcite . however , there is only a small amount of material involved in the streams . in the coplanar case , the tidal truncation radius of the disk may be estimated by the radius at which ballistic particle orbits cross . in a circular binary with our parameters this would be at a radius of @xmath41 @xcite . for our parameters this is @xmath42 . in an eccentric binary , the size of the disk is further reduced . for our parameters , the truncation radius is around @xmath43 @xcite . in an eccentric binary , the disk eccentricity ( with no dissipation ) grows at all radii but the rate is strongly dependent on radius . thus large eccentricity growth is limited to high inclination disks where reduced tidal torques allow the disk to expand to large radii ( nixon et al . in prep ) . for smaller disks ( lower inclinations ) , dissipation can keep the disk close to circular with @xmath44 @xcite . we have run several sph simulations with varying misalignment angle and we find that the disk becomes significantly eccentric only for @xmath45 . thus , type ii outbursts are expected only in systems with a large misalignment angle . here we consider properties of the accretion disk that forms around the neutron star . due to the misalignment between the be star disk and the neutron star orbit it is likely that accretion on to the neutron star proceeds through an initially misaligned disk ( c.f . * ) . for numerical tractability we can not model the neutron star disk in its entirety and it remains under - resolved even during the peak of the type ii outburst . this will be investigated in more detail in a future publication . however , we have run several sph simulations with a much smaller neutron star accretion radius . we do not find significant evidence for a neutron star accretion disk during the type i activity , only during the type ii . there is much less material and its circularization radius is much smaller during the type i outburst phase and thus it accretes more rapidly ( see also * ? ? ? * ) . the viscous timescale in the neutron star disk is @xmath46 , where @xmath47 is the distance from the neutron star , the viscosity is @xmath48 , @xmath49 is the @xcite parameter and the sound speed is @xmath50 where @xmath51 is the gas constant and @xmath52 is the gas mean molecular weight . the temperature of the disk surface is found by assuming a steady profile @xmath53 @xcite . thus , the accretion timescale of the disk scaled by the orbital period is @xmath54 if the disc is significantly optically thick this is an overestimate . while a radially extended neutron star disk is present , the accretion on to the neutron star is smoothed on a timescale of several orbital periods . this explains why there is no orbital modulation observed during a type ii x - ray outburst . we do not expect smoothing on this timescale to take place during the type i activity because the accreting material has lower initial angular momentum and therefore a smaller circularisation radius ( @xmath55 ) and viscous time . these numbers are approximate and tend to over - predict the viscous time compared to observations as we assume a steady disk . time dependent processes , such as tides or interactions of the neutron star disk with streams from the be star disk could shorten the viscous timescale . the corresponding x - ray luminosity of the disk is @xmath56 , where @xmath57 is the radius of the neutron star and thus @xmath58 thus , for the accretion rates to agree with observations of the x - ray luminosities , we require an initial disk mass of around @xmath59 or smaller . this agrees well with the be star mass loss rates that are in the range @xmath60 to @xmath61 ( e.g. * ? ? ? * ; * ? ? ? * ) and the lifetimes of the be star disks ( of a few years ) . the simulations presented here model a warped , eccentric , precessing be star disk . the ejection mechanism for material from the be star to the disk is poorly understood and thus we have considered only the evolution following disk formation . thus , the disc is an accretion rather than a decretion disc , but over the timescales we simulate this should not affect our conclusions . the ejection of material from the be star surface in a circular orbit at the inner edge of the disc could affect the eccentricity growth and inclination of the disc . however , because eccentricity growth is driven more strongly in the outer parts of the disc , this should not affect the type ii outburst mechanism . during the type ii outburst in our simulation , the be star disk becomes significantly depleted . about @xmath62 of the initial disk mass is accreted on to the be star . this depletion could explain the cycles observed in the optical light curves of be / x - ray binaries ( e.g. * ? ? ? . with mass ejection from the be star included , the disk would build up again and we would expect the type ii outbursts to recur . the timescale of recurrence would depend on the accretion rate on to the be star disk . parameters such as the eccentricity , misalignment and binary separation could be varied to model other be / x - ray binary systems and predict the types of outbursts that may occur . observations suggest that the type ii outbursts occur when the disk becomes warped . however we have found that this is not a sufficient condition , it is necessary for the disk to become eccentric . an inclined circular disk , which can expand freely as the tidal torques are smaller , becomes eccentric before it becomes large enough for a ns encounter . if there are no outbursts observed in a be binary system , the binary must have small eccentricity and misalignment . conversely , if type ii outbursts are observed , then the system must be highly misaligned . we find that eccentricity may also be driven in counter rotating discs . the frequency of outbursts in be / x - ray binary systems will be investigated in future with a disc model including mass injection into the disc from the be star in order to determine critical inclination angles required for outbursts . the tidal torque aligns the disk with the binary orbital plane on a timescale of years @xcite . in a long term disk simulation this alignment may also play a role in ending the outburst cycle . we have performed 3d sph simulations of a moderately eccentric and highly misaligned be / x - ray binary . we find that for a binary with orbital period of @xmath63 and eccentricity @xmath64 that type i outbursts occur for all misalignments , including the aligned case . if the disk is sufficiently misaligned , @xmath45 , the disk is much larger because the tidal torque is weaker . this allows strong eccentricity growth in the outer parts of the disk from the gravitational forcing of the eccentric companion . thus , the disk becomes eccentric on a timescale of around ten binary orbital periods . we suggest that the eccentric disk causes a type ii x - ray outburst as a large amount of material is transferred to the neutron star . rgm s support was provided under contract with the california institute of technology ( caltech ) funded by nasa through the sagan fellowship program . support for cjn was provided by nasa through the einstein fellowship program , grant pf2 - 130098 . pja acknowledges support from nasa s atp program under awards nnx11ae12 g and nnx14ab42 g . djp is supported by future fellowship ft130100034 from the australian research council . we acknowledge the use of splash @xcite for the rendering of the figures . this work utilised the janus supercomputer , which is supported by the national science foundation ( award number cns-0821794 ) , the university of colorado boulder , the university of colorado denver , and the national center for atmospheric research . the janus supercomputer is operated by the university of colorado boulder . , a. t. & hayasaki , k. 2007 , in astronomical society of the pacific conference series , vol . 361 , active ob - stars : laboratories for stellare and circumstellar physics , ed . a. t. okazaki , s. p. owocki , & s. stefl , 395	be / x - ray binary systems exhibit both periodic ( type i ) x - ray outbursts and giant ( type ii ) outbursts , whose origin has remained elusive . we suggest that type ii x - ray outbursts occur when a highly misaligned decretion disk around the be star becomes eccentric , allowing the compact object companion to capture a large amount of material at periastron . using 3d smoothed particle hydrodynamics simulations we model the long term evolution of a representative be / x - ray binary system . we find that periodic ( type i ) x - ray outbursts occur when the neutron star is close to periastron for all disk inclinations . type ii outbursts occur for large misalignment angles and are associated with eccentricity growth that occurs on a timescale of about 10 orbital periods . mass capture from the eccentric decretion disk results in an accretion disk around the neutron star whose estimated viscous time is long enough to explain the extended duration of type ii outbursts . previous studies suggested that the outbursts are caused by a warped disk but our results suggest that this is not sufficient , the disk must be both highly misaligned and eccentric to initiate a type ii accretion event . [ firstpage ]
the stereo mission provides us an unprecedented view of the solar corona , enabling us for the first time to fully constrain the three - dimensional ( 3-d ) geometry of the coronal magnetic field . stereoscopic triangulation of coronal loops has been conducted at small stereo spacecraft separation angles ( @xmath5 ) , for several active regions observed with stereo a(head ) and b(ehind ) in april and may 2007 ( aschwanden et al . 2008a , b ; 2009 ) . the reconstructed 3-d geometry of stereo - observed coronal loops has been compared with theoretical magnetic field models based on extrapolations from photospheric magnetograms , using nonlinear force - free field ( nlfff ) models ( derosa et al . 2009 ) , as well as potential and stretched potential field models ( sandman et al . 2009 ) , but surprisingly it turned out that the two types of magnetic field lines exhibited an average misalignment angle of @xmath6 , regardless of what type of theoretical magnetic field model was used . from this dilemma it was concluded that a more realistic physical model is needed to quantify the transition from the non - force - free photospheric boundary condition to the nearly force - free field at the base of the solar corona ( derosa et al . 2009 ) . at this junction , it is not clear what a viable method is to obtain a force - free boundary of the magnetic field at the coronal base , or how to correct the non - force - free magnetograms . however , the stereoscopic triangulation supposedly provides the correct 3-d directions of the magnetic field @xmath7 , which together with maxwell s equation of divergence - freeness ( @xmath8 , constrain also the absolute values of the field strengths . in this paper i we choose a magnetic field model that is defined in terms of multiple unipolar charges . an approach in terms of multiple dipoles is employed in paper ii ( sandman and aschwanden 2010 ) . since both unipolar or dipolar magnetic fields represent potential magnetic fields that fulfill the divergence - free condition , the superposition of multiple unipolar and dipolar magnetic field components fulfill the same condition . we develop a numerical code of such a parameterized divergence - free magnetic field that can be forward - fitted to the 3-d geometry of stereoscopically triangulated coronal loops . so , the simple goal of this study is to evaluate how closely the stereoscopically observed loops can be modeled in terms of potential fields , a goal that was already attempted with skylab observations ( sakurai and uchida 1977 ) . modeling with non - potential fields , such as nonlinear force - free field ( nlfff ) models , will be considered in future studies . this paper is organized as follows : the definition of a parameterized potential field is described in section 2 , the development and tests of a numeric magnetic field code and the results of forward - fitting to stereoscopically triangulated loops is presented in section 3 , and conclusions follow in section 4 . an alternative approach with dipolar magnetic fields is the subject of paper ii ( sandman and aschwanden 2010 ) . since the coronal plasma-@xmath9 parameter is generally less than unity ( gary 2001 ) , the magnetic pressure exceeds the thermal pressure , and thus all soft x - ray or euv - emitting plasma that fills or flows through coronal flux tubes traces out the coronal field . consequently , stereoscopic triangulation of euv loops provides the correct 3-d field directions along coronal loops . we denote the normalized 3-d field direction along a loop with the unity vector @xmath10 , which is parameterized as a function of a loop length coordinate @xmath11 , starting at the footpoint position @xmath12 at the base of the corona , @xmath13 \over \sqrt{b_x^2+b_y^2+b_z^2 } } \ , \ ] ] where the magnetic field is defined by the cartesian components @xmath14 $ ] . however , the absolute magnitude of the magnetic field strength , @xmath15 , is not known a priori . for a physical solution of the magnetic field , maxwell s equation of a divergence - free magnetic field has to be satisfied , @xmath16 which ( in its integral form ) corresponds to the theorem of magnetic flux conservation along a fluxtube , @xmath17 where @xmath18 is the integral over the cross - sectional area of a fluxtube defined in perpendicular direction to the magnetic field line ( at loop position @xmath19 ) . therefore , since the stereoscopic triangulation defines the magnetic field directions in adjacent flux tubes , it defines also the divergence of the field and the relative change of the magnetic field strength @xmath20 along the flux tubes , and this way implicitly defines also the isogauss surfaces perpendicular to each fluxtube , and therefore the full 3-d vector field @xmath21 , except for a scaling constant . the derivation of a 3-d magnetic field @xmath21 in an active region , @xmath22 requires also the knowledge of the scalar function @xmath23 in every 3-d location @xmath24 , which we constrain with a forward - fitting method of a divergene - free field model . a divergence - free 3-d magnetic field model @xmath7 can be parameterized by a superposition of divergence - free fields , because the divergence - free condition is linear ( or abelian ) , i.e. , if two components @xmath25 and @xmath21 fulfill @xmath26 and @xmath27 , then also their sum is divergence - free , @xmath28 . divergence - free magnetic field components are , for instance , a parallel field , unipolar fields ( a magnetic charge with spherical isogauss surfaces and a field that falls off with the square of the distance ) , dipole fields , quadrupolar fields , other multi - pole representations , or any potential field . the abelian property warrants that any superposition of divergence - free fields is also divergence - free . specifically , we will use divergence - free potential field models that consist of either _ i ) _ multiple unipolar charges ( in this paper i ) , or _ ii ) _ multiple dipoles ( in paper ii ) . our philosophy is the following : we will employ magnetic field models of the category of potential - field models , which are divergence - free by definition . we use particular potential field models that can be quantified with a finite number of free parameters . potential field models are not as general as non - potential and force - free field ( nlfff ) models . however , since both potential and nlllf models currently exhibit an equally poor misalignment with observed euv loops , we need first to investigate whether the misalignement between observations and _ any _ theoretical model can be mimized for the simplest class of magnetic field models , such as potential field models . if our approach proves to be successfull , refinements with non - potential or nlllf models can then be pursued in the future along the same avenue ( e.g. , see conlon and gallagher 2010 ) . unipolar potential fields often provide a good approximation to the magnetic field of sunspots , and thus can be used also for an active region that is composed of a finite number of spot - like magnetic polarities . conceptually , a unipolar field can be considered as an approximation to the upper half of a vertically positioned dipole . the simplest representation of a unipolar magnetic field that is a potential field , and hence fulfills maxwell s divergence - free condition , is a spherically symmetric field that drops off with an @xmath29-dependence with distance , so it has a potential function that drops off with @xmath30 , @xmath31 where @xmath32 is the potential field value at the solar surface vertically above the buried magnetic charge in depth @xmath33 . of course , the extrapolated magnetic field is only computed in the coronal domain @xmath34 , so that no `` magnetic monopole '' exists in the solar corona . the magnetic field model of a single magnetic charge requires 4 parameters : the maximum value of the potential field @xmath32 and the location @xmath35 of the buried charge , having a distance @xmath36 to any point @xmath37 in the solar corona , @xmath38^{1/2 } \ .\ ] ] the resulting magnetic field has then only a radial component @xmath39 in direction of @xmath40 , @xmath41 with the surface field strength @xmath42 . this unipolar potential field fulfills the divergence - free condition , as it can be calculated from the laplacian operator of the potential function , @xmath43 the fulfillment of the divergence - free condition can also be verified from the conservation of the magnetic flux theorem ( eq . 3 ) , if the envelope of a fluxtube is defined by radial field lines , so that the cross - sectional area @xmath44 remains constant for @xmath45 . in our first model we employ a superposition of @xmath46 multiple unipolar charges , @xmath47 in terms of the vector @xmath48 $ ] . for a single unipolar charge , the field lines will all be straight lines in radial direction away from the buried charge , which can approximate open - field regions . burying multiple magnetic charges of opposite magnetic polarity , however , can mimic closed - field regions . an example is given in fig . 1 , where we compare the magnetic field of a dipole with that of a combination of two unipolar charges with opposite magnetic polarity . actually , the two magnetic field models become identical when the two unipolar charges are moved close together at the location of the dipole moment , as it can be shown mathematically . although the two models are equivalent in the far - field approximation , a combination of two unipolar charges ( with @xmath49 free parameters ) allows more general solutions than a single dipole ( with 6 free parameters ) , especially in the case of strongly asymmetric fields ( sunspots ) or open - field regions , as they exist is most active regions . the forward - fitting of our analytical magnetic field model to a set of observed magnetic field vectors @xmath50 ( e.g. , using stereoscopically triangulated loop coordinates ) , is the task of optimizing the free parameters of the analytical model until the best match with the observed field lines is obtained . for the evaluation of the goodness or consistency of the analytical magnetic field models @xmath51 with the observed field line model @xmath52 , we define the 3-d misalignment angle @xmath53 , which is defined by the scalar product between the two field vectors @xmath54 and @xmath55 , @xmath56 or equivalently , between the unity field vectors @xmath57 and @xmath58 , @xmath59 this parameter is a single value at every spatial point @xmath24 , which can be averaged at the coronal base or over the lengths of the observed field lines , at @xmath60 positions , @xmath61^{1/2 } \ , \ ] ] which is similar to a @xmath62-criterion . since a unipolar magnetic charge can be parameterized with 4 free parameters ( @xmath63 ( eq . 9 ) , a model with @xmath60 components has @xmath64 free parameters . our strategy to bootstrap the coronal magnetic field with stereoscopic data consists of the following steps : ( 1 ) we create a model of subphotospheric magnetic charges by deconvolving an observed photospheric magnetogram into point charges , which defines the unit vectors of our parameterized theoretical magnetic field model @xmath65 ; ( 2 ) we perform stereoscopic triangulation for a set of coronal loops observed with stereo / euvi , which are quantified in terms of directional field vectors @xmath66 ; ( 3 ) we forward - fit the theoretical magnetic field model @xmath67 with a number of free parameters in the setup of unipolar magnetic charges by minimizing the mean misalignment angle @xmath68 , and which yields a best - fit solution @xmath51 . we can then compare the minimized misalignment of the bootstrapped best - fit model with those of standard methods based on extrapolation of the the photospheric boundaries using a magnetogram , e.g. , with the _ potential field source surface ( pfss ) _ model . the procedure is illustrated in fig . 2 , where a dipolar euv loop is observed and a magnetic field model is constructed from two unipolar charges . by adjusting the field strength of the second unipolar charge from @xmath69 to @xmath70 , the misalignment angle between the observed euv loop and the model field can be reduced from @xmath71 ( fig . 2 top ) to @xmath72 ( fig . 2 , bottom ) . note that the adjusted field is still a potential field and divergence - free , but represents a better match to the observed euv loop . we select four active regions observed with stereo / euvi and the _ michelson doppler imager ( mdi ) _ ( scherrer et al . 1995 ) onboard the _ solar and heliospheric observatory ( soho ) _ : 2007 april 30 , may 9 , may 19 , and dec 11 . the first ar 10953 ( 2007 apr 30 ) is identical to the case previously analyzed with stereo and hinode ( derosa et al . 2009 ; sandman et al . the second ar 10955 ( 2007 may 9 ) was subject of the first stereoscopically triangulated coronal loops , temperature and density measurements , and stereoscopic tomographic reconstruction ( aschwanden et al . 2008a , b , 2009 ; sandman et al . the third ar 10953 ( 2007 may 9 ) displayed a small flare ( during 12:40 - 13:20 ut ) as well as a partial filament eruption during the time of observations , and was featured in a few studies ( li et al . 2008 ; liewer et al . 2009 ; sandman et al . the fourth ar 10978 ( 2007 dec 11 ) is also subject of recent magnetic field modeling ( aad van ballegooijen and alec engell , private communication 2010 ) . some details of these four active regions are given in table 1 , such as the heliographic position of the ar center , the magnetic area for fluxes of @xmath73 g , the minimum and maximum field strengths , and the total unsigned magnetic flux . all four active regions were observed with the _ michelson doppler imager ( mdi ) _ onboard the _ solar and heliospheric observatory ( soho ) _ , which provides full - disk mdi magnetograms with a pixel size of 2 . subimages encompassing the active region of interest are shown in fig . 3 ( left column ) , with quadratic field - of - view sizes ranging from 145 to 339 pixels , or 0.3 to 0.7 solar radii , respectively . in order to create a realistic 3d magnetic field model we decompose the partial magnetograms into a number of @xmath74 positive and negative gaussian 2-d components , using an iterative 2-d gaussian fitting scheme that determines local maxima in decreasing order of field strengths . the composite magnetogram of these 200 decomposed sources is shown in fig . 2 ( middle column ) , and the difference to the original magnetogram is also shown ( right column ) . for each of the 200 magnetic source components we store the peak magnetic field value @xmath75 , the center position ( @xmath76 ) , and the half widths @xmath77 at full maximum . for a parameterization in terms of unipolar charges , however , we need to convert the half width @xmath77 into the corresponding depth @xmath78 at which the unipolar charge is buried . from the definition of the full width at half maximum ( fwhm ) at @xmath79 of a unipolar field ( eq . 9 ) , we have for the vertical magnetic field component @xmath80 ( see geometry in fig . 4 ) , @xmath81 with @xmath82 the angle between the vertical and a surface ring with radius @xmath83 ( fig . 4 ) , from which we can calculate the dipole depth @xmath78 , @xmath84 which corresponds approximately to the half width @xmath83 of the fitted gaussian component . since we have now all input parameters @xmath85 , @xmath86 , for a definition of a magnetic field model with multiple unipolar magnetic charges ( eq . 9 ) , we can calculate the full 3d magnetic field by superimposing the fields of all components . a particular field line is simply calculated by starting with the field at a footpoint position and by iterative stepping in the field direction , until the field line returns to the solar surface ( in case of closed field lines ) or to a selected boundary of the computation box ( for open field lines ) . for each of the four active regions we triangulate as many loops as can be discerned with a highpass filter in image pairs from stereo / euvi a and b. the method of stereoscopic triangulation is described in detail in aschwanden ( 2008a ) , from which we use identical loop coordinates for ar 10955 observed on 2007 may 9 . the stereoscopic triangulation requires accurate coalignment of both euvi / a and b images in an epipolar coordinate system . furthermore , we transform the 3-d loop coordinates measured in the epipolar coordinate system with the line - of - sight of the euvi / a spacecraft into the coordinate system of mdi , which sees the sun from the langrangian point l1 , almost in the same direction as seen from earth . the advantage of transforming the euvi loop coordinates into the mdi reference frame is the direct modeling of the longitudinal magnetic field component @xmath87 in the mdi reference frame , without requiring any knowledge of the absolute magnetic field strength @xmath88 , which is model - dependent , i.e. , it depends on the choice of the extrapolation method ( potential , force - free , or nlfff ) from photospheric magnetograms ( which measures only the longitudinal component @xmath87 ) . some parameters of the stereoscopic triangulation procedure are listed in table 1 , such as the heliographic coordinates of the active region , the spacecraft separation angle , and the number of triangulated euv loops ( varying between 70 and 200 per active region , combined from all three coronal wavelengths , 171 , 195 , and 284 ) . 3-d representations of the stereoscopically triangualted euv loops are shown in figs . 5 - 8 ( with blue color ) , seen along the line - of - sight of mdi ( in greyscale maps of figs . 5 - 8 ) and in the two orthogonal directions ( sideview and topview in figs . 5 - 8 ) . the height range of stereosopically triangulated loops generally does not exceed 0.1 solar radii , due to the drop of dynamic range in flux for altitudes in excess of one hydrostatic scale height . after we have parameterized the magnetic field with @xmath89 unipolar charges , each one defined by 4 parameters @xmath85 , using the procedure of iterative decomposition of a photospheric magnetogram as described in section 3.2 , we can vary these free model parameters to adjust it to the 3d geometry of the stereoscopically triangulated loops . however , since we typically represent an mdi magnetogram with @xmath90 components , we have @xmath91 free parameters , which are too many to optimize independently . standard optimization procedures , such as the powell algorithm ( press et al . 1986 ) , allow for optimization of @xmath92 free parameters , with good convergence behavior if the initial guesses are suitably chosen . therefore we choose 10 - 20 small circular zones ( with typical radii of @xmath93 solar radii ) containing the strongest field regions in the magnetogram of equal magnetic polarity and optimize the magnetic field strengths @xmath94 with a common correction factor @xmath95 in each zone . this is an empirical optimization of the lower boundary of the coronal magnetic field , optimized by varying the ( decomposed ) photospheric mdi magnetogram in such a way that the resulting potential field more closely matches the stereoscopically triangulated loops . the results of the best fits , based on the mimization of the median misalignment angle ( eqs . 11 - 12 ) , are shown in figs . 5 - 9 , in form of ( red ) model field lines that are extrapolated at the same locations as the footpoints of the stereoscopic loops ( blue in figs . 5 - 9 ) . the distribution of all misalignment angles , evaluated at about 80 locations of every loop is shown in form of histograms at the bottom of figs . 5 - 9 . each histogram is characterized with a gaussian peak , which are found to have a mean value and standard deviation of @xmath96 ( fig.5 : 2007 apr 30 ) , @xmath97 ( fig.6 : 2007 may 9 ) , @xmath98 ( fig.7 : 2007 may 19 ) , and @xmath99 ( fig.8 : 2007 dec 11 ) . if we compare these misalignment angles in the range of @xmath100 with the previously measured values using a potential field source surface ( pfss ) model ( @xmath101 ; sandman et al . 2009 ) , or using a nonlinear force - free field ( nlfff ) model ( @xmath102 ; derosa et al . 2009 ) , we see an improvement of the misalignment angle by about a factor of two . this is a remarkable result that demonstrates that the magnetic field at the lower boundary of the corona can be bootstrapped with stereoscopic measurements and with a suitable parameterization of a potential field model . although the extrapolation from a photospheric magnetogram should be unique , if there is no data noise present , an infinite number of potential field solutions can be obtained depending on how the boundary field is parameterized ( e.g. , by unipolar charges or dipolar magnetic moments ) . in our case , for every variation of the @xmath103 free parameters , a slightly different field with a different misalignment to the stereosocpic loops is obtained . the pfss code is designed to compute the potential field of the entire ( front and backside ) sun , and thus has a relatively coarse spatial resolution of typically @xmath104 ( one heliographic degree , i.e. , 12 mm ) for standard computations , but the small - scale magnetic field should not matter too much for our large - scale loops ( @xmath105 mm ) . hence , the best - fit potential - field bootstraps a force - free photospheric boundary field that is significantly different from the observed photospheric magnetograms which is observed in a non - force - free zone . the remaining misalignment between our best - fit potential field solution and the stereoscopic loops could be due to three sources of errors : ( 1 ) the bootstrapping method did not converge to the best solution ; ( 2 ) the 3d coordinates of the stereoscopically triangulated loops have some error ; or ( 3 ) the real coronal magnetic field that is represented by the stereoscopic loops could be non - potential . we consider the convergence of our code as satisfactory , because we ran many attempts for each case with different initial conditions and obtained about the same minimum misalignment . one possibility to vary the initial conditions is to vary the number of ( gaussian ) unipolar components . 9 shows the convergence behavior as a function of the number of ( decomposed ) unipolar components , where we computed a best - fit potential field solution for @xmath106 unipolar components for each of the 4 active regions . convergence to the minimum misalignment value typically recquires @xmath107 components for the simplest dipolar active regions ( 2007 may 9 or dec 11 ) , and @xmath108 for more complex active regions ( 2007 apr 30 or may 19 ) . we investigate the second possible source of errors that could contribute to the measured misalignment , i.e. , errors associated with the stereoscopic triangulation method . the errors of stereoscopic triangulation have been discussed in aschwanden et al . ( 2008a ) , which depend on ( 1 ) the ratio of the stereoscopic parallax angle to the spatial resolution of the instrument , and thus on the spacecraft separation angle , ( 2 ) the angle between the loop segment and east - west direction in the epipolar plane , being largest for loop segments parallel to the epipolar plane , and ( 3 ) the proper identification of a corresponding loop in image b to a selected loop in image a. while the first two sources of errors can be formally calculated , the correspondence problem is difficult to quantify . since stereoscopic triangulation can not be accomplished in an automated way at present time , the error of identifying corresponding loops could be estimated from the scatter obtained with different observers , but this is time - consuming . here we pursue another approach that is based on a self - consistency test . the procedure works as follows . if a sufficient large number of loops are triangulated in an active region , there should be for every triangulated loop ( which we call primary loop ) a neighbored ( secondary ) loop with an almost parallel direction . whatever the direction of the true magnetic field in the same neighborhood is , both the primary @xmath109 and the secondary stereo loop @xmath110 should have a similar misalignment angle with the local magnetic field , @xmath111 , or @xmath112 , in order to be self - consistent . therefore we can define a stereoscopic error ( se ) angle @xmath113 by averaging these differences in misalignements over all stereo loop positions with suitable weighting factors @xmath114 , @xmath115 where the index @xmath116 runs over all loop positions for each primary loop and the index @xmath117 runs over all loop positions of secondary loops , excluding the primary loop . for the weighting factor @xmath118 we should give less weight to more distant neighbors , because the true magnetic field and thus the local misalignment angle is likely to increase with distance . thus we should choose a negative power of the relative distance @xmath119 , @xmath120^{-p/2 } \ , \ ] ] where @xmath121 is a power index . if the index @xmath121 is small , say @xmath122 the weighting is reciprocal to the distance and the long - range neighbors are relatively strongly weighted . if the index @xmath121 is large , say @xmath123 , the short - range neighbors are relatively strongly weighted while the longe - range neighbors have almost no weight . so we expect that the so - defined stereoscopic error monotonously decreases with short - range weighting towards higher power indices @xmath124 . in fig . 10 we show the stereoscopic error calculated with eqs . ( 15 - 16 ) as a function of the power index from @xmath122 to @xmath125 , for all 4 active regions . indeed , the stereoscopic error monotonously decreases from @xmath122 to @xmath126 , because we give progressively less weight to the long - range misalignments . however , from @xmath126 towards @xmath125 the error increases again , probably because of the inhomogeneity of the closest neighbors at the shortest range and the excessive weighting to the very nearest neighbors . however , the plateau in the range of @xmath127 is a good indication that we measure a stable value at the minumum of @xmath128 . from this we obtain a misalignment contribution of @xmath129 ( 2007 apr 30 ) , @xmath130 ( 2007 may 9 ) , @xmath131 ( 2007 may 19 ) , and @xmath132 ( 2007 dec 11 ) , as listed in table 2 . a diagram of the various misalignment angles is shown in fig . 11 , which we plot as a function of the maximum goes soft x - ray flux during the observing interval . the goes flux was dominanted by the emission from the active regions analyzed here , since there was no other comparable active region present on the solar disk during the times of the observations . the case of 2007 may 9 with the lowest goes shows a single dipolar structure , while the case of 2007 may 19 with the highest goes flux exhibits multiple dipolar groups . the misalignment angles @xmath133 indicate those obtained with the _ potential field source surface ( pfss ) _ code , @xmath134 those with the _ nonlinear force - free field ( nlfff ) _ code ( only for 2007 apr 30 ) , @xmath135 those optimized with the _ potential field with unipolar charge ( pfu ) _ code , and @xmath113 indicates the stereosocpic errors . as a working hypothesis we might attribute the remaining residuals @xmath4 to the non - potentiality of the active regions , @xmath136 for which we measure the values : @xmath137 ( 2007 apr 30 ) , @xmath138 ( 2007 may 9 ) , @xmath139 ( 2007 may 19 ) , and @xmath140 ( 2007 dec 11 ) , listed also in table 2 . we consider these values as a new method to quantify the nonpotentialiy of active regions . we find that quiescent active regions that contain simple dipoles ( 2007 apr 30 and may 9 ) have small misalignments of @xmath141 with best - fit potential field models , while more complex active regions ( 2007 may 19 and dec 11 ) have somewhat larger non - potential misalignments in the order of @xmath142 compared with best - fit potential field models . while we express thge degree of nonpotentiality in terms of a mean misalignment angle here , other measures are the ratios of the total nonpotential to the potential energy in an active region , which amounts up to @xmath143 in one flaring active region ( schrijver et al . 2008 ) . since the four investigated active regions have quite different misalignment angles , and also the inferred nonpotentiality of the magnetic field varies significantly , we quantify the activity level of the active regions from the soft x - ray flux measured by goes . 12 shows the goes light curves for the 4 active regions during the times of stereoscopic triangulation . the lowest goes flux is measured for the ar of 2007 may 9 , with a level of @xmath144 w m@xmath145 ( goes class a4 ) , while the highest goes flux is measured for the ar of 2007 may 19 , which has a flare occurring during the observing period with a goes flux of @xmath146 w m@xmath145 ( goes class c0 ) . the three active regions with the low soft x - ray flux levels appear to be quiescent , judging from the goes flux profile , or may be subject to micro - flaring at a low level . there is a clear correlation between the soft x - ray level of the active region ( when it was on disk ) and the overall misalignment angle ( fig . 11 ) , as well as with the misalignment angle @xmath4 attributed to the non - potentiality , varying from @xmath147 for the lowest goes a - class levels to @xmath148 for an active region with a goes c - class flare . a higher soft x - ray flux generally means a higher heating rate with stronger impulsive heating or flaring . a higher degree of non - potentiality , on the other hand , indicates the presence of a higher level of electric currents ( which are non - potential ) . therefore , the observed correlation suggests a physical relationship between the electric currents in an active region and the amount of heating input . this is not surprising , since evidence for current - carrying emerging flux was demonstrated previously for h@xmath149 and soft x - ray structures that are non - potential ( e.g. , leka et al . 1996 ; jiao et al . 1997 ; schmieder et al . our measurement of the degree of non - potentiality with the magnetic field misalignment averaged over the entire active region , is a very coarse technique , but a more detailed investigation of the misalignment in separate parts of the active region that are quiescent or flaring will be pursued in paper ii . the agreement between theoretical magnetic field models of active regions in the solar corona with the true 3-d magnetic field as delineated from the stereoscopic triangulation of coronal loops in euv wavelengths has never been quantified until the recent advent of the stereo mission . to everybody s surprise , the average misalignment between the theoretical and observed magnetic field was quite substantial , in the amount of @xmath150 for both potential and nonlinear force - free field models ( derosa et al . 2009 ; sandman et al . 2009 ) . in this study we investigate the various contributions of this large misalignment for four different active regions observed with stereo and arrive at the following conclusions : 1 . the amount of misalignment can be reduced to about half of the values for potential - field models optimized by a bootstrapping method that minimizes the field directions with the stereoscopically triangulated loops . our potential - field model is parameterized with @xmath151 unipolar charges per active region , whose positions and field strengths are approximately derived from a gaussian decomposition of a photospheric magnetogram , and then varied until a best fit is obtained . the best - fit potential field model has an improved misalignment of @xmath152 . because the best - fit potential field model defines an improved magnetic field boundary condition at the bottom of the corona , the difference to the observed photospheric magnetogram contains information on the currents between the photosphere and the base of the force - free corona . 2 . we estimate the misalignment contriubtion caused by stereoscopic correlation errors from self - consistency measurements between the magnetic field misalignments of adjacent loops . we find contributions in the order of @xmath153 . we estimate the contributions to the field misalignment due to non - potentiality caused by electric currents from the residuals between the best - fit potential field and the stereoscopic triangulation errors and find misalignment contributions in the order of @xmath154 . 4 . the overall average misalignemnt angle between potential field models and stereoscopic loop directions , as well as the contribution to the misalignment due to non - potentiality , are found to correlate with the soft x - ray flux of the active region , which suggests a correlation between the amount of electric currents and the amount of energy dissipation in form of plasma heating in an active region . in this study we identify for the first time the contributions to the misalignment of the magnetic field , in terms of optimized potential field models , non - potentiality due to electric currents , and stereoscopic triangulation errors . these results open up a number of new avenues to improve theoretical modeling of the coronal magnetic field . first of all , optimized potential field models can be found that represent a suitable lower boundary condition at the base of the force - free corona , which provides a less computing - expensive method than nonlinear force - free codes . second , methods can be developed that allow us to localize electric currents in the non - force - free photophere and chromosphere . third , the misalignment angle can be used as a sensitive parameter to probe the evolution of current dissipation , energy build - up in form of non - potential magnetic energy in different quiescenct and flaring zones of active regions . the high - resolution magnetic field data from hinode and _ solar dynamics observatory _ provide excellent opportunities to obtain better theoretical models of the coronal magnetic field using our bootstrapping method , which is not restricted to stereoscopic data only , but can also be applied to single - spacecraft observations . acknowledgements : we are grateful to helpful discussions with marc derosa and allen gary . this work was partially supported by the nasa contract nas5 - 38099 of the trace mission and by nasa stereo under nrl contract n00173 - 02-c-2035 . the stereo / secchi data used here are produced by an international consortium of nrl , lmsal , ral , mpi , isas , and nasa . the mdi / soho data were produced by the mdi team at stanford university and nasa . # 1 [ aschwanden , m.j . , 2004 , _ physics of the solar corona - an introduction _ , praxis publishing ltd . , chichester uk , and springer , berlin . ] [ aschwanden , m.j . , wuelser , j.p . , nitta , n. , and lemen , j.r . : 2008a , , 827 . ] [ aschwanden , m.j . , nitta , n.v . , wuelser , j.p . , and lemen , j.r . : 2008b , , 1477 . ] [ aschwanden , m.j . , nitta , n.v . , wuelser , j.p . , lemen , j.r . , and sandman , a. : 2009 , , 12 . ] [ conlon , p.a . and gallagher , p.t . 2010 , apj ( submitted ) . ] [ derosa m.l . , schrijver , c.j . , barnes , g . , leka , k.d . , lites , b.w . , aschwanden , m.j . , amari , t . , canou , a . , mctiernan , j.m . , regnier , s . , thalmann , j . , valori , g . , wheatland , m.s . , wiegelmann , t . , cheung , m.c.m . , conlon , p.a . , fuhrmann , m . , inhester , b . , and tadesse , t . 2009 , , 1780 . ] [ gary , a. 2001 , solar phys . 203 , 71 . ] [ jiao , l. , mcclymont , a.n . , and mikic , z. 1997 , 174 , 311 . ] [ leka , k.d . , canfield , r.c . , mcclymont , a.n . , and van driel - gesztelyi , l. 1996 , 462 , 547 . ] [ li , y. , lynch , b.j . , stenborg , g. , luhmann , j.g . , huttunen , k.e.j . , welsch , b.t . , liewer , p.c . , and vourlidas , a. 2008 , , l37-l40 . ] [ liewer , p.c . , dejong , e.m . , hall , j.r . , howard , r.a . , thompson , w.t . , culhane , j.l . , bone , l . , and vandriel - gesztelyi , l . 2009 , , 57 - 72 . ] [ press , w.h . , flannery , b.p . , teukolsky , s.a . , and vetterling , w.t . 1986 , _ numerical recipes . the art of scientific computing _ , cambridge university press : new york . ] [ sakurai , t. and uchida , y. 1977 , , 397 . ] [ sandman , a.w . , aschwanden , m.j . , derosa , m.l . , wuelser , j.p . , and alexander , d. 2009 , , 1 . ] [ sandman , a.w . and aschwanden , m.j . , 2010 , ( in preparation ) , ( paper ii ) . ] [ scherrer , p.h . 1995 , solar phys . 162 , 129 . ] [ schmieder , b. , demoulin , p. , aulanier , g. , and golub , l . 1996 , 467 , 881 . ] [ schrijver , c.j . , derosa , t. metcalf , g. barnes , b. lites , t. tarbell , j. mctiernan , g. valori , t. wiegelmann , m.s . wheatland , t. amari , g. aulanier , p. demoulin , m. fuhrmann , k. kusano , s. regnier , and j.k . thalmann 2008 , 675 , 1637 . ] lllrrrrr 10953 ( s05e20 ) & 2007-apr-30 & 23:00 - 23:20 & 5.966 & 200 & 24.2 & [ -3134,+1425 ] & 8.7 + 10955 ( s09e24 ) & 2007-may-9 & 20:30 - 20:50 & 7.129 & 70 & 4.4 & [ -2396,+1926 ] & 1.6 + 10953 ( n03w03 ) & 2007-may-19 & 12:40 - 13:00 & 8.554 & 100 & 12.2 & [ -2056,+2307 ] & 4.0 + 10978 ( s09e06 ) & 2007-dec-11 & 16:30 - 16:50 & 42.698 & 87 & 8.2 & [ -2270,+2037 ] & 4.8 + lllll misalignment nlfff@xmath155 & 24@xmath15644 & & & + misalignment pfss@xmath157 & 25 @xmath158 8 & 19 @xmath1586 & 36 @xmath15813 & 32 @xmath15810 + misalignment pfu@xmath159 & 14.3@xmath15811.5 & 13.3@xmath1589.3 & 20.3@xmath15816.5 & 15.2@xmath15812.3 + median pfu@xmath159 & 20.0 & 16.2 & 25.8 & 15.7 + stereoscopy error@xmath160 & 9.4 & 7.6 & 11.5 & 8.9 + nonpotentiality@xmath161 & 4.9 & 5.7 & 8.8 & 7.2 + goes soft x - ray flux@xmath162&@xmath163 & @xmath144 & @xmath146 & @xmath164 + goes class & a7 & a4 & c0 & b1 +	we investigate the recently quantified misalignment of @xmath0 between the 3-d geometry of stereoscopically triangulated coronal loops observed with stereo / euvi ( in four active regions ) and theoretical ( potential or nonlinear force - free ) magnetic field models extrapolated from photospheric magnetograms . we develop an efficient method of bootstrapping the coronal magnetic field by forward - fitting a parameterized potential field model to the stereo - observed loops . the potential field model consists of a number of unipolar magnetic charges that are parameterized by decomposing a photospheric magnetogram from mdi . the forward - fitting method yields a best - fit magnetic field model with a reduced misalignment of @xmath1 . we evaluate also stereoscopic measurement errors and find a contribution of @xmath2 , which constrains the residual misalignment to @xmath3 , which is likely due to the nonpotentiality of the active regions . the residual misalignment angle @xmath4 of the potential field due to nonpotentiality is found to correlate with the soft x - ray flux of the active region , which implies a relationship between electric currents and plasma heating .
the investigation of ordered double perovskite materials @xmath12 with @xmath3 an alkaline earth such as sr , ba , or ca and @xmath13 two different transition metals has been strongly stimulated by the discovery of a large room temperature magnetoresistive effect at low magnetic fields in sr@xmath6femoo@xmath1 @xcite . the fact that the double perovskites seem to be ferromagnetic metals with high curie temperatures @xmath4 of up to 635k @xcite and apparently have highly spin polarized conduction band makes these materials interesting for applications in spintronic devices such as magnetic tunnel junctions or low - field magnetoresistive sensors @xcite . however , the double perovskites are also of fundamental interest since both their basic physics and materials aspects are not well understood . it is evident that intensive research has been dedicated to both the variation of the metallic / magnetic ions on the @xmath14- and @xmath15-site as well as electron doping studies , where the divalent alkaline earth ions on the @xmath3-site are partially replaced by a trivalent rare earth ion such as la , in order to understand the electronic structure and the magnetic exchange in the double perovskites in detail . furthermore , these studies aimed for the tailoring and optimization of the magnetic properties of the double perovskites for their use in magnetoelectronic devices such as spin valves , magnetic information storage systems , or as sources for spin polarized electrons in spintronics . here , important aspects are the achievement of sufficiently high values for @xmath4 and the spin polarization to allow for the operation of potential devices at room temperature . along this line , in the compound sr@xmath6crreo@xmath1 a curie temperature of @xmath16k has been obtained @xcite . furthermore , in sr@xmath6femoo@xmath1 an increase in @xmath4 of about 70k has been reported as a result of electron doping by partial substitution of divalent sr@xmath9 by trivalent la@xmath10 @xcite . with respect to magnetoresistance , up to now large low - field magnetoresistive effects have been found not only in sr@xmath6femoo@xmath1@xcite , but also in many other double perovskites as for example sr@xmath6fereo@xmath1 @xcite , sr@xmath6crwo@xmath1 @xcite , and ( ba@xmath17sr@xmath18)@xmath19la@xmath20femoo@xmath1 @xcite . the origin of magnetism in the double perovskite still is discussed controversially . historically , the ferrimagnetism in the system sr@xmath6femoo@xmath1 has been explained in terms of an antiferromagnetic superexchange interaction between the mo@xmath21 ( @xmath22 ) spin and the fe@xmath10 ( @xmath23 ) spins @xcite . however , more recently moritomo _ et al . _ found a strong correlation between the room temperature conductivity and the curie temperature , implying that the mobile conduction electrons mediate the exchange interaction between the local fe@xmath10 spins @xcite . therefore , in analogy to the doped manganites it was tempting to explain the ferromagnetic coupling between the fe sites based on a double exchange mechanism , where the delocalized electron provided by the mo @xmath24 configuration plays the role of the delocalized @xmath25 electron in the manganites . however , as pointed out by sarma @xcite there are important differences between the manganites and the double perovskites . in the former , both the localized mn @xmath26 electrons and the delocalized mn @xmath25 electron reside at the same site and their spins are coupled ferromagnetically by a strong on - site hund s coupling . in the latter , the localized fe @xmath27 electrons ( fe@xmath10 : @xmath28 ) and the delocalized mo @xmath29 electron ( mo@xmath21 : @xmath30 ) nominally are at two different sites although the mo @xmath29 electron obtains a finite fe character by sizable hopping interaction . at first glance this seems to support a double exchange szenario . however , in sr@xmath6femoo@xmath1 according to band structure calculations the fe @xmath27 spin - up band is completely full making it impossible for another spin - up electron to hop between fe sites and thus forcing the delocalized electrons to be spin - down electrons . therefore , the hund s coupling strength , which provides the energy scale for the on - site spin coupling in the double exchange mechanism for the manganites , can not be invoked for sr@xmath6femoo@xmath1 . this demonstrates that in the double perovskites the antiferromagnetic coupling between the localized and the delocalized electrons , which according to the large @xmath4 is strong , must originate from another mechanism . such mechanism has been proposed by sarma @xcite for sr@xmath6femoo@xmath1 and extended to many other systems by kanamori and terakura @xcite . in this model , the hybridization of the mo @xmath29 ( @xmath26 ) and fe @xmath27 ( @xmath26 ) states plays the key role in stabilizing ferromagnetism at high curie temperatures @xcite . if the fe@xmath10 spins are ferromagnetically ordered , the hybridization between the fe @xmath27 ( @xmath26 ) and the mo @xmath29 ( @xmath26 ) states pushes up and down the mo @xmath31 and mo @xmath32 states , respectively . the essential point in this scenario , which is discussed in more detail below , is that the hybridized states are located energetically between the exchange - split fe @xmath33 and the fe @xmath34 levels . we note that in this model the magnetic moment at the mo site is merely induced by the fe magnetic moments through the hybridization between the fe @xmath27 and the mo @xmath29 states what can be considered as a magnetic proximity effect . in this sense the double perovskites should be denoted ferromagnetic and not ferrimagnetic @xcite . despite the recent progress in understanding the physics of the double perovskites there is still an open debate on the adequate theoretical modelling . in particular , the details of the interplay between structural , electronic , and magnetic degrees of freedom in the double perovskites is not yet clearly understood . recently , the @xmath4 in double perovskites was discussed to depend sensitively on the band structure and band filling in contrast to experimental results @xcite . however , since in experiments often changes in the number of conduction electrons achieved by partially replacing the divalent alkaline earth ions on the @xmath3- and @xmath35-site by trivalent rare earth ions are accompanied by significant steric effects , it is difficult to distinguish between doping and structural effects . it is therefore desirable to investigate both the influence of structural changes ( changes in the bond length and bond angles ) and the effect of carrier doping . here , we present a study of the effect of structural changes and doping in the double perovskite @xmath0crwo@xmath1 ( @xmath36 ) on the magnetotransport , the magnetic and optical properties together with band structure calculations . we note that a major difference between the system @xmath0crwo@xmath1 and the systems @xmath0femoo@xmath1 or @xmath0fereo@xmath1 is the fact that in the former the majority spin band is only partially filled ( cr@xmath10 , @xmath37 , only the cr @xmath27 ( @xmath26 ) levels are occupied ) , whereas in the latter the majority spin band is completely filled ( fe@xmath10 , @xmath28 , both the fe @xmath27 ( @xmath26 ) and fe @xmath27 ( @xmath38 ) levels are occupied ) . we discuss that despite this difference the magnetism in @xmath0crwo@xmath1 and @xmath0femoo@xmath1 is similar . for both systems the exchange gap is important for the magnetic exchange and half metallicity . however , whereas for the former system also the crystal field gap between the @xmath26 and the @xmath25 levels plays a key role , for the latter the crystal field gap is irrelevant . a key result of our study is that optimum magnetic properties such as half metallicity and high @xmath4 can only be obtained in the undoped compounds close to the ideal undistorted perovskite structure that is characterized by a tolerance factor @xmath39 of the perovskite unit cell . furthermore , we find that electron doping tends to decrease @xmath4 in @xmath0crwo@xmath1 . in our study , we have used both polycrystalline and epitaxial thin film samples . polycrystalline samples were prepared from stoichiometric mixtures of srco@xmath40 , baco@xmath40 , caco@xmath40 , cr@xmath6o@xmath40 , la@xmath6o@xmath40 , and wo@xmath40 with a purity ranging between 99.99% and 99.999% . these powders were thoroughly mixed , placed in al@xmath6o@xmath40 crucibles , and then repeatedly heated in a thermobalance under reducing atmosphere ( h@xmath6/ar : 5%/95% ) with intermediate grinding . the final sintering temperatures @xmath41 were increased from 1200@xmath42c for the first up to 1550@xmath42c for the final firing for the first series of samples . in order to increase the curie temperature we prepared a second series of undoped sr@xmath6crwo@xmath1 samples with more sintering steps and longer final sintering times at lower ( 1300@xmath42c ) temperatures . for this series we find the highest curie temperatures , which correspond to the values reported in literature . this most likely is caused by a slightly higher oxygen content in these samples compared to those sintered at higher temperatures in the same reducing atmosphere . the use of the thermobalance allowed to monitor the ongoing reaction process due to the associated weight loss of the samples . the exact oxygen content of the investigated samples has not been determined . however , due to the very similar preparation procedure for the samples used for the study of electron doping , we can assume a similar oxygen content . the polycrystalline samples were characterized by x - ray powder diffractometry to detect parasitic phases , e.g. the insulating compound srwo@xmath43 . interestingly , even in samples containing srwo@xmath43 no chromium containing parasitic phases could be detected . the most simple explanation for this observation is the high vapor pressure of cr@xmath6o@xmath40 resulting in a loss of chromium . however , no weight loss has been found in the thermogravimetric measurement . therefore , it is more likely that the formation of the mixed phase @xmath0cr@xmath44w@xmath45o@xmath46 compensates for the missing sr and w due to the srwo@xmath43 impurity phase . while the double perovskite sr@xmath6femoo@xmath1 can be grown as single crystal @xcite , our attempts to grow single crystals of sr@xmath6crwo@xmath1 by floating zone melting failed so far due to the high vapor pressure of cr@xmath6o@xmath40 at the melting point . epitaxial thin films of sr@xmath6crwo@xmath1 were fabricated in a uhv laser molecular beam epitaxy ( l - mbe ) system @xcite . the details of the thin film deposition have been described recently @xcite . [ tem ] shows a transmission electron micrograph of an epitaxial sr@xmath6crwo@xmath1 thin film grown on a srtio@xmath40 ( 001 ) substrate . the image shows that the thin film grows epitaxially with high crystalline quality . the @xmath47-axis oriented films have been analyzed using a high resolution 4-circle x - ray diffractometer . only ( @xmath48 ) peaks could be detected and typically , rocking curves of the ( 004 ) peak had a full width at half maximum ( fwhm ) of only 0.02 to 0.03@xmath49 , which is very close to the fwhm of the substrate peak . moreover , an afm analysis of the sr@xmath6crwo@xmath1 films showed that they have a very small surface roughness of the order of one unit cell @xcite . in this study the thin films were used for optical measurements in order to determine the band structure . it is well known that the tolerance factor @xmath50 determines the crystal structure of perovskites @xmath51o@xmath40 @xcite . only for @xmath50 close to unity a cubic perovskite structure is obtained . for @xmath52 a tilt and rotation of the oxygen octahedra is obtained compensating for the misfit of the ionic radii of the involved @xmath3 and @xmath14 cations . hence , the deviation of the tolerance factor from the ideal value @xmath53 can be used as a measure for the internal strain in perovskites induced by the different radii of the @xmath3 and @xmath14 cations . this can be seen from the definition of @xmath50 given by @xmath54 here , @xmath55 denotes the average ionic radius for the ions on the @xmath14-site . for @xmath56 , the strain is compensated by a tilt and rotation of the oxygen octahedra . this results in a deviation of the @xmath14-o-@xmath14 bond angles from the ideal value of @xmath57 . for @xmath58 the connecting pattern of the oxygen octahedra is rhombohedral , whereas it is orthorhombic for lower values of @xmath50 @xcite . a simple consequence of the deviation of the bond angles from @xmath57 is a decrease in the hopping amplitude because the electron transfer between the @xmath14 sites is via o @xmath59-states . this , in turn , results in a decrease of the one - electron bandwidth @xmath60 . for @xmath61 , a hexagonal structure is expected which is classified by the stacking sequence of the @xmath14o@xmath1 octahedra @xcite . .overview on the crystallographic properties of the investigated samples of the @xmath0crwo@xmath1 series with @xmath2 . the tolerance @xmath50 factor was calculated using the spuds simulation software @xcite . [ cols="^,^,^,^ " , ] [ bigtable ] fig . [ fig : res ] shows the temperature dependence of the resistivity for the double perovskites @xmath0crwo@xmath1 with @xmath62 . all samples show an increase of the resistivity with decreasing temperature . since the investigated samples are polycrystalline , the influence of grain boundaries plays an important role . hence , the observed semiconductor like resistivity vs temperature curves may be related to the grain boundary resistance , whereas the intrinsic resistance of the double perovskites may be metallic . although the intrinsic resistivity behavior can not be unambiguously derived from our measurements , the fact that @xmath63 for @xmath64 ( see inset of fig . [ fig : res ] ) provides significant evidence for a metallic behavior in the sr@xmath6crwo@xmath1 sample . here , @xmath65 is the electrical conductivity . the observed trend that the resistivity increases , if sr@xmath6 is replaced by ca@xmath6 and even more by ba@xmath6 , can be understood as follows : in contrast to the cubic perovskite sr@xmath6crwo@xmath1 , ca@xmath6crwo@xmath1 has a distorted perovskite structure with a @xmath14-o-@xmath14 bonding angle deviating from 180@xmath42 . this result in a reduction of the overlap between the relevant orbitals and , hence , the hopping amplitude . finally , the ba@xmath6crwo@xmath1 compound has the highest resistivity most likely due to its hexagonal structure ( see table [ table : samples ] ) . the magnetotransport properties of sr@xmath6crwo@xmath1 already has been discussed elsewhere @xcite . we found , that polycrystalline samples containing a large number of grain boundaries show a large negative low - field magnetoresistance , @xmath66/r(0)$ ] , of up to -41% at 5k . at room temperature , this effect is reduced to a few percent . the large grain boundary @xmath67 effect at low temperatures indicates that sr@xmath6crwo@xmath1 has a large spin polarization of the charge carries and due to its high curie temperature may be an interesting candidate for magnetoelectronic devices operating room - temperature . in fig . [ fig : f ] we plot the curie temperature @xmath4 , the saturation magnetization @xmath68 , and the ionic radii versus the tolerance factor for the series @xmath0crwo@xmath1 . it is evident that the curie temperature is largest for sr@xmath6crwo@xmath1 ( @xmath5k ) , whereas it is suppressed strongly for ca@xmath6crwo@xmath1 ( @xmath69k ) . we attribute this fact to the small ionic radius of ca@xmath9 , which results in @xmath70 and , in turn , in a distorted perovskite structure . this results in a reduction of the effective hopping interactions between cr @xmath27 and w @xmath71 states , leading to a reduced spin splitting of the conduction band @xcite . this naturally reduces the magnetic coupling strength and hence the @xmath4 . the saturation magnetization @xmath68 is known to depend strongly on the amount @xmath72 of antisites , with @xmath73 for no antisites and @xmath74 for 50% antisites or complete disorder . by simply assuming the presence of antiferromagnetically coupled cr and w sublattices , a maximum saturation magnetization of @xmath75 is expected for @xmath73 , which is decreasing to zero for @xmath74 . that is , this assumption leads to @xcite @xmath76 where @xmath68 is the saturation magnetization in units of @xmath77 ; @xmath78 and @xmath79 are the magnetic moments of the cr@xmath10 and w@xmath21 ions in units of @xmath80 , respectively . comparing the experimental data to this simple model prediction leads to a surprisingly good agreement for sr@xmath6crwo@xmath1 . the measured value of @xmath81f.u . is very close to the value of @xmath82f.u . expected from eq.([eq : msat ] ) for @xmath83 . [ fig : f]b also shows that the saturation magnetization of ca@xmath6crwo@xmath1 is larger than for sr@xmath6crwo@xmath1 despite the much lower curie temperature of the latter . the obvious reason for that is the lower amount of antisites in ca@xmath6crwo@xmath1 . from the measured value of @xmath84 we expect @xmath85f.u . for ca@xmath6crwo@xmath1 which is slightly larger than the measured value of @xmath86f.u .. the reason for the observation that the experimental @xmath87 value is below the one predicted by eq.([eq : msat ] ) most likely is the distorted crystal structure of ca@xmath6crwo@xmath1 . the magnetic properties of the ba@xmath6crwo@xmath1 compound are completely different from those of sr@xmath6crwo@xmath1 and ca@xmath6crwo@xmath1 . here , the large ionic radius of ba@xmath9 enlarges @xmath50 well above unity . this causes a structural phase transition towards a hexagonal structure , where the ferromagnetic interaction is strongly suppressed . therefore , not only @xmath4 is strongly suppressed ( @xmath88k , in contrast to @xmath89k for sr@xmath6crwo@xmath1 ) , but also the saturation magnetization ( @xmath90f.u . ) is close to zero . this clearly indicates the strong effect of the structural phase transition on the magnetic interaction . we note that we can exclude the possibility that the very small @xmath68 value is due to strong disorder . evidently , there should be almost complete disorder to suppress @xmath68 close to zero . however , in the ba@xmath6crwo@xmath1 compound the two different @xmath14 sites establish a certain amount of order , since almost all w ions occupy the @xmath91 site . whether the observed behavior is related to a canted antiferromagnetic phase or simply to the presence of minority phases has to be clarified . [ fig : acwomag ] shows that the magnetic interactions for all three compounds @xmath0crwo@xmath1 is ferromagnetic . however , the strongly reduced saturation magnetization of ba@xmath6crwo@xmath1 and the upturn in magnetization vs. temperature curve at low temperatures indicates that ferromagnetic interactions are small and that there may be paramagnetic regions in the sample . the substitution of sr by ca evidently results in a strong reduction of @xmath4 , however , the ca@xmath6crwo@xmath1 samples are still clearly ferromagnetic . again , the important point is that only the compound with @xmath39 has optimum magnetic properties with respect to applications in magnetoelectronics . we also have performed measurements of the coercive field @xmath92 . at low temperature we obtained @xmath93oe for the sr@xmath6crwo@xmath1 compound and @xmath94oe for the ca@xmath6crwo@xmath1 compound . furthermore , for the sr@xmath6crwo@xmath1 the coercive field was found to depend on the preparation conditions . for example , lowering the final firing temperature from 1550@xmath42c to 1300@xmath42c was found to increase @xmath92 from 450 to 1200oe . in agreement with the findings for the doped manganites @xcite , the coercive field increases with decreasing @xmath50 . this observation can be easily understood , since the induced structural distortions can effectively act as pinning centers for domain wall movement . in those cases where a large remanent magnetization and/or coercive field are important for applications , the use of double perovskites with reduced values of @xmath50 obtained by suitable substitution on the @xmath3-site may be desirable . comparing our results for the series @xmath0crwo@xmath1 with @xmath95 to other double perovskite compounds summarized in table [ bigtable ] , it is evident that the suppression of @xmath4 as a function of the deviation of the tolerance factor from its ideal value of @xmath53 is a general trend : it is only weak for the series @xmath0femoo@xmath1 with @xmath36 , where @xmath4 varies between 310k and 420k @xcite . however , it is also strong for the series @xmath0crreo@xmath1 with @xmath96k for sr@xmath6crreo@xmath1 and @xmath97k for ca@xmath6crreo@xmath1 @xcite . in general , a high curie temperature can only be realized in double perovskites of the composition @xmath98o@xmath1 having a tolerance factor close to unity . this is realized in the different systems for @xmath99 . for @xmath50 well below unity , the curie temperature is drastically reduced in agreement with what is found for the doped manganites @xcite . for the double perovskites , the system ( sr@xmath100ca@xmath101)@xmath6fereo@xmath1 is an exception of the general rule @xcite : here , the ca@xmath6fereo@xmath1 compound has the highest @xmath4 , although the tolerance factor decreases continuously from @xmath102 for sr@xmath6fereo@xmath1 to @xmath103 for ca@xmath6fereo@xmath1 on substituting sr by ca . we note , however , that ca@xmath6fereo@xmath1 is a unique material as it is a ferromagnetic insulator and that there may be an other mechanism causing the high ordering temperature @xcite . in fig . [ fig : tctoleranceall ] we have plotted the curie temperatures of the double perovskite materials listed in table [ bigtable ] versus their tolerance factor omitting the ferromagnetic insulator ca@xmath6fereo@xmath1 . despite the significant spread of data that may be partially caused be different sample quality , it is evident that a maximum curie temperature is obtained for the systems having a tolerance factor of @xmath8 . these systems have a cubic / tetragonal symmetry . as shown in fig . [ fig : tctoleranceall ] , for @xmath56 there is a transition to orthorhombic structures for @xmath104 , whereas for @xmath105 there is a transition to a hexagonal structure at @xmath106 . however , according to data from literature this transition seems to be smeared out as indicated by the shaded area . we briefly summarize the measured structural , transport and magnetic properties and discuss the underlying physics . recently , sarma _ et al . _ proposed an interesting model explaining the origin of the strong antiferromagnetic coupling between fe and mo in sr@xmath6femoo@xmath1 in terms of an strong effective exchange enhancement at the mo site due to a fe @xmath27-mo @xmath29 hybridization @xcite . kanamori and terakura extended this idea to explain ferromagnetism in many other systems , where nonmagnetic elements positioned between high - spin @xmath27 elements contribute to the stabilization of ferromagnetic coupling between the @xmath27 elements @xcite . the essence of this model is summarized in fig . [ fig : models]a for the case of sr@xmath6femoo@xmath1 . without any hopping interactions , the fe@xmath10 @xmath23 configuration has a large exchange splitting of the @xmath27 level in the spin - up and spin - down states and there is also a crystal field splitting @xmath107 into the @xmath26 and the @xmath25 states ( see fig . [ fig : models]a ) . the exchange splitting of the mo@xmath21 @xmath24 configuration ( better the mo-@xmath29-o-@xmath59 hybridized states ) is vanishingly small , however , there is a large crystal field splitting ( the @xmath25 states are several ev above the @xmath26 states and not shown in fig . [ fig : models]a ) . the interesting physics occurs on switching on hopping interactions , which result in a finite coupling between states of the same symmetry and spin . the hopping interaction not only leads to an admixture of the fe @xmath27 to the mo @xmath29 states , but more importantly to a shift of the bare energy levels . as shown in fig . [ fig : models]a , the delocalized mo @xmath26 spin - up states are pushed up , whereas the mo @xmath26 spin - down states are pushed down . this causes a finite spin polarization at the fermi level ( actually 100% in fig . [ fig : models]a ) resulting from the hopping interactions . this kinetic energy driven mechanism leads to an antiferromagnetic coupling between the delocalized mo @xmath29 and the localized fe @xmath27 electrons , since the energy is lowered by populating the mo @xmath29 spin - down band @xcite . the magnitude of the spin polarization derived from this mechanism obviously is governed by the hopping strength and the charge transfer energy between the localized and the delocalized states @xcite . the question now arises , whether the above model also applies for the @xmath0crwo@xmath1 compounds . the key concept of the model is the energy gain contributed by the spin polarization of the nonmagnetic element ( now w ) induced by the hybridization with the magnetic transition metal ( now cr ) . it has been pointed out by sarma @xcite that the underlying mechanism will always be operative , whenever the conduction band is placed within the energy gap formed by the large exchange splitting of the localized electrons at the transition metal site . as shown in fig . [ fig : models]b , this is also the case for the @xmath0crwo@xmath1 compounds : the w @xmath71 band resides in between the exchange split cr @xmath27 ( @xmath26 ) spin - up and spin - down bands . the schematic band structure of fig . [ fig : models]b has been confirmed by band structure calculations presented below . the only difference between the system @xmath0crwo@xmath1 and the systems @xmath0femoo@xmath1 or @xmath0fereo@xmath1 is the fact that in the former the majority spin band is only partially full . for cr@xmath10 , ( @xmath108 ) only the cr @xmath26 levels are occupied . in contrast , for fe@xmath10 ( @xmath28 ) both the fe @xmath26 and fe @xmath38 levels are occupied , that is , the majority spin band is completely filled . band structure calculations show that the crystal field splitting in the cr compounds ( @xmath109ev ) is slightly larger than in the fe compounds @xcite . on the other hand , the exchange splitting in the cr @xmath27 bands is somewhat smaller than for the fe @xmath27 bands due to the valence configuration cr @xmath110 with less electrons and weaker hund s coupling . taking these facts into account we have to split up the cr @xmath27 spin - up and spin - down band into two separate @xmath27 ( @xmath26 ) and @xmath27 ( @xmath38 ) bands with the fermi level lying in the gap between the bands as shown in fig . [ fig : models]b . indeed band structure calculations ( see below ) show that the cr @xmath27 ( @xmath25 ) spin - up band is about 0.5ev above the fermi level . however , the above mechanism still works as long as the w @xmath71 ( @xmath26 ) band is placed within the energy gap between the cr @xmath27 ( @xmath26 ) spin - up and the cr @xmath27 ( @xmath26 ) spin - down band . then , again the w @xmath71 ( @xmath26 ) levels would hybridize with the cr @xmath27 ( @xmath26 ) levels resulting in a negative spin polarization of the nonmagnetic element w @xmath71 ( @xmath26 ) and a stabilization of ferromagnetism and half - metallic behavior . we note that no hybridization takes place between the cr @xmath27 ( @xmath25 ) spin - up band and the w @xmath71 ( @xmath26 ) spin - up band due to the different symmetry of these levels . therefore , the exact position of the cr @xmath27 ( @xmath25 ) spin - up band is not relevant . we also note that due to the large crystal field splitting for the w @xmath71 band , the w @xmath27 ( @xmath25 ) band is several ev above the w @xmath27 ( @xmath26 ) band and not shown in fig . [ fig : models]b . summarizing our discussion we can state that the essential physical mechanism leading to ferromagnetism is very similar for the @xmath0crwo@xmath1 and the @xmath0femoo@xmath1 compounds . comparing fig . [ fig : models]a and b we see that the only difference is an upward shift of the @xmath27 bands for the @xmath0crwo@xmath1 compounds . we now discuss the experimental results in context with the models discussed above . we first discuss the strong dependence of @xmath4 on the tolerance factor , which in turn is intimately related to the bond angles . it is well known that in perovskite type transition metal oxides in general the increase of the @xmath14-o bond length and the deviation of the @xmath14-o-@xmath14 bond angle from @xmath57 has the effect of a reducing the hybridization matrix element @xmath111 . this is caused by the reduction of the overlap of the oxygen 2@xmath112 and transition - metal @xmath113 states @xcite , which in turn results in a reduction of the bare electron bandwidth @xmath60 . since the weakening of the hybridization causes a reduction of the energy gain stabilizing ferromagnetism , we expect a significant decrease of @xmath4 with increasing deviation of the tolerance factor from @xmath53 , or equivalently with an increasing deviation of the @xmath14-o-@xmath14 bond angle from @xmath57 . this is in good qualitative agreement with our results on the @xmath0crwo@xmath1 compound ( see fig . [ fig : f]a ) and the collected data plotted in fig . [ fig : tctoleranceall ] . theoretical models providing a quantitative explanation have still to be developed . it is known that for a @xmath57 bond angle , @xmath111 decreases with increasing bond length @xmath114 roughly as @xmath115 . therefore , the effective hopping amplitude between the transition metal ions @xmath14 has an even stronger dependence on @xmath114 . hence , a significant increase of @xmath111 and , in turn , @xmath4 is expected with a reduction of the bond length . applying this consideration to the investigated series @xmath0crwo@xmath1 we would expect the largest @xmath4 for the ca@xmath6crwo@xmath1 compound due to its smallest cell volume and , hence , shortest bond length . however , experimentally the largest @xmath4 was found for sr@xmath6crwo@xmath1 . this is caused by the fact that @xmath111 not only depends on the bond length but also strongly on the bond angle . for the cubic double perovskite sr@xmath6crwo@xmath1 with @xmath39 the bond angle has the ideal value of @xmath57 , whereas for the distorted double perovskite ca@xmath6crwo@xmath1 with @xmath56 the bond angle significantly deviates from this ideal value . for the double perovskites lattice effects come into play also for large tolerance factors @xmath116 , where in the hexagonal structure the formation of dimers suppresses strongly ferromagnetism and enhances antiferromagnetic interactions . furthermore , the magnetic interactions are also weakened due to the increased distance ( bond length ) between the magnetic ions . we finally would like to mention that in a recent work on @xmath0femoo@xmath1 , a linear correlation , @xmath117 , between @xmath4 and the bare one - electron bandwidth @xmath118 has been found by ritter _ et al . _ @xcite . summarizing our discussion we can state that our findings for the series @xmath0crwo@xmath1 can be extended to the double perovskites in general . the data summarized in table [ bigtable ] and fig . [ fig : tctoleranceall ] clearly indicated that for most double perovskites a maximum @xmath4 is obtained for a tolerance factor of @xmath8 corresponding to an about cubic perovskite structure with a bond angle close to @xmath57 . this optimum situation in most cases is realized in the sr@xmath119o@xmath1 compounds . the requirement @xmath39 for optimum @xmath4 in the double perovskite is different for the doped manganites . here , a maximum @xmath4 is achieved for compounds with @xmath120 , that is , for a significantly distorted perovskite structure . et al . _ @xcite have shown that the highest @xmath4 in doped manganites is obtained for @xmath121 ( in a more precise analysis by zhou _ et al . _ @xcite slightly larger values for @xmath50 have been derived based on a coordination number of 9 ) . for a tolerance factor below the optimum value , a dramatic decrease of @xmath4 has been found . whether or not the differences between the manganites and the double perovskites are related to the different mechanism stabilizing ferromagnetism in these compounds has to be clarified . we note , however , that the different behavior of the manganites and the double perovskites is likely to be related to different mechanisms . in manganites the polarization of the conduction band is driven by the hund s coupling ( intra atomic ) between the @xmath26 and the @xmath25 states , which is not sensitive to the details of the band structure . in contrast , in the double perovskites the spin polarization of the conduction band is itself dependent on the band structure and therefore is more intimately linked to it . the effect of carrier ( electron ) doping in sr@xmath6crwo@xmath1 was studied in a series of sr@xmath19la@xmath20crwo@xmath1 samples with @xmath123 , @xmath124 , @xmath125 and @xmath126 . in our experiments trivalent la@xmath10 is chosen to replace the sr@xmath9 ions because the ionic radius of la@xmath10 ( @xmath127 ) is similar to that of sr@xmath9 ( @xmath128 ) . as a result , there are only a small variations of the lattice parameters and the tolerance factor @xmath50 on changing the doping level from @xmath123 to @xmath129 . x - ray analysis showed that the lattice parameter slightly decreases from @xmath130 ( @xmath123 ) to @xmath131 ( @xmath129 ) as expected since the ionic radius of la@xmath10 is smaller than that of sr@xmath9 . however , the crystal structure of all samples remained cubic . this means that the structural changes are small on varying the doping level . in this way the effect of doping can be studied without being strongly influenced by structural effects . in fig . [ fig : la ] the magnetization curves of the series sr@xmath19la@xmath20crwo@xmath1 are shown for @xmath132k . all samples of this series have been fired between 1530 and @xmath133c in reducing atmosphere . for the undoped sample this results in a smaller curie temperature of about 390k as compared to about 460k for the samples fired at @xmath134c . the reason for that most likely is a slightly smaller oxygen content in the sample fired at a higher temperature . since a reduced oxygen content corresponds to an effective electron doping , for the doping series we compare only samples prepared under identical conditions . we find that the saturation magnetization @xmath68 decreases from @xmath135/f.u . for @xmath123 to @xmath136/f.u . for @xmath129 . as shown in the inset of fig . [ fig : la ] , with increasing doping level also the amount of antisites is increasing . at present , we only have a plausible explanation for the observed increase of the amount of antisites with increasing doping level , which so far could not be unambiguously proven . however , since la doping is expected to result in a reduction of the differences in the valence states of cr and w , it evidently results in a reduction of the differences in the ionic radii of cr and w ( @xmath137 and @xmath138 ) . therefore , increasing the doping level results in more similar ionic radii of cr and w paving the way for the creation of cr / w antisites . since the substitution of sr@xmath9 by la@xmath10 results both in electron doping and an increase of antisites , it is not possible to unambiguously attribute the measured decrease in @xmath68 to either the increasing doping level or the increase of antisites alone . as will be discussed in the following , for the series sr@xmath19la@xmath20crwo@xmath1 the measured reduction of @xmath68 most likely is caused by both electron doping and disorder . we first discuss the expected reduction of @xmath68 due to the increasing amount of antisites . we note that that several authors have found a reduction of @xmath68 following the increase of antisites @xcite . this behavior is consistent with simple monte carlo simulation studies @xcite . however , also more complicated models @xcite predict a reduction of @xmath68 with increasing amount of antisites due to charge transfer effects @xcite . in a first approach we can analyze our data using eq.([eq : msat ] ) . with the measured @xmath139 values for the amount of antisites we can calculate @xmath140 . we find that the calculated @xmath68 values are significantly larger than the measured ones . this suggests that the observed reduction of the saturation magnetization can not be explained by the increasing amount of antisites alone ( at least within the simple model yielding eq.([eq : msat ] ) ) . however , we also have to keep in mind that electron doping itself contributes to the reduction of @xmath68 . according to the illustration given in fig . [ fig : models]a it is evident that electron doping in the @xmath0femoo@xmath1 system increases the spin - down magnetic moment at the mo site , indicating that the electrons are filled into the mo @xmath141 band and thereby reduce @xmath68 @xcite . according to fig . [ fig : models]b , the same mechanism holds for the sr@xmath19la@xmath20crwo@xmath1 system . that is , our results suggest that the observed reduction of the saturation magnetization with increasing la@xmath10 substitution is caused both by an increase of the amount of antisites and an increase of the number of conduction electrons . summarizing our discussion of the saturation magnetization we would like to emphasize that a variation of the doping level in most cases is correlated with a variation of the amount of disorder , since doping contributes to a reduction of the difference of the valence states of cr and w ( fe and mo ) what in turn results in an increasing amount of antisites @xcite unfortunately , due to this fact the experimental situation is not completely clear . while in @xcite for sr@xmath6femoo@xmath1 and also in our study for sr@xmath6crwo@xmath1 la doping is clearly correlated with higher disorder , in @xcite the amount of antisites seems to be constant in sr@xmath19la@xmath20femoo@xmath1 for almost the whole doping series from @xmath123 to @xmath142 . for the system sr@xmath6crwo@xmath1 , the suppression of @xmath68 is stronger than for the system sr@xmath6femoo@xmath1 , probably due to the fact that the cr and w ions are easier to disorder because of their similar ionic radii . we finally note that in general both increasing doping and disorder can destroy the underlying half - metallic ferromagnetic state in double perovskites , thereby significantly populating / depopulating the different spin channels leading to a sharp decrease in @xmath68 @xcite . we next discuss the influence of la doping on the curie temperature . it is evident that an enhancement of @xmath4 by doping would be of great importance for possible applications of the double perovskites in spintronic devices . however , both the experimental data on the variation of @xmath4 with electron doping as well as the theoretical interpretation is controversial at present . in single crystals of sr@xmath19la@xmath20femoo@xmath1 moritomo _ et al . _ have found that @xmath4 does not change as a function of la doping @xmath143 @xcite . in contrast , navarro _ _ have reported a considerable increase of @xmath4 of about 70k in ceramic samples of sr@xmath19la@xmath20femoo@xmath1 investigating a wider doping range @xmath144 @xcite . in our study on the system sr@xmath19la@xmath20crwo@xmath1 we have found a reduction of @xmath4 of about 80k in the doping range @xmath145 as can be seen in fig . [ fig : la - tc ] . we note that part of the experimental discrepancies may be related to large error bars in the determination of @xmath4 . in particular , great care has to be taken over the determination of @xmath4 , since the transition from zero magnetization to finite magnetization is smeared out considerably due to finite applied magnetic fields and parasitic phases . it is evident from fig . [ fig : la - tc ] that in all our samples there are minor phases with optimum @xmath4 close to about 400k . however , it is also evident that la doping reduces @xmath4 of the major part of the sample . that is , different experimental results on @xmath4 may be in part related to different ways of measuring and analyzing the data . from the theoretical point of view one expects both an increase and decrease of @xmath4 with increasing doping . on the one hand , within the model presented in @xcite @xmath4 is expected to be rather reduced than enhanced by electron doping due to the fact that the possible energy gain by shifting electrons from spin - up band into the spin - down band is reduced @xcite . this , in turn , reduces the stability of the ferromagnetic phase in agreement with our results . however , the role of the increasing amount of antisites with increasing doping still has to be clarified . on the other hand , in a double exchange model the increase of the number of conduction electrons promoted by la doping is expected to enhance the double exchange interaction leading to an increase of @xmath4 . within this model the ferromagnetic interaction arises from the double exchange interaction between the localized moments on cr@xmath10 sites ( @xmath110 , @xmath146 ) mediated by itinerant electron provided by the w@xmath21 ions ( @xmath22 , @xmath147 ) . according to the band structure calculation presented below , the cr @xmath26 spin - up subband is completely filled and it is the electron in the @xmath26 spin - down subband of cr and w which mediates the double exchange interaction . in general , an increasing number of electrons in the spin - down subband is expected to strengthen the double exchange interaction leading to an increase of @xmath4 in conflict with our experimental findings . that is , the observed decrease of @xmath4 with increasing doping level seems to support the model presented in @xcite ( see also fig . [ fig : models ] ) . however , we have to take into account that by la doping besides the number of electrons we also increase the amount of disorder . the latter may weaken the double exchange interaction sufficiently to result in an effective decrease of @xmath4 also within a double exchange based model . further work is required to clarify this in more detail . the half - metallicity of ferromagnetic materials is one of the important ingredients required for applications . in order to obtain a conclusive picture regarding the spin polarization at the fermi level it is important to compare band structure calculations with experimental results . here , we compare the results of band structure calculations based on _ ab initio _ methods to experimental results on the optical reflectivity and transmissivity . in fig . [ fig : bandstructure ] the density of states of sr@xmath6crwo@xmath1 is plotted versus energy . the band structure has been calculated @xcite within the linear muffin - tin orbital ( lmto ) method using the atomic sphere approximation ( asa ) . a detailed discussion of this method can be found elsewhere @xcite . as a key result of the calculation we obtain a gap of about @xmath148ev in the spin - up band around the fermi level ( compare also fig . [ fig : models]b ) . this gap corresponds to gap between the crystal field split cr @xmath27 ( @xmath26 ) and cr @xmath27 ( @xmath38 ) subband . on the other hand , the spin - down states are available at the fermi - level . the broad band corresponds to the hybridized cr @xmath27 ( @xmath26 ) and w @xmath71 ( @xmath26 ) levels . that is , according to the band structure calculation we expect a half - metallic behavior for the double perovskite sr@xmath6crwo@xmath1 . a similar result but with slightly larger band gaps has been reported recently by jeng _ the result of the band structure calculation is in agreement with the model assumptions made above . the cr @xmath27 ( @xmath26 ) spin - up subband is completely filled and there is a gap to the cr @xmath27 ( @xmath38 ) spin - up subband lying about 0.5ev above the fermi level . the w @xmath71 ( @xmath26 ) spin - up band also is mainly above the fermi level . the hybridized cr @xmath27 ( @xmath26 ) spin - down states and the w @xmath71 ( @xmath26 ) spin - down states form a broad spin - down band with the fermi level lying in this band . this is fully consistent with the model presented in fig . [ fig : models ] . furthermore , the band structure calculation shows that la doping is expected to add electrons in the spin - down band . in order to verify the band structure calculations experimentally , we have performed optical reflection and transmission measurements of sr@xmath6crwo@xmath1 thin films with photon energies from 0.38ev to 7ev . the srtio@xmath40 substrates , on which the epitaxial quality of the films is high ( see fig . [ tem ] ) , are unfortunately transparent only in the range of photon energies of 0.20 - 3.2ev . therefore , we have also investigated strained epitaxial sr@xmath6crwo@xmath1 films on laalo@xmath40 substrates , which are transparent from 0.17 to 5.5ev , and polycrystalline sr@xmath6crwo@xmath1 films on mgal@xmath6o@xmath40 substrates , which are transparent from to 0.22 - 6.5ev , however have a larger lattice mismatch to sr@xmath6crwo@xmath1 . on each substrate , films with thicknesses of @xmath149 , 80 , and 320 nm were investigated . almost identical optical absorption spectra were obtained for the epitaxial films , suggesting that the observed optical features are indeed related to the sr@xmath6crwo@xmath1 films , as summarized in fig . [ fig : absorption ] . the dominant optical features , which are an absorption shoulder around 1ev , and a strong increase of the absorption above 4ev , are also found for the polycrystalline samples grown on mgal@xmath6o@xmath43 . the transmission @xmath150 at 4.6ev is around 0.1% for the films with @xmath151 nm , and around 30% for the films with @xmath149 nm . the error in the calculation of the absorption coefficient @xmath152/d $ ] with the reflection @xmath153 comes mostly from uncertainties in @xmath154 , which is sufficiently small for the 320 nm films at lower energies , and for the 30 nm films at higher energies . for the films with @xmath151 nm , reflectivity oscillations with a period @xmath155ev were observed at photon energies of 1.5 - 3.5ev , indicating a refractory index @xmath156 . this is consistent with the measured reflection data and @xmath157 in this energy range . as determined from fourier transform infrared ( ftir ) transmission measurements , the films remain transparent down to 0.2ev , however , @xmath153 is increasing significantly towards lower photon energies . the optical measurements agree fairly well with the band structure calculations . the increase of the absorption coefficient of sr@xmath6crwo@xmath1 above 4ev can most probably be attributed to a charge transfer transition between the p - like spin - up and -down oxygen bands at @xmath158ev below the fermi - level into the oxygen / metal bands at + 1ev above the fermi - level . the absorption shoulder around 1ev coincides roughly with the energy gap at the fermi level in the spin - up band , and is therefore probably caused by transitions from the cr spin - up @xmath26 states below the fermi - level into the oxygen / metal bands around + 1ev above the fermi - level . unfortunately , due to the substrate absorption below 0.2ev , no optical information is available in the infrared region which would allow a more definitive statement with respect to the density of states at the fermi level and the half - metallic character of sr@xmath6crwo@xmath1 . however , from the refractory index @xmath159 , one can estimate an optical conductivity of @xmath160@xmath161 at 1ev . comparing this result with recent measurements of sr@xmath6crreo@xmath1 @xcite , one can classify sr@xmath6crwo@xmath1 as a ( very ) bad half metal . this is also in agreement with recent transport measurements @xcite . we note , however , that the transport data may be ambiguous , since the conductivity of the sr@xmath6crwo@xmath1 thin film is comparable to that of the substrate . this is for example the case for srtio@xmath40 substrates , which obtain a conductive surface layer at the reducing atmosphere of the thin film deposition process @xcite . that is , more conclusive transport data are required to settle this issue . we have performed a detailed analysis of the structural , transport , magnetic and optical properties of of the double perovskite @xmath0crwo@xmath1 with @xmath2 . in agreement with band structure calculations the double perovskite sr@xmath6crwo@xmath1 is a half - metal with a high curie - temperature above 450k . the measured saturation magnetization of @xmath162f.u . is well below the optimum value of @xmath163f.u . due to 23% of cr / w antisites . the large amount of antisites most likely is caused by the similar ionic radii of the cr@xmath10 and w@xmath21 ions resulting in a low threshold for the formation of antisites . the substitution of sr by ba and ca in the double perovskite system @xmath0crwo@xmath1 showed that the maximum curie temperature is obtained for the compound with a tolerance factor close to one . this is in agreement with a large variety of data reported in literature on other double perovskite systems @xmath98o@xmath1 . the fact that the maximum @xmath4 is obtained for a tolerance factor close to one differs from the behavior of the doped manganites where a maximum @xmath4 was found for a tolerance factor of about 0.93 . the observation that a tolerance factor @xmath39 yields the highest @xmath4 has been explained within a model , where ferromagnetism is stabilized by the energy gain contributed by the negative spin polarization on the nonmagnetic w - site due to hybridization of the cr @xmath27 ( @xmath26 ) and the w @xmath71 ( @xmath26 ) states . this hybridization is weakened by bond angles deviating from @xmath57 or equivalently a tolerance factor deviating from unity . electron doping of sr@xmath6crwo@xmath1 by partial substitution of sr@xmath9 by la@xmath10 was found to decrease both the curie temperature and the saturation magnetization . the decrease in the saturation magnetization was found to be caused both by an increase in the amount of antisites and by increasing band filling . although the decrease of @xmath4 with increasing doping can be explained qualitatively within the same model , it seems to be in conflict with double exchange type models predicting an increase of @xmath4 with increasing band filling . however , an unambiguous conclusion can not be drawn at present . the reason for that is the fact that on increasing the doping level one also obtains an increasing amount of disorder . in order to clarify this issue it is required to study systems allowing for the variation of the doping level without changing disorder . this work was supported by the deutsche forschungsgemeinschaft and the bundesministerium fr bildung und forschung ( project 13n8279 ) . the authors acknowledge fruitful discussions with m. s. ramachandra rao . r. gross , j. klein , b. wiedenhorst , c. hfener , u. schoop , j. b. philipp , m. schonecke , f. herbstritt , l. alff , yafeng lu , a. marx , s. schymon , s. thienhaus , w. mader , in _ superconducting and related oxides : physics and nanoengineering iv _ , d. pavuna and i. bosovic eds . , spie conf . proc . * vol . 4058 * ( 2000 ) , pp . 278294 . for a review see : j. b. goodenogh , j. m. longo , in : k .- h . hellwege , o. madelung ( eds . ) , _ magnetic and other properties of oxides and related compounds _ , landolt - brnstein , new series , group iii , vol . * 4 * , springer , berlin , 1970 .	the structural , transport , magnetic and optical properties of the double perovskite @xmath0crwo@xmath1 with @xmath2 have been studied . by varying the alkaline earth ion on the @xmath3 site , the influence of steric effects on the curie temperature @xmath4 and the saturation magnetization has been determined . a maximum @xmath5k was found for sr@xmath6crwo@xmath1 having an almost undistorted perovskite structure with a tolerance factor @xmath7 . for ca@xmath6crwo@xmath1 and ba@xmath6crwo@xmath1 structural changes result in a strong reduction of @xmath4 . our study strongly suggests that for the double perovskites in general an optimum @xmath4 is achieved only for @xmath8 , that is , for an undistorted perovskite structure . electron doping in sr@xmath6crwo@xmath1 by a partial substitution of sr@xmath9 by la@xmath10 was found to reduce both @xmath4 and the saturation magnetization @xmath11 . the reduction of @xmath11 could be attributed both to band structure effects and the cr / w antisites induced by doping . band structure calculations for sr@xmath6crwo@xmath1 predict an energy gap in the spin - up band , but a finite density of states for the spin - down band . the predictions of the band structure calculation are consistent with our optical measurements . our experimental results support the presence of a kinetic energy driven mechanism in @xmath0crwo@xmath1 , where ferromagnetism is stabilized by a hybridization of states of the nonmagnetic w - site positioned in between the high spin cr - sites .
production of the doubly strange baryon @xmath0 has been studied in heavy ion collisions at various energies . for ultrarelativistic energies at sps , the measured @xmath0 abundance is significantly enhanced compared to that expected from initial nucleon - nucleon collisions @xcite . explanations for this enhancement include exotic mechanisms due to the formation of the quark - gluon plasma , topological defects , and color ropes as well as more conventional processes of strangeness - exchange reactions such as @xmath9 and @xmath2 spal01 . in the latter case , the experimental data can , however , only be explained if a relatively high dense hadronic matter exists in these collisions . observations of @xmath0 production at lower energies at the ags , e.g. , au+au collisions at 6 agev @xcite , have also been reported recently , and the measured yield is comparable to the predictions from the transport model based on strangeness - exchange reactions @xcite . again , the hadronic density reached in these collisions is quite large with maximum density reaching 7 - 8 times normal nuclear matter density . in all these cases , the @xmath0 yield seems to be consistent with that given by the statistical model @xcite . although both @xmath10 and @xmath11 are infrequently produced in heavy ion collisions at lower gsi energies , i.e. , about 1 - 2 agev , which are below the threshold for producing these particles from nucleon - nucleon collisions in free space , studies of their production have provided valuable information about the nuclear equation of state at high densities as well as the in - medium properties of hadrons ko96,cassing99 . the latter is believed to be related to the partial restoration of chiral symmetry in hot dense matter . in the present study , @xmath12 production in heavy - ion collisions at energies , which are below the threshold of about @xmath13 for its production in nucleon - nucleon collisions in free space is studied in the framework of a relativistic transport model . we include @xmath0 production from the strangeness - exchange reactions @xmath14 and @xmath15 with their cross sections taken from the predictions of a coupled - channel approach based on a flavor su(3)-invariant hadronic lagrangian @xcite . the @xmath0 yield from these collisions thus depends on the cross sections for these reactions as well as the abundance of @xmath10 , @xmath16 and @xmath17 in the collisions . medium effects due to modified kaon properties are also studied as it is known to affect the production of @xmath10 , @xmath16 , and @xmath18 in dense matter @xcite . we find that the final @xmath12 yield is not much affected by the medium effects but is sensitive to the cross sections for the strangeness - exchange reactions , thus offering the possibility of testing the predicted cross sections from the hadronic model . the transport model used in present study is taken from that of refs . gqli97,gqli98 based on the relativistic vlasov - uehling - uhlenbeck equation ( rvuu ) @xcite . in this model , only the nucleon , delta resonance , and pion are treated explicitly . besides undergoing elastic and inelastic two - body scatterings , these particles also propagate in mean - field potentials . for nucleons , their potential is taken from the effective chiral lagrangian of ref.@xcite instead of the nonlinear walecka model used in ref.@xcite . as a result , their motions are given by the hamilton equation of motion , @xmath19where @xmath20 with @xmath21 and @xmath22 , and @xmath23 and @xmath24 are the scalar and vector mean - field potentials . for mean - field parameters , we use the set that gives a nuclear matter incompressibility @xmath25 and a nucleon effective mass @xmath26 at normal nuclear matter density @xmath27 @xmath28 . the @xmath29 resonance is assumed to have the same mean - field potential as the nucleon , while the mean - field potential for the pion is neglected . kaons together with its partners ( hyperons or antikaons ) are produced in this model from pion - baryon and baryon - baryon reactions , i.e. , @xmath30 and @xmath31 . in refs.@xcite , the cross section for the reaction @xmath32 is taken from the resonance model of ref.@xcite . here , we use that from an improved resonance model @xcite . the cross section for the reaction @xmath33 is , on the other hand , taken from the one - boson - exchange model of refs . antikaons in this model are produced not only from pion - baryon and baryon - baryon reactions , i.e. , @xmath34 and @xmath35 , but also from the pion - hyperon reactions @xmath36 @xcite , where @xmath37 denotes either @xmath38 or @xmath39 . their cross sections are taken either from predictions of the boson - exchange model or from the empirical values as in ref . annihilation of produced antikaons is included via the inverse reactions of pion - hyperon reactions , i.e. , @xmath40 , as other absorption reactions involve the rarely produced kaons and are thus unimportant . however , the annihilation of kaons is neglected as it has little effect on kaon production @xcite . because of the small production probabilities of kaons , hyperons , and antikaons in heavy - ion collisions at energies , the above discussed reactions are treated perturbatively . in this method , the dynamics of nucleons , @xmath29 resonances , and pions that produce these particles are not affected , and the produced particle is given a probability determined by the ratio of its production cross section to the total two - body scattering cross section . after being produced , these rare particles also undergo elastic and inelastic scattering as well as propagate in mean - field potentials . for @xmath41 and @xmath39 , their mean - field potentials are taken to be @xmath42 of the nucleon potential according to their light quark content . for kaon and antikaon , their mean - field potentials are obtained from the chiral lagrange including both scalar and vector interactions @xcite , i.e. , @xmath43 ^{1/2}\pm b_{k}\rho .\ ] ] in the above , we have @xmath44 @xmath45 corresponding to a pion decay constant @xmath46 mev ; @xmath47 and @xmath48 are , respectively , the nuclear matter and scalar density ; and the `` @xmath49 '' and `` @xmath50 '' signs in the last term are for kaon and antikaon , respectively . the two parameters @xmath51 and @xmath52 , which determine the strength of the attractive scalar potential for kaon and antikaon , respectively , can in principle be evaluated from the chiral lagrangian @xcite . here , we follow the method of ref.@xcite by determining their values from fitting the experimental data on kaon and antikaon yield in heavy - ion collisions using the rvuu model , and this will be described in detail later . the perturbative method for rare particle production in heavy ion collisions was first introduced in ref.@xcite and is now extensively used in transport models @xcite . since the yield of @xmath0 particle is expected to be small at sis energies , it is treated perturbatively as well . the produced @xmath0 from the strangeness - exchange reactions @xmath53 is thus given a probability equal to the ratio of this cross section to the @xmath54 total cross section , which is taken to be @xmath55 , multiplied by the probabilities carried by @xmath10 and @xmath37 . since the final @xmath0 yield is given by the product of its production probability in the reaction @xmath56 and the number of @xmath54 collisions , which is proportional to the @xmath54 total cross section , it depends only on the cross section for the reaction @xmath57 but not on the @xmath54 total cross section . during its propagation in the nuclear medium , the @xmath0 is subjected to a mean - field potential that is @xmath58 of the nucleon potential as it has only one light quark . , ( b ) @xmath61 , ( c ) @xmath62 , and ( d ) @xmath63 . solid curves are from the coupled - channel calculations of ref . @xcite with a cutoff parameter of 1 gev , and dashed curves are parameterized ones based on eq.([param]).,width=307,height=288 ] for the strange - exchange reactions @xmath56 , there is no empirical information on their cross sections . these reactions have , however , been studied in the coupled - channel approach based on a gauged flavor -invariant hadronic lagrangian @xcite . the spin- and isospin - averaged cross sections for these reactions can be written as @xmath64where @xmath65 and @xmath66 are initial antikaon and final pion momenta in the center - of - mass system , and @xmath67 and @xmath68 are squared invariant matrix elements with summation over the spins and isospins of both initial and final particles . the results from the coupled - channel approach with cutoff parameter of @xmath7 gev can be fitted by using the following parameterization for the squared invariant matrix elements : @xmath69where the threshold energy @xmath70 in the center - of mass system is @xmath71 and @xmath72 for the reactions @xmath9 and @xmath2 , respectively . in figs . [ crsc](a ) and ( b ) , we show by solid curves the cross sections for the two reactions @xmath73 and @xmath2 from the coupled - channel approach and by dashed curves those using the parameterizations of eq.([param ] ) . except near threshold , these cross sections are of the order of @xmath74-@xmath75 . the cross sections for the inverse reactions @xmath76 and @xmath77 , which are needed for treating @xmath12 annihilation , are related to those for @xmath12 production by the principle of detailed balancing , i.e. , @xmath78 these cross sections are shown in figs . [ crsc](c ) and ( d ) by solid curves for those from the coupled - channel approach and by dashed curves for the parameterized ones . and ( b ) @xmath79 with ( solid lines ) and without ( dashed lines ) kaon medium effects in near - central collisions of @xmath5ni + @xmath5ni at @xmath6 gev . the experimental data are from the fopi @xcite and kaos @xcite collaborations.,width=288,height=288 ] since the effective masses of kaons and antikaons are modified in nuclear medium , their production cross sections differ from those in free space . following refs.@xcite , their values at center - of - mass energy @xmath80-@xmath81 above the threshold energy @xmath81 in the medium are taken to have same values as those at center - of - mass energy @xmath82-@xmath83 above the threshold energy @xmath84 in free space . taking into account also propagation of kaons and antikaons in their mean - field potentials given by eq.([kaon ] ) , we have studied kaon and antikaon production in near - central collisions of @xmath4ni + @xmath5ni at @xmath6 . we find that the measured rapidity distributions of kaons and antikaons @xcite , shown by squares in figs . [ dndykpkm](a ) and ( b ) , respectively , can be fitted with the parameters @xmath85 @xmath86@xmath28 and @xmath87 @xmath86@xmath28 . these values are somewhat smaller than those used in ref . @xcite since slightly different cross sections are used in the present study as discussed in the previous section . the above parameters give a repulsive potential of about @xmath55 for kaons and an attractive potential of about @xmath88 for anitkaons with zero momentum in a nuclear matter at normal density . this value of kaon potential is consistent with that known empirically from the kaon - nucleus scattering @xcite and also from a more recent study of soft kaon production in p+a reactions at cosy - jlich @xcite . the magnitude of the antikaon potential is smaller than that extracted from the kaonic atom data @xcite , which is about @xmath89 but depends strongly on the extrapolation procedure from the surface of nuclei to their interiors @xcite . it is , however , comparable to the moderate attractive potential with magnitude of about @xmath90-@xmath91 predicted by recent studies based on self - consistent calculations with a chiral lagrangian @xcite or meson - exchange potentials tolos01 . the relatively shallow antikaon - nucleus potential has also been found to reproduce reasonably the experimental data on kaonic atoms in studies based on the su(3 ) chiral unitary model @xcite . the present values of kaon and antikaon potentials determined from fitting the experimental data on kaon and antikaon yield in heavy - ion collisions using the rvuu model are thus reasonable . in fig . [ dndykpkm ] , we also show the results from the rvuu model without kaon medium effects . it is seen that the repulsive kaon potential reduces kaon production due to the higher threshold while the attractive antikaon potential enhances antikaon production as a result of lower threshold . including medium effects on kaon and antikaon production in heavy ion collisions are thus important in describing the measured kaon and antikaon yields . , @xmath92 ; ( b ) abundances of @xmath93 , @xmath41 , @xmath39 , and @xmath10 ; and ( c ) the @xmath12 abundance in cases with and without kaon medium effects and different cross sections ( see text ) in collisions of @xmath5ni + @xmath5ni at @xmath6 gev at @xmath94 fm.,width=307,height=403 ] we first show in fig . [ abndtime ] the time evolutions of ( a ) central baryon density and the abundances of @xmath95 , @xmath92 , and ( b ) @xmath93 , @xmath96 , @xmath39 , @xmath10 in collisions of @xmath5ni + @xmath5ni at energy @xmath97 and at impact parameter @xmath94 . it is seen that the colliding system reaches the highest compression of central baryon density of about @xmath98 at about @xmath99 , and most particles are produced during this compression stage . the final @xmath100 multiplicity is about @xmath101 , corresponding to an averaged charged pion multiplicity @xmath102 of about @xmath103 , which is in good agreement with the experimental data @xcite . the abundances of @xmath93 , @xmath41 , and @xmath39 reach their respective final values of @xmath104 , @xmath105 , and @xmath106 quickly at about @xmath75 , while the @xmath10 abundance first increases and then decreases to its final value of about @xmath107 at around @xmath108 . the decrease of the @xmath109 abundance is due to the strong @xmath10 absorption through the strangeness - exchange reactions @xmath110 . the final @xmath111 ratio is about @xmath112 and is again in good agreement with experimental data and is also comparable to the predictions from the statistical model @xcite . the time evolution of the @xmath0 abundance from this reaction is shown by the solid curve in fig . [ abndtime](c ) . the @xmath0 abundance is seen to reach its final value of about @xmath113 at about @xmath55 . the resulting @xmath114 ratio is thus about @xmath115 and is measurable in experiments at gsi . to see how the @xmath0 abundance depends on the cross sections for the strangeness - exchange reactions @xmath116 and @xmath117 , we also show in fig . [ abndtime ] ( c ) the time evolution of the @xmath0 abundance from same heavy - ion collisions when the cross sections for the strangeness - exchange reactions are varied by a factor @xmath8 . this changes the @xmath118 abundance to @xmath119 and @xmath120 , corresponding to effects of about @xmath121 and @xmath122 , when the cross sections for the strangeness - exchange reactions are doubled ( dashed line in fig . abndtime ( c ) ) and halved ( dotted line in fig . [ abndtime ] ( c ) ) . our results thus show that the @xmath0 yield in heavy ion collisions does not reach chemical equilibrium , similar to the conclusion of ref.@xcite on subthreshold @xmath10 production in heavy - ion collisions at sis energies . in fig . [ abndtime](c ) , we also show by the dash - dotted curve the @xmath0 abundance in the case of neglecting kaon medium effects . in this case , the final @xmath0 yield is about @xmath123 and is only about @xmath124 less than that from including kaon medium effects . the small effect is due to the fact that kaon medium effects increase the @xmath10 abundance but decrease the abundance of @xmath38 and @xmath39 . the @xmath0 abundance is thus more sensitive to the cross sections for strangeness - exchange reactions than to the medium effects due to the modification of kaon properties . study of @xmath0 production in heavy - ion collisions at sis energies thus provides the possibility to understand the strangeness - exchange reactions @xmath125 and @xmath126 , and to test the hadronic model used in ref.@xcite for studying these reactions . , @xmath10 , and @xmath12 , and ( b ) @xmath127 and @xmath114 ratios as functions of impact parameter @xmath128 in collisions of @xmath4ni + @xmath5ni at @xmath6 gev.,width=288,height=288 ] to see the centrality dependence of particle production , we show in figs . [ bdep](a ) the abundances of @xmath93 , @xmath10 , and @xmath0 , and ( b ) the @xmath129 and @xmath114 ratios as functions of impact parameter @xmath128 in collisions of @xmath5ni + @xmath5ni at @xmath6 . it is seen that the abundances of @xmath93 , @xmath10 and @xmath0 decrease almost linearly with @xmath128 while the @xmath127 and @xmath114 ratios only depends weakly on centrality . for the @xmath127 ratio , its value is in the range @xmath130 for impact parameters between 1 and 4 fm and agrees well with the experimental data @xcite . the weak centrality dependence of @xmath127 was also observed in previous calculations @xcite . the predicted variation of the @xmath114 ratio with impact parameter is in the range @xmath131 . , @xmath10 , and @xmath118 , and ( b ) @xmath127 and @xmath114 ratios from central collisions of @xmath133au + @xmath132au at @xmath94 fm.,width=288,height=288 ] it is also of interest to study @xmath0 production from other collision systems and at different energies . shown in fig . [ aue](a ) are excitation functions for the abundances of @xmath93 , @xmath10 , and @xmath0 from central collisions of @xmath132au + @xmath132au at @xmath94 . the @xmath0 abundance is about @xmath134 , @xmath135 , and @xmath136 at @xmath137 , @xmath138 , and @xmath139 , respectively . therefore , the @xmath0 yield depends on the incident energy and mass of the colliding system . figs . [ aue](b ) shows the @xmath127 and @xmath114 ratios as functions of incident energy in collisions of @xmath132au + @xmath132au at @xmath94 . it is seen that the @xmath127 and @xmath114 ratios increase with increasing incident energy . in particular , large @xmath0 abundance of about @xmath140 and @xmath114 ratio of about @xmath141 are obtained in central @xmath132au + @xmath132au collisions at @xmath142 , which can be studied at ags e866e917,e895 . in summary , we have studied subthreshold @xmath0 production in heavy - ion collisions at energies within the framework of a relativistic transport model that includes explicitly the nucleon , @xmath92 , and pion , and perturbatively the kaon , antikaon , and hyperons . the production of @xmath12 is also treated perturbatively through the strangeness - exchange reactions @xmath143 and @xmath2 with their cross sections taken from the predictions from a hadronic model . we find that the @xmath0 yield is about @xmath3 per event in central collisions of @xmath5ni + @xmath5ni at @xmath6 and can reach to about @xmath144 per event in central collisions of @xmath132au + @xmath132au at @xmath142 . the @xmath0 yield is further found to be more sensitive to the cross sections for the strangeness - exchange reactions than to the kaon medium effects . study of subthreshold @xmath12 production in heavy ion collisions at energies thus provides the possibility of studying the strangeness - exchange reactions and testing the hadronic model used for evaluating their cross sections . this paper was based on work supported by the u.s . national science foundation under grant no . phy-0098805 and the welch foundation under grant no . was also supported by the national natural science foundation of china under grant no . y.t . would like to express his gratitude to the cyclotron institute at tamu for the warm hospitality during his visit , and his work was supported by taiwan s national science council grant nsc92 - 2112-m001 - 64	production of the doubly strange @xmath0 baryon in heavy - ion collisions at energies is studied in a relativistic transport model that includes perturbatively the strangeness - exchange reactions @xmath1 and @xmath2 . taking the cross sections for these reactions from the predictions of a hadronic model , we find that the @xmath0 yield is about @xmath3 in central collisions of @xmath4ni + @xmath5ni at @xmath6 . the @xmath0 yield is further found to be more sensitive to the magnitude of the cross sections for strangeness - exchange reactions than to the medium effects due to modified kaon properties . we have also made predictions for @xmath0 production in au+au collisions at energies from @xmath7 to @xmath8 gev / nucleon .
the most widely used family of exchange - correlation ( xc ) functionals in density - functional theory ( dft ) @xcite , the local density approximation ( lda ) and its gradient - corrected semilocal relatives , has been largely responsible for the enormous popularity of the theory for electronic - structure calculations . despite many successes , there are important properties of the exact xc potential that are missed by the lda and many other popular approximations . in this paper , we shall deal with one such property , namely the discontinuity of the xc potential upon change of particle number @xcite . it is well known that the exact xc energy functional must vary linearly as a function of the total number of particles @xmath0 , displaying a derivative discontinuity every time the system passes through an integral value of @xmath0 @xcite . the delocalization error @xcite , however , causes the lda to predict a nonlinear curvature for the change of xc energy with particle number . this has important practical consequences , for instance leading to a description of molecular dissociation where the resulting isolated atoms end up with unphysical fractional electron numbers . therefore , attempts to model molecular dissociation processes , or any phenomenon that is associated with the transport of charges between well - separated spatial regions or subsystems , must take the discontinuity of the xc functional into account @xcite . in the framework of time - dependent density - functional theory ( tddft ) @xcite , discontinuities of the time - dependent xc potential upon change of particle number have recently been observed numerically @xcite , and shown to be essential for a proper description of strong - field ionization processes . in such processes featuring rapid electron escape , the formation of steps in the xc potential was observed , which eventually develop into fully - fledged discontinuities . these previous time - dependent studies dealt with one - dimensional model atoms , either putting in the potential discontinuity by hand @xcite and studying its influence on single versus double ionization probabilities , or calculating the exact xc potential , including steps and discontinuities , by inverting an exact solution of the time - dependent schrdinger equation @xcite . in this paper we shall work along somewhat different lines : our primary interest is in the question how these features of the time - dependent xc potential influence the behavior of a dissociating system . rather than carrying out a numerically exact calculation of a simple , few - electron model system , we shall work with physically more realistic ( yet still simple ) systems , namely doped semiconductor quantum wells in effective - mass approximation , and study the electron loss of single quantum wells , and the dissociation of an asymmetric double quantum well system . in our tddft simulations of these processes , we shall compare the adiabatic lda with and without two types of self - interaction correction ( sic ) . we will first show that the lda plus sic indeed leads to jumps in the xc potential under the appropriate conditions . then , we will investigate how these jumps affect the dissociation of the double well system . the paper is organized as follows : section 2 gives some theoretical background on the concept of self - interaction error and its correction in dft and tddft , as well as the basics of our quantum well model system . in sections 3 a and b we present results of time - dependent simulations of electron loss and dissociation of single and double quantum wells , respectively . finally , in section 4 we give our conclusions . in static spin - dft ( sdft ) , the noninteracting kohn - sham ( ks ) electrons move in a spin - dependent effective potential given by @xmath1({\bf r } ) = v_{\rm ext,\sigma}({\bf r})+ v_{\rm h}[n]({\bf r } ) + v_{\rm xc,\sigma}[n_\uparrow , n_\downarrow]({\bf r}).\ ] ] the density is obtained via @xmath2 , where the orbital densities are related to the ks orbitals by @xmath3 . the first term in eq . ( [ vsstatic ] ) is the external potential , followed by the hartree and the xc potential . one of the exact conditions that the xc potential must satisfy is to be free of the self - interaction error . this means , in particular , that the hartree and xc potential must cancel each other in a one - electron system . this is automatically true if , for instance , we approximate @xmath4 by the exact - exchange ( exx ) potential @xcite . however , for lda and semilocal functionals this is a hard constraint to be satisfied . therefore , various procedures for removing the self - interaction error of an approximate xc functional have been developed . perdew and zunger @xcite proposed the most widely known sic for approximate xc energy functionals : @xmath5=e_{\rm xc}^{approx}[n_\uparrow , n_\downarrow ] - \sum_{\sigma=\uparrow,\downarrow } \sum_k^{occ } \left(e_{\rm h}[n_{k\sigma } ] + e^{approx}_{\rm xc}[n_{k\sigma},0]\right ) . \label{pzsic}\end{aligned}\ ] ] for a one - electron system it is immediately seen that the xc contributions cancel each other , and @xmath6 becomes equal to the negative of the hartree energy , thus cancelling @xmath7 . the xc potential associated with @xmath6 , formally defined by @xmath8 ( { \bf r } ) = \frac{\delta e_{\rm xc}^{pzsic}[n_\uparrow , n_\downarrow]}{\delta n_\sigma ( { \bf r } ) } , \ ] ] is not straightforward to obtain by direct variation with respect to the density , since @xmath6 is an explicit functional of the orbital densities but only an implicit functional of @xmath9 @xcite . instead , to obtain a local , state - independent pzsic xc potential one must apply the so - called optimized - effective potential ( oep ) @xcite methodology or one of its simplifications such as the krieger - li - iafrate ( kli ) approach @xcite . the latter is defined as follows : @xmath10 \right\ } + \textrm{c.c . } , \ ] ] where @xmath11 } { \delta \varphi_{k\sigma}({\bf r})}\ ] ] and the bars denote orbital averages , e.g. @xmath12 . being self - interaction free , the pzsic formulation also ensures the correct @xmath13 asymptotic behavior for the xc potential , contrasting with the exponential decay predicted by the lda . such a feature is particularly important when dealing , e.g. , with the delocalization error @xcite , since in lda the outermost electrons are not bound strongly enough . moreover , it has been shown that the oep / kli formalism is able to predict the expected potential discontinuity when the particle number crosses an integer @xcite . the pzsic approach is by no means the only method to get rid of self - interaction errors . a more recent version of sic , proposed by lundin and eriksson @xcite , removes the self - interaction directly in the effective potential , which in this approach is given by @xmath14({\bf r } ) = - v_{\rm h}[n_{k\sigma}]({\bf r } ) + v^{approx}_{\rm xc}[n_\sigma - n_{k\sigma},n_{\bar{\sigma}}]({\bf r } ) , \label{lesicv}\ ] ] where the xc potential acting on the @xmath15th orbital depends on all orbital densities except the @xmath15th . just like pzsic , the lesic approach produces a state - dependent xc potential , which also requires the implementation of the oep / kli procedure in order to obtain a local multiplicative ks potential . however , contrary to pzsic , the lesic xc approach is formulated directly for the potential , instead of for the energy , so that in principle we can not use eq . ( [ uxc ] ) of the oep / kli formalism . nevertheless , one can _ assume _ that there exists an energy functional ( even though we do not know its structure ) which gives rise to the le orbital potential ( [ lesicv ] ) , and adopt @xmath16({\bf r})$ ] in eqs . ( [ kli ] ) and ( [ uxc ] ) . other features such as the asymptotic behavior and potential discontinuity are also recovered by lesic , but unlike pzsic it predicts an erroneous correction even for the hypothetic exact functional @xcite . let us now consider dynamical situations . in tddft , the time - dependent analog of eq . ( [ vsstatic ] ) is @xmath17({\bf r},t ) = v_{\rm ext,\sigma}({\bf r},t)+ v_{\rm h}[n]({\bf r},t ) + v_{\rm xc,\sigma}[n_\uparrow , n_\downarrow]({\bf r},t),\ ] ] where @xmath18 follows from the time - dependent ks orbitals . @xmath19({\bf r},t)$ ] is an even more complex object than the static xc potential , since at any given time @xmath20 it must retain memory effects from all previous times @xmath21 . modeling this is a formidable task , and many commonly used tddft approaches therefore make use of the adiabatic approximation , which takes an approximate xc potential from static dft and evaluates it at the instantaneous time - dependent density @xmath22 @xcite . accordingly , our previous equations ( [ vsstatic])-([lesicv ] ) , dealing with static sic and kli calculations , are readily brought under the adiabatic time - dependent umbrella by simply replacing @xmath23 . clearly , adiabatic approximations misrepresent the history of the system , because the resulting td functional does not have a memory of the past . one way to build memory in the functional has been outlined in reference @xcite . here , however , we will focus on aspects where memory effects are not expected to be crucial . therefore , all time - dependent calculations we present in the following will make use of the adiabatic approach . testing density functionals in a well controlled and modeled environment is of crucial importance for the development of ( td)dft . in this sense , the quasi two - dimensional electron gases which occur in doped semiconductor heterostructures are useful since they represent an essentially one - dimensional problem : in the effective - mass approximation , only the direction of quantum confinement needs to be analyzed , whereas the lateral or in - plane electronic degrees of freedom can be treated analytically . in particular , gaas / al@xmath24ga@xmath25as heterostructures have attracted considerable attention since the effective - mass approximation for conduction band electrons works very well in these systems @xcite . we assume that such quantum wells have been grown along the @xmath26 direction and are translationally invariant in the @xmath27 plane . considering only the conduction band , the ks eigenfunctions can be written as @xmath28 where @xmath29 is an area and @xmath30 is a subband index , and @xmath31 and @xmath32 denote in - plane position and wave vectors . the envelope functions @xmath33 satisfy the following one - dimensional ks equation : @xmath34(z ) + v_{\rm xc,\sigma}[n](z ) \right ] \varphi_{j\sigma } ( z ) = \varepsilon_{j\sigma } \varphi_{j\sigma } ( z)\ ] ] with @xmath35 , and @xmath36 being the conduction band fermi level ( @xmath37 from this point on , so that the effective mass @xmath38 is a pure number ; further , the bare electric charge @xmath39 is replaced by an effective charge @xmath40 ) . the external potential @xmath41 is prescribed by the design of the quantum well . to describe the dynamics we then propagate the subband envelope functions @xmath42 using the time - dependent ks equation @xmath43(z , t ) + v_{\rm xc,\sigma}[n](z , t ) \right ] \varphi_{j\sigma}(z , t)\ ] ] with the initial condition @xmath44 . we use lda+sic for @xmath45(z , t)$ ] , and the corresponding tdkli potential @xcite can be written as @xmath46 \right\ } + \textrm{c.c.}\ ] ] with , e.g. , @xmath47 ) retains desirable features of the exact sic , such as the potential discontinuity each time a new subband @xmath30 is filled and @xmath48 asymptotic behavior . in the following , the ks orbitals are propagated on a real - space grid in real time with a crank - nicholson algorithm and a time step of 0.05 a.u . mundt and kmmel @xcite showed for a one - dimensional lithium model atom that the time - dependent exx potential displays humps at its edges as the system loses electrons in an ionization process . in the limit of complete ionization , these humps become more and more steplike , reaching an asymptotic constant value corresponding to the intrinsic derivative discontinuity of the xc potential as the particle number passes through an integer value . here we study an analogous situation , letting electrons escape from a semiconductor quantum well . , the well depth is abruptly reduced to @xmath49 of the initial value . lower panel : quantum well geometry for studying double - well dissociation . the barrier width grows steadily from 0.2 nm at @xmath50 to 10 nm at @xmath51 a.u . [ fig1 ] ] we consider a @xmath52 nm square gaas / al@xmath53ga@xmath54as quantum well with conduction band effective mass @xmath55 and charge @xmath56 , and with an electronic density such that the two lowest subbands ( @xmath57 ) are occupied . initially ( @xmath50 ) , the difference between the conduction band edges of gaas and al@xmath53ga@xmath54as is taken as 220mev . once we have established the ks ground states @xmath58 , for @xmath59 we suppress the barrier height to one third of the initial value . in other words , we consider a sudden switching process which reduces the quantum well depth by 66% , but keeps the square well shape otherwise intact ( see upper panel of fig . [ fig1 ] ) . such a process is , of course , highly idealized and would be difficult to realize experimentally . in practice , a static electric field could be used to suppress the quantum well barrier on one side . applying an absorbing boundary condition , we then let the system evolve such that the electrons in the second subband are almost completely ionized , and focus our attention on the behavior of the lda and sic potentials . we limit ourselves in the following to the exchange - only lda , @xmath60(z , t ) = -\left[6/\pi\ n_\sigma(z , t)\right]^{1/3}$ ] , applying the tdkli / pzsic and tdkli / lesic approaches . number of electrons in the second subband versus time , illustrating electron escape following a suppression of the quantum well depth to one third of its initial value at @xmath50 . [ fig2 ] ] figure [ fig2 ] shows the number of electrons per unit area in the second subband , @xmath61 , as a function of time . due to the lowering of the barrier , electrons are able to escape , move away from the quantum well and are absorbed at the boundary , so that the norm of the envelope functions steadily decreases , @xmath62 . the electrons in the first subband are much more tightly bound , so that the overall ionization of the well is almost exclusively a consequence of emptying the second subband . as one can observe , for pure lda the ionization occurs much faster than for lda+pzsic and lda+lesic . this feature is a consequence of the delocalization error @xcite and the incorrect asymptotic behavior of the lda : since the electrons are not strongly enough bound to the system , they escape more easily . snapshots of the time - dependent lda , pzsic and lesic xc potentials for the electron escape process shown in figure [ fig2 ] . the ionization of the second subband produces a jump of the xc sic potentials which builds up via a step structure that moves rapidly away from the well . [ fig3 ] ] figures [ fig3]a , [ fig3]b and [ fig3]c show snapshots of the xc potentials at different times , in the lda , pzsic and lesic approaches . starting with two occupied subbands at @xmath50 , the second subband is almost completely depleted at @xmath63 and accordingly the xc sic potentials exhibit a rigid shift between these two situations . as ionization sets in , step structures form and quickly migrate away from the quantum well , along with the escaping electrons . similar effects were observed previously in the exx functional for the one - dimensional lithium atom @xcite . as expected , the lda xc potential exhibits neither the formation of the steps nor the resulting jump . figure [ fig3 ] also clearly illustrates the long - range behavior of the lda and sic potentials . the lda exhibits the well - known rapid exponential decay ; both sic potentials have a much longer spatial range , approaching a @xmath48 behavior in the barrier region , sufficiently far away from the well . once the appearance of steps and jumps in the time - dependent xc potential has been confirmed , the next question is whether and how this affects any physical observables . previous studies @xcite were focused mainly on the xc potential itself . in the next section , we present an example where the impact of the discontinuities of the time - dependent xc potential plays an essential role , namely the dissociation process of a double quantum well . a better understanding and more accurate description of dissociation processes is a central issue of much modern research in dft and tddft @xcite . traditionally this problem has been approached from a static point of view @xcite . molecular dissociation displays a basic behavior : the separated atomic systems must have integral charge , and this must be reproduced by any accurate xc functional , either in the static or time - dependent case . unfortunately , most local and semilocal functionals fail to satisfy this requirement and lead to final dissociation products with fractional charges . it has been shown @xcite that , as a molecule dissociates , the xc potential develops a sharp peak at the bond midpoint followed by the buildup of step structures . however , these insights were obtained in a quasistatic picture , using ground - state calculations in some model systems . here , we use a double quantum well system as a simple _ time - dependent _ model of a dissociating heteroatomic molecule . we analyze both the xc potential itself and the effect of jumps or steps on the total number of particles placed in each side of a double well system . we start at @xmath50 with an asymmetric double quantum well divided by an initially very thin ( 0.2 nm ) barrier , such that electrons can be viewed as sharing both wells , analogous to a molecular orbital . the left well has width 32 nm , and the right one has width 8 nm , and both have the same depth of 220 mev ( see lower panel of fig . [ fig1 ] ) . the system is populated with a given total number of electrons , @xmath64 . for @xmath59 , we let the system dissociate , separating the quantum wells from each other by gradually increasing the width of the dividing barrier up to a width of @xmath65 nm . the whole process is allowed to take a total time of 250 a.u . , as indicated in the lower panel of figure [ fig1 ] . we chose such a slow dissociation speed in order to avoid the strong density fluctuations that would be induced if the wells were torn apart too rapidly ; nevertheless , the process is far from being adiabatic , and some charge - density oscillations can not be avoided . at the final separation of 10 nm at @xmath51 a.u . , however , these charge - density oscillations have become small ripples , as we will illustrate below . for all practical purposes , the dissociation can then be considered complete . number of electrons in the left well , @xmath66 , as a function of total electron number @xmath67 of a dissociated double quantum well . both sic approaches predict an abrupt decrease of @xmath66 approaching the region where the second subband would be filled in the isolated well ( dotted lines ) . [ fig4 ] ] we monitor the total number of particles in the left and right quantum well , @xmath66 and @xmath68 , while the system dissociates . in figure [ fig4 ] we display @xmath66 as a function of @xmath67 when the separation of the two wells is complete at @xmath51 a.u . the horizontal dotted lines indicate the number of electrons @xmath66 at which , in a ground - state calculation , the second subband in the isolated left quantum well would start to become occupied . this value of @xmath66 is the same for both sic calculations , but a smaller value is predicted in lda . the reason for this difference is that sic ( as well as exx ) leads to larger intersubband level spacings due to the different asymptotic behavior of the potential @xcite ( see also figure [ fig3 ] ) . as figure [ fig4 ] shows , the lda predicts a continuous and rather smooth increase of @xmath66 . by contrast , the sic results behave dramatically differently when @xmath66 approaches the region where the second subband would start to become occupied in the isolated left well . the electrons seem to resist filling the left well , and one even sees a marked decrease of @xmath66 . if @xmath67 continues to increase , @xmath66 picks up again , and eventually crosses the threshold of the second subband . time evolution of @xmath66 for the parameters of point a and b indicated in figure [ fig4 ] . the drop of particle number in the left well occurs as soon as the system is sufficiently separated , so that the subband structure of the isolated left well starts to become relevant . [ fig5 ] ] in order to understand this behavior , we plot in figure [ fig5 ] the time evolution of @xmath66 calculated with pzsic , up until the dissociated limit of @xmath65 nm , with parameters corresponding to the points a and b in figure [ fig4 ] . situation b starts with a slightly larger density in the left well , but displays a dramatic drop as soon as the two wells are sufficiently separated . the same behavior is found in the lesic approach , comparing points c and d in figure [ fig4 ] . this indicates a repulsive force causing electrons to flow from the left to the right well a clear signature of steps in the potential . snapshots of the pzsic xc potential corresponding to curve b in figure [ fig5 ] . due to the rise of the xc potential in the the left well relative to the right , electrons are prevented from filling the second subband and are forced to move to the right well instead . a potential peak builds up in region between the wells , leading to a repulsive force . the dashed lines indicate @xmath69 , and the dash - dotted lines show the density @xmath70 . [ fig6 ] ] in figure [ fig6 ] we plot snapshots of the pzsic xc potential and of the density @xmath70 at different times , for parameters corresponding to point b in figures [ fig4 ] and [ fig5 ] , i.e. , @xmath71 at the initial time . the snapshots of the xc potential clearly illustrate the mechanism preventing the left quantum well from being more filled : once the double well starts to dissociate , the xc potential builds up steps structures , with a very pronounced sharp peak in the region between the wells . the system resists putting electrons in the second subband : the moment this happens , the potential on the left side shoots up , which means that electrons flow back to the right , as seen from the plots of the density . this effect is clearly absent in the lda . it is evident that in a tddft computational treatment of molecular dissociation such a behavior of xc potential and density must play a crucial role . the key physical requirement is the neutrality of the isolated atomic system . by building up steps and jumps in the potential , the system avoids incorrect fractional - charge dissociation , forcing electrons to flow from an atom to another . this is the main mechanism which restores the nature of neutral atoms during the dissociation processes , with the correct integral final charges . the results presented here demonstrate that the essential features of the lda plus sic approach , namely correct asymptotic behavior and discontinuities upon change of particle number , appear to carry over from static dft into the time domain . comparing two different formulations of sic , pz and le , we find only very little difference between them . our tddft calculations were performed in the tdkli approximation , which leads to an adiabatic xc potential that is state independent . it is known that the tdkli approach may lead to difficulties with violations of the zero - force theorem @xcite , but in our calculations this was not a problem ( see also ref . @xcite ) . the time - dependent sic was used in simulating the electron dynamics in doped semiconductor quantum well systems . we first studied electron escape of a single square well upon sudden reduction of the well depth . the main effect of sic is here due to its asymptotic behavior , which in general makes electron escape proceed at a slower pace compared to the lda . we observed clear indications of a buildup of a step structure in the xc potential , eventually leading to a discontinuity . these observations are in agreement with what kmmel and coworkers found earlier for their model atoms @xcite . the most dramatic effect of the xc potential steps and discontinuities was observed in the simulated dissociation of an asymmetric double quantum well . this system can be compared with a dissociating heteroatomic molecule . of course , quantum wells are extended systems , so that it is not possible to observe the absolute change in particle number . the corresponding feature in quantum wells is instead the subband population . we found that the dissociation of the double well depends dramatically on the electronic structure of the resulting isolated single wells , namely , whether the second subband in the wider well would be populated or not . it turns out that in sic the system tends to resist populating the second subband level as long as possible ; a pronounced peak structure develops between the two wells , and the depth of the xc potential of the wider well jumps with respect to the potential in the other well . as a result , the total particle numbers of the left and the right well after dissociation are dramatically different from what one would find with the lda . our calculations were carried out for relatively slowly dissociating quantum wells . nevertheless , we were quite far away from the adiabatic limit , as indicated by the presence of noticeable fluctuations and charge - density oscillations which , however , subsided once the wells were properly separated . the step structures and jumps of the xc potential become more and more difficult to discern the more nonadiabatic and abrupt the dynamics becomes , and tend to be washed out by strong fluctuations of densities and currents . further systematic studies to identify dynamical regimes and observables for which the time - dependent xc potential discontinuities are relevant remain an important task . however , there is no doubt that the effect is crucial for generic dissociation or fragmentation processes , and can be captured in tddft by simple , self - interaction corrected adiabatic functionals . t. grabo , t. kreibich , s. kurth and e.k.u . gross , in _ strong coulomb correlations in electronic structure calculations : beyond the local density approximation _ , edited by v.i . anisimov ( gordon and breach , 2000 ) , p.203 .	the exact exchange - correlation ( xc ) potential in time - dependent density - functional theory ( tddft ) is known to develop steps and discontinuities upon change of the particle number in spatially confined regions or isolated subsystems . we demonstrate that the self - interaction corrected adiabatic local - density approximation for the xc potential has this property , using the example of electron loss of a model quantum well system . we then study the influence of the xc potential discontinuity in a real - time simulation of a dissociation process of an asymmetric double quantum well system , and show that it dramatically affects the population of the resulting isolated single quantum wells . this indicates the importance of a proper account of the discontinuities in tddft descriptions of ionization , dissociation or charge transfer processes .
the ads / cft correspondence @xcite gives new insight for understanding strongly coupled gauge theories or strongly correlated condensed matter systems . in particular , high temperature superconductors in the framework of the ads / cft correspondence attract much attention . the simple model of holographic superconductors initiated in @xcite has been extended into more realistic models with inhomogeneity @xcite . anisotropic models of superconductors also have been investigated in the context of p - wave superconductors , where a non - abelian gauge field condenses in the superconducting state @xcite . since real - world superconductors exhibit various types of anisotropy , it will be valuable to consider other models of anisotropic holographic superconductors . we then consider in this paper the holographic model in the einstein - maxwell - dilaton theory analyzed in @xcite , where the bulk solution corresponding to the normal state was constructed . this work was motivated by the ir geometry , i.e. , the near - horizon geometry , of an asymptotically anti - de sitter ( _ ads _ ) black _ brane _ spacetime , since the behavior of a holographic superconductor strongly depends on the ir geometry , as we see in the analyses of the fermi surface of a non - fermi liquid @xcite . while anisotropy in the ir geometry has been actively investigated also in the context of lifshitz geomerty @xcite , where anisotropy between time and space is considered , the anisotropy in @xcite is between two spatial directions induced by spatial dependence of the dilaton field , i.e. , the bianchi type @xcite . in this paper , we construct the superconducting state in the einstein - maxwell - dilaton theory by turning on a charged scalar field , and analyze the effect of the anisotropy on the optical conductivity . it will be worthwhile mentioning here the difference between our model and the p - wave holographic model of superconductors @xcite . in our model , the bulk solution is anisotropic not only in the superconducting state but also in the normal state , as if the anisotropy arose due to structures of a superconductor , such as crystal structure and doping . in contrast , the anisotropy from the non - abelian gauge field vanishes in the normal state in the p - wave model of superconductors @xcite . we also point out that , in our model , the anisotropy is described by a parameter in the _ solution _ of the dilaton field and hence it is under our control , while the parameter of the anisotropy in @xcite is a coupling constant of the _ theory_. therefore , it seems natural to view our model as corresponding to a superconductor with the anisotropy due to crystal structure and doping . we numerically construct four - dimensional black brane background bulk solutions dressed with a charged scalar field and investigate the properties of the superconducting states in sec . [ sec : background ] . we then perturb the solutions by the gauge field and investigate the optical conductivities in sec . [ sec : conductivity ] . from the low frequency behavior we will determine the energy gap . conclusion and discussions are devoted in sec . [ sec : discussion ] . in this section , we numerically construct anisotropic black brane solutions in the einstein - maxwell - dilaton theory coupled to a charged scalar field @xmath0 . the solutions with a charged scalar hair correspond to a superconducting state in the dual field theory . when @xmath0 is identically zero , they are reduced to the solutions constructed in @xcite , which correspond to a normal state in the dual field theory . not only the normal state solutions , we construct here the superconducting state solutions . the action we consider is @xmath1 , \notag \\ d_a & \equiv & \nabla _ a - i q a_a , \qquad f_{a b } \equiv \partial _ a a_b - \partial _ b a_a , \label{eqn : action}\end{aligned}\ ] ] where @xmath2 is dilaton , @xmath3 is the ads length scale and an asterisk denotes complex conjugate . the field equations derived from the action take the following form @xmath4 where @xmath5 in order to make the definitions of physical quantities clear , here we start with the ansatz of the metric @xmath6 together with the forms of the gauge potential @xmath7 and the charged scalar field @xmath0 as @xmath8 which we assume describes an asymptotically ads black brane spacetime . then , we should have the asymptotic forms of the metric functions as @xmath9 the asymptotic behavior of the gauge potential and the charged scalar field is shown to take the form @xmath10 where @xmath11 , @xmath12 , @xmath13 and @xmath14 are constants . the constants @xmath11 and @xmath12 are the chemical potential and the charge density in the dual field theory , respectively . on the other hand , for the charged scalar field with the potential given in eq . , either constant @xmath13 or @xmath14 corresponds to the expectation value of an operator @xmath15 in the dual field theory and the other should vanish @xcite . with the time coordinate @xmath16 in eq . , the timelike killing vector @xmath17 is expressed as @xmath18 and then the surface gravity @xmath19 is defined by @xmath20 , with which the hawking temperature @xmath21 is given by @xmath22 . now we transform to the coordinate system employed in @xcite , i.e. , we introduce the new radial coordinate @xmath23 defined as @xmath24 where @xmath25 is the horizon radius , and we rescale the coordinate variables @xmath26 in eq . as @xmath27 thus , the metric and the gauge potential are rewritten as @xmath28 , \label{eqn : metric } \\ a_a & = & \phi ( z ) \frac{l^2}{r_+ } \ : ( d t ) _ a \equiv \phi ( z ) \ : ( d t ) _ a . \label{eqn : gaugers } \end{aligned}\ ] ] in what follows , we exclusively use the new coordinate system . with the new coordinates , the asymptotic behavior and is rewritten as @xmath29 where @xmath30 , @xmath31 , @xmath32 and @xmath33 . notice that in our coordinate system the horizon and infinity correspond to @xmath34 and @xmath35 respectively . then , the temperature @xmath22 is calculated as @xmath36 where a prime denotes the derivative with respect to @xmath23 . we also introduce here the temperature normalized by the charge density @xmath12 as @xmath37 for later convenience . for the dilaton field @xmath2 , we make the ansatz @xmath38 where @xmath39 is a dimensionless constant . this is a simple way to consider spatial anisotropy in holographic models , which was originally used in @xcite . under this assumption , eq . is automatically satisfied . we note that this holographic model becomes isotropic if @xmath39 is zero . we now explain how to solve the field equations . in the coordinate system , eqs . are written as follows @xmath40 we note that the equations are invariant under the scale transformation @xmath41 which rescales the metric by @xmath42 . by employing this rescaling , we set as @xmath43 in what follows , unless we explicitly restore @xmath3 . as the boundary condition on the horizon , @xmath44 , we have @xmath45 . in addition , we set @xmath46 so that @xmath47 is finite . then , the regularity of eqs . , and yields @xmath48 therefore , free parameters at the horizon are chosen as @xmath49 and the rest is determined by eqs . . on the other hand , we assume that all the variables are sufficiently differentiable and @xmath50 vanishes at infinity . then , eq . requires @xmath51 since we are concerned with asymptotically ads solutions , we impose @xmath52 , or equivalently @xmath53 . the asymptotic solutions to the remaining equations are then derived as @xmath54 in order for the metric eq . to take the manifestly ads form , we need to require @xmath55 when we construct the family of solutions with a _ fixed _ value of @xmath39 , it is not allowed to realize the condition by rescaling . it is because the rescaling of @xmath56 necessarily involves the rescaling of @xmath39 , which results in a family of solutions with different values of @xmath39 . hence , we impose the condition as the boundary condition at infinity . for @xmath0 , both of the two terms in are normalizable @xcite and we can impose the boundary condition that either one vanishes . in this paper , we focus on the case where @xmath13 vanishes . furthermore , in the numerical calculations below , we consider solutions with the normalized temperature @xmath57 fixed . we then need to impose the boundary condition at infinity on the gauge potential @xmath58 such that eq . is satisfied for a fixed @xmath57 . thus , the boundary condition at infinity is summarized as @xmath59 therefore , to obtain numerical solutions of eqs . with a fixed temperature @xmath57 , we search for the five parameters in eq . satisfying the boundary condition at infinity eq . , in general . in our holographic model , the charged scalar field @xmath0 plays the role of the order parameter . a normal state of the dual field theory corresponds to the only solution to eqs . with @xmath60 , which occurs when the temperature @xmath21 is higher than a certain temperature @xmath61 . on the other hand , when @xmath62 , there emerges a solution with @xmath63 , and it corresponds to a superconducting state . therefore , @xmath61 is identified with the critical temperature of the dual field theory . here we analyze the properties associated with the superconducting states , which include the critical temperature , the condensate of @xmath0 , and the horizon area . first we determine the critical temperature . to do so , we construct background solutions with @xmath64 by restricting to @xmath65 in eq . , which we have confirmed is numerically consistent with the perturbed solution studied in @xcite . we then consider perturbation on the background solutions and analyze whether a hair , i.e. , non - trivial configuration of @xmath0 , can reside on it . as in the isotropic case @xcite , we expect that the superconducting phase transition is second - order , as actually confirmed below . then , we define the inifinitesimal parameter @xmath66 as @xmath67 and write @xmath0 , near the phase transition point , @xmath68 since the backreaction of @xmath69 onto the metric and the gauge potential is second order in @xmath66 , we only have to solve @xmath70 on the background solution with @xmath64 , in order to determine the critical temperature @xmath61 ( @xmath71 ) . as we saw in sec . [ subsec : ns ] , we impose the boundary condition on @xmath69 at infinity as @xmath72 for a non - trivial solution of @xmath69 , its value at the horizon , @xmath73 , is non - vanishing but arbitrary , since eq . is linear . we then regard eq . as an eigenvalue equation with the eigenvalue being @xmath74 . we solve it on a _ fixed _ background solution , and identify the critical temperature as the temperature of the background , and @xmath74 is _ derived _ as the lowest eigenvalue . since the only free parameter in eq . is @xmath74 , we note that the critical temperature @xmath61 is determined as a function of @xmath74 . as a function of @xmath74 . from top to bottom , the various curves correspond to @xmath75 ( black solid line ) , @xmath76 ( green dotted line ) , @xmath77 ( blue dot - dashed line ) , @xmath78 ( red dashed line ) . ] we show in fig . [ fig : critical_temperature ] the critical temperature normalized by the charge density , @xmath79 , for various values of the anisotropic parameter @xmath39 . we find that for a fixed @xmath12 , @xmath61 monotonically decreases with increasing @xmath39 . it has been reported in @xcite recently that an iron - based superconductor has the same property in the sense that anisotropy lowers the critical temperature . we note , however , that the anisotropy in @xcite is between the vertical and parallel directions with respect to the layered superconductor , in contrast to our holographic model , where we consider the anisotropy within the layered superconductor ( the @xmath56-@xmath80 plane ) . next we investigate the behavior of the condensate of the charged scalar field @xmath0 . according to the ads / cft dictionary , we can read off the value of the condensate @xmath81 as @xmath82 where @xmath14 is defined in eq . . thus , now we let @xmath83 in eq . be back to a free parameter , and we construct background solutions with @xmath63 by solving eqs . . we then determine @xmath14 from their asymptotic forms . in fig . [ fig : condensate ] , we show the value of the condensate as a function of the temperature for a variety of the anisotropic parameter @xmath39 . near the phase transition point @xmath84 , the value of the condensate @xmath81 is numerically found to behave as @xmath85 which implies that the phase transition is second - order , as we expected above . we see from fig . [ fig : condensate ] also that the ratio @xmath86 of the energy gap @xmath87 to the critical temperature @xmath61 monotonically increases with increasing @xmath39 , since the value of the condensate is related to the energy gap of the superconducting state as @xmath88 @xcite . , @xmath89 , further normalized by @xmath90 , as a function of @xmath91 for @xmath92 . from bottom to top , the various curves correspond to @xmath75 ( black solid line ) , @xmath76 ( green dotted line ) , @xmath77 ( blue dot - dashed line ) , @xmath78 ( red dashed line ) . ] it is well - known that the horizon area per unit length of the reissner - nordstrm black brane solution is finite in the extremal limit . this feature is not in accord with the nernst formulation of the third law of thermodynamics and hence the solution is not suitable as a gravity dual of holographic superconductor . the previous study @xcite resolves this problem by considering the normal state in the present holographic model . here we consider whether it remains true also for the superconducting state . in our system , the horizon area per unit length is given by @xmath93 which goes to zero as @xmath94 , if the nernst theorem holds . in fig . [ fig : horizon_area ] , we show the area in the superconducting state as a function of the temperature . we see that the horizon area vanishes in the extremal limit . therefore , our holographic model of superconductor is consistent with the nernst formulation of the third law of thermodynamics also in the superconducting state . for @xmath92 . from top to bottom , the various curves correspond to @xmath95 ( green solid line ) , @xmath96 ( orange dotted line ) , @xmath77 ( blue dot - dashed line ) , @xmath78 ( red dashed line ) . ] in this section , we calculate the optical conductivity and investigate the energy gap of the superconducting state . in order to calculate the conductivity , we consider the fluctuation of the gauge field around the background constructed in sec . [ sec : background ] . we thus write as @xmath97 , \ ] ] where @xmath98 is the perturbation parameter . then , the metric @xmath99 is assumed to be perturbed as @xmath100 , \end{aligned}\ ] ] where @xmath101 is the background metric . these perturbation variables are assumed to have the time dependence as @xmath102 where @xmath103 is a dimensionless constant , with which the frequency in the dual field theory is given by @xmath104 . if we did not consider perturbation of other degrees of freedom , the first - order perturbation of eq . would yield @xmath105 this gives @xmath106 , and then we obtain @xmath107 from eqs . below , which implies that the conductivity in the @xmath56-direction is not well - defined . this difficulty is circumvented by adding a new degree of freedom of purturbation . here we assume that the dilaton @xmath2 is also perturbed as @xmath108 under the assumptions , the perturbation equations are written as @xmath109 whereas eq . is written in a decoupled form , we could not decouple eqs . . in sec . [ subsec : nc ] , we will transform these perturbation equations into a more practical form . we note that the perturbation equations are linear differential equations having a singular point at the horizon @xmath34 and hence they are numerically unstable . here , we transform eqs . into a numerically stable form and explain how to solve the equations . we first introduce the new variables as @xmath110 such that primed and hatted variables possess the same characteristic exponent @xmath111 for their series expansions at the horizon . we thus write as @xmath112 in terms of these new variables , we can rewrite the perturbation equations as @xmath113 which are written , near the horizon , in the form of the eigenvalue equations as @xmath114 where @xmath115 . here we impose the ingoing boundary condition at the horizon , which corresponds to the retarded green s function in the dual field theory . then the eigenvectors of eq . which satisfy this boundary condition are derived as @xmath116 with the corresponding eigenvalues being given by @xmath117 and @xmath118 , respectively . we numerically solve eqs. , by substituting each of these eigenvalues @xmath111 and taking the corresponding eigenvectors in eq . as the initial condtion . we thus obtain in this way two linearly independent solutions . similarly , the eigenvectors of eq . satisfying the ingoing boundary condition are found to be @xmath119 where the eigenvalue corresponding to the first eigenvector is given by @xmath117 , and the eigenvalues for the other two are found to be degenerate and given by @xmath118 . again , we substitute each eigenvalue into eqs . and impose the initial condition as required by the corresponding eigenvectors in eq . , to obtain three linearly independent solutions . the asymptotic behavior at infinity of the perturbation variables are derived from eqs . as @xmath120 where @xmath121 denotes @xmath56 or @xmath80 , and @xmath122 , @xmath123 , @xmath124 , @xmath125 , @xmath126 and @xmath127 are constants . in the ads / cft correspondence , a slower fall - off corresponds to a source of the dual field theory . in particular , @xmath125 corresponds to the expectation value of the current @xmath128 of the dual field theory for the external electric field @xmath129 , and then the optical conductivity in the @xmath121-direction is computed @xcite as @xmath130 on the other hand , @xmath122 represents a thermal gradient which induces an energy flow in the @xmath121-direction . here we do not consider such sources except for the gauge potential , and require @xmath131 as the boundary condition at infinity . the linearly independent solutions constructed as above , i.e. , those solutions whose initial conditions are given by the eigenvectors in eq . and , respectively , may not satisfy the boundary condition eq . . however , eqs . and eqs . are linear , and hence we superimpose these linearly independent solutions , so that eq . is satisfied . now we show the numerical results of the optical conductivity @xmath132 , and analyze the energy gap of our holographic superconductor based on the behavior of the optical conductivity . we first consider the conductivity in the @xmath80-direction . we show in fig . [ fig : y - gap1 ] the real part of conductivity in the @xmath80-direction . each curve in fig . [ fig : y - gap1 ] is the limit curve , to which the conductivity curves converge as the temperature is lowered . we see from fig . [ fig : y - gap1 ] that the conductivity in the @xmath80-direction exhibits behavior similar to the isotropic case @xcite . in particular , we find that @xmath133 drops to zero as the frequency @xmath134 decreases , and hence that the energy gap is well - defined in this case . as a function of the frequency normalized by @xmath12 , @xmath135 , further normalized by @xmath90 , for @xmath92 . from left to right , the various curves correspond to @xmath75 ( black solid line ) , @xmath76 ( green dotted line ) , @xmath77 ( blue dot - dashed line ) , @xmath78 ( red dashed line ) . this figure shows that the energy gap increases with increasing @xmath39 . ] in contrast , the conductivity in the @xmath56-direction shows rather different behavior even for the sufficiently low temperature , as shown in fig . [ fig : x - gap ] . we particularly find that @xmath136 remains non - vanishing even in the low frequency region . this feature of the conductivity is very similar to the psuedogap observed in the p - wave superconductor @xcite . closer observation of fig . [ fig : x - gap ] reveals also that the conductivity within the psuedogap becomes large for the increasing values of the anisotropy parameter @xmath39 . as a function of @xmath137 normalized by @xmath90 for @xmath92 . from left to right , the various curves correspond to @xmath75 ( black solid line ) , @xmath76 ( green dotted line ) , @xmath77 ( blue dot - dashed line ) , @xmath78 ( red dashed line ) . ] since the psuedogap appears in the conductivity in the @xmath56-direction , we estimate the energy gap of our holographic superconductor based only on the conductivity in the @xmath80-direction . for simplicity , in this paper , we shall define the energy gap @xmath87 as the frequency at the inflection point of the conductivity curve . as in the isotropic case @xcite , @xmath138 will get steeper for the larger value of @xmath74 , and then the energy gap will not be so sensitive to @xmath74 . within this accuracy , we find that the energy gap @xmath87 is approximately estimated as @xmath139 as in the isotropic case @xcite , @xmath86 is found to be large enough compared to the one predicted by bcs theory , and hence this feature remains true also in the anisotropic case . as we saw in sec . [ subsec : c - temp ] , the critical temperature @xmath61 depends on the anisotropic parameter @xmath39 . it may be helpful to take into account this @xmath39 dependence of @xmath61 for understanding of the total dependence of @xmath87 on @xmath39 . as we see from fig . [ fig : y - gap2 ] , the curves of @xmath138 are almost degenerate for various values of @xmath39 . thus , we find that the energy gap itself is insensitive to the anisotropy parameter @xmath39 . as a function of @xmath137 for @xmath92 . from left to right , the various curves correspond to @xmath75 ( black solid line ) , @xmath76 ( green dotted line ) , @xmath77 ( blue dot - dashed line ) , @xmath78 ( red dashed line ) . ] in this paper , we considered the anisotropic holographic superconductors in the einstein - maxwell - dilaton theory . we first constructed numerically the bulk background black brane solutions corresponding not only to the normal state , but also to the superconducting state . we then calculated the critical temperature @xmath61 , and found that it decreases as the anisotropy becomes large . it is very interesting to note that a real - world iron - based superconductor with the same property has been reported recently @xcite . we also computed the optical conductivities in both of the @xmath56- and @xmath80-directions . we estimated the energy gap @xmath87 from the conductivity in the @xmath80-direction and found that @xmath86 increases as the anisotropy becomes large , and hence that @xmath87 is larger than @xmath140 . as it is much larger than the bcs prediction @xmath141 , the strong coupling effect for the holographic superconductor model is not broken by the anisotropy , but rather enhanced . on the other hand , the conductivity in the @xmath56-direction never drops to zero even at the sufficiently low temperature . this indicates that a pseudogap appears much like in the p - wave superconductors @xcite , and we found that the magnitude of the pseudogap seems to increase as the anisotropy becomes large . it will be interesting if a peudogap appears generally in an anisotropic model of holographic superconductors , without depending on details of holographic models . we are grateful to t. okamura for helpful discussion . we also thank m. miura for enlightening us on the experiment @xcite . the work of k. m. was supported in part by jsps kakenhi grant number 23740200 . 99 j. maldacena , `` the large n limit of superconformal field theories and supergravity '' adv . phys . 2 , 231 - 252 ( 1998 ) ; o. aharony , s. gubser , j. maldacena , h. ooguri , and y. oz , `` large n field theories , string theory and gravity '' phys . 323 , 183 - 386 ( 2000 ) . s. a. hartnoll , c. p. herzog , and g. t. horowitz , building an ads / cft superconductor " , phys . * 101 * 031601 ( 2008 ) . s. a. hartnoll , c. p. herzog , and g. t. horowitz , holographic superconductors " , jhep * 0812 * 015 ( 2008 ) . k. maeda , t. okamura , and j. koga , `` inhomogeneous charged black hole solutions in asymptotically anti - de sitter spacetime '' , phys . * d 85 * 066003 ( 2012 ) . g. t. horowitz , j. e. santos , and d. tong , `` optical conductivity with holographic lattices , '' jhep * 07 * 168 ( 2012 ) . g. t. horowitz , j. e. santos , and d. tong , `` further evidence for lattice - induced scaling , '' jhep * 11 * 102 ( 2012 ) . g. t. horowitz and j. e. santos , `` general relativity and the cuprates , '' [ arxiv:1302.6586 [ hep - th ] ] . s. s. gubser and s. s. pufu , the gravity dual of a p - wave superconductor " , jhep * 0811 * 033 ( 2008 ) . m. m. roberts and s. a. hartnoll , pseudogap and time reversal breaking in a holographic superconductor " , jhep * 0808 * 035 ( 2008 ) . j. a. hutasoit , g. siopsis , j. therrien , conductivity of strongly coupled striped superconductor " , jhep * 1401 * 132 ( 2014 ) . n. iizuka and k. maeda , study of anisotropic black branes in asymptotically anti - de sitter " , jhep * 1207 * 129 ( 2012 ) . n. iizuka , n. kundu , p. narayan , and s. p. trivedi , holographic fermi and non - fermi liquids with transitions in dilaton gravity " , jhep * 1201 * 094 ( 2012 ) . s. kachru , x. liu , and m. mulligan , gravity duals of lifshitz - like fixed points " , phys . * d 78 * 106005 ( 2008 ) . m. taylor , non - relativistic holography " , arxiv : 0812.0530[hep - th ] . k. goldstein , s. kachru , s. prakash , and s. p. trivedi , holography of charged dilaton black holes " , jhep * 1008 * 078 ( 2010 ) . k. goldstein , n. iizuka , s. kachru , s. prakash , s. p. trivedi , and a. westphal , holography of dyonic dilaton black branes " , jhep * 10 * 027 ( 2010 ) . n. iizuka , s. kachru , n. kundu , p. narayan , n. sircar , and s. p. trivedi , bianchi attractors : a classification of extremal black brane geometries " , jhep * 1207 * 193 ( 2012 ) . d. mateos and d. trancanelli , the anisotropic n=4 super yang - mills plasma and its instabilities " , phys . rev . lett . * 107 * 101601 ( 2011 ) . d. mateos and d. trancanelli , thermodynamics and instabilities of a strongly coupled anisotropic plasma " , jhep * 1107 * 054 ( 2011 ) . i. r. klebanov and e. witten , ads / cft correspondence and symmetry breaking " , nucl . phys . * b 556 * 89 ( 1999 ) . m. miura et al , `` anisotropy and superconducting properties of bafe@xmath142(as@xmath143p@xmath144)@xmath142 films with various phosphorus contents '' , appl . express * 6 * 093101 ( 2013 ) .	we investigate an anisotropic model of superconductors in the einstein - maxwell - dilaton theory with a charged scalar field . it is found that the critical temperature decreases as the anisotropy becomes large . we then estimate the energy gap of the superconductor , and find that the ratio of the energy gap to the critical temperature increases as the anisotropy increases and so it is larger than that in the isotropic case . we also find that peudogap appears due to the anisotropy .
relativistic heavy - ion collisions provide the opportunity to study strongly interacting nuclear matter at different thermodynamic conditions . quantum chromodynamics ( qcd ) , a fundamental theory to describe the interactions of quarks and gluons , has anticipated the transition from hadronic matter to a new state of matter called quark - gluon plasma ( qgp ) [ 1 ] phase at high temperature and high energy density ( @xmath11 gev/@xmath12 ) . the qcd phase diagram is characterized by the temperature ( t ) and the baryon chemical potential ( @xmath13 ) or the ( net ) baryon density ( @xmath14 ) , and it contains the information about the phase boundary that separates the qgp and hadronic phases [ 2,3 ] . finite temperature lattice qcd calculations [ 4 ] predict a cross - over from hadronic to qgp phase at vanishing baryon chemical potential and large t while several qcd - based calculations [ 5 ] show that at lower t and @xmath15 a first - order phase transition may take place . the point in the qcd phase diagram , where the first order phase transition ends would be the qcd critical point [ 6 ] . the bes program at rhic has been carried out with the specific aim to explore several features of qcd phase diagram such as to search for the phase boundaries and the location of qcd critical point by colliding nuclei at several center - of - mass energies . when heavy ions collide at ultra - relativistic energies , the state of the system at the time of the final interaction is reflected by the observed single particle spectra . the process of hadron decoupling from an interacting system is called freeze - out . freeze - out are of two types - kinetic and chemical freeze - out . chemical freeze - out occurs at a temperature ( t@xmath16 ) when inelastic collisions cease and the particle yields become fixed and thermal ( kinetic ) freeze - out occurs at a temperature ( t@xmath17 ) when elastic collisions cease and particle transverse momenta spectra are fixed . + in this paper we discuss the identified transverse momentum spectra and particle ratios produced in au+au collisions at @xmath18 = 27 gev . similar measurements at other bes center - of - mass energies @xmath19 , 11.5 , and 39 gev are reported in [ 7 ] . statistical thermal models , such as thermus [ 8 ] have been proven to be successful in describing the particle production in heavy - ion collisions [ 9 , 10 , 11 ] . experimental particle ratios are used in a statistical thermal model in both grand canonical ensemble ( gce ) and strangeness canonical ensemble ( sce ) approach to extract various chemical freeze - out parameters such as chemical freeze - out temperature ( t@xmath16 ) , baryon chemical potential ( @xmath13 ) , strangeness chemical potential ( @xmath20 ) and strangeness saturation factor ( @xmath21 ) . in this study we have used mid - rapidity particle ratios that include measured yields for charged pions ( @xmath0 ) , charged kaons ( @xmath1 ) , protons ( @xmath2 , @xmath3 ) , @xmath4 , lambdas ( @xmath5 , @xmath6 ) and cascades ( @xmath7 , @xmath8 ) [ 7 , 12 ] . the energy and centrality dependence of extracted chemical freeze - out parameters in au+au collisions at above bes energies are studied . figures 1 and 2 show the transverse momentum spectra of @xmath22 and @xmath23 respectively in au+au collisions at @xmath24 gev in different centralities at mid - rapidity @xmath25 . the @xmath10-integrated pion yields obtained from the bose - einstein fit to @xmath10-spectra have been corrected for feed - down from weak decays . the @xmath26-exponential fit is used to obtain the @xmath10-integrated kaon yields . the collision - energy dependence of the anti - particle to particle ratios in central heavy - ion collisions for @xmath27 and @xmath28/@xmath23 are shown in fig . the @xmath27 ratio at lower beam energies have values larger than unity , which could be due to significant contributions from resonance decays ( such as from @xmath29 baryons ) . the @xmath3/@xmath2 ratio , not shown in this paper , increases with increasing collision energy and approaches unity for top rhic energies [ 21 ] . this indicates that at higher beam energies , the @xmath2 ( @xmath3 ) production at mid - rapidity is dominated by pair production . the @xmath28/@xmath23 ratio approaches unity as collision energy increases , indicating the dominance of kaon - pair production while at lower bes energies associated production of @xmath23 dominates . [ cols="^,^ " , ] the study of the hadron particle ratios at bes energies in star within the framework of a statistical model yields the following conclusions . a centrality dependence of the chemical freeze - out parameters is observed at the lower energies . baryon chemical potential ( @xmath15 ) range extends from 20 to 400 mev in the qcd phase diagram in the new bes measurements at @xmath30 , 11.5 , 27 and 39 gev . a detailed study using particle yields and different choices of ensembles in the statistical model is ongoing : this is expected to provide a better understanding of the freeze - out dynamics . 50 adams j _ et al _ ( star collaboration ) 2005 _ nucl.phys . _ a * 757 * 102 braun - munzinger p _ et al _ 2011 _ arxiv:1101.3167 _ mohanty b 2009 _ nucl . phys . _ a * 830 * 899c aoki y _ et al _ 2006 nature * 443 * 675 ejiri s 2008 _ phys . _ d * 78 * 074507 ; bowman e s and kapusta j i 2009 _ phys . _ c * 79 * 015202 gupta s _ et al _ 2011 _ science _ * 332 * 1525 kumar l ( star collaboration ) 2011 _ j. phys_. g _ nucl . phys . _ * 38 * 124145 cleymans j _ et al _ 2009 _ computer physics communications _ * 180 * 84 cleymans j _ et al _ 2005 _ phys.rev . _ c * 71 * 054901 aggarwal m m _ et al _ ( star collaboration ) 2011 _ phys . _ c * 83 * 024901 ; kumar l ( star collaboration ) 2013 _ nucl . phys . _ a * 904 - 905 * 256c cleymans j _ et al _ 2006 _ phys . _ c * 73 * 034905 zhu x ( star collaboration ) 2012 _ acta phys . _ b * proc . . 5 * 213 - 218 das s ( star collaboration ) 2013 _ nucl . _ a * 904 - 905 * 891c ahle l__et al _ _ ( e866 collaboration and e917 collaboration ) 2000 _ phys . _ b * 490 * 53 klay j l _ et al _ ( e895 collaboration ) 2002 _ phys . lett . _ * 88 * 102301 ahle l _ et al _ ( e866 collaboration and e917 collaboration ) 2000 _ phys . _ b * 476 * 1 afanasiev s v _ et al _ ( na49 collaboration ) 2002 _ phys . _ c * 66 * 054902 alt c _ et al _ ( na49 collaboration ) 2008 _ phys . _ c * 77 * 024903 anticic t _ ( na49 collaboration ) 2004 _ phys.rev . _ c * 69 * 024902 abelev b i _ et al _ ( star collaboration ) 2009 _ phys . _ c * 79 * 034909 adams j _ et al _ ( star collaboration ) 2004 _ phys . lett . _ * 92 * 112301 abelev b i _ et al _ ( star collaboration ) 2010 _ phys . _ c * 81 * 024911	the star detector at rhic , due to its large uniform acceptance and excellent particle identification capabilities , has measured a variety of hadron species ( @xmath0 , @xmath1 , @xmath2 , @xmath3 , @xmath4 , @xmath5 , @xmath6 , @xmath7 , @xmath8 ) produced in au+au collisions at @xmath9 , 11.5 , 27 and 39 gev . these data are part of the beam energy scan ( bes ) program at rhic and provide an opportunity to measure the yields and transverse momentum spectra ( @xmath10 ) of the particles produced in the collisions . the corresponding measurements allow to study the freeze - out properties and dynamics of heavy ion collisions . a statistical thermal model analysis of particle production in bes energies in both grand canonical and strangeness canonical ensembles , is used to extract the chemical freeze - out parameters . the @xmath10 spectra , particle ratios , and the energy and centrality dependence of freeze - out parameters determined from the thermal fit of particle ratios are discussed .
since sumer time @xcite , and much later turing s machine tape @xcite , read / write objects are certainly the most basic memory - based communication objects . such an object , usually called a _ register _ , provides its users ( processes ) with a write operation which defines the new value of the register , and a read operation which returns the value of the register . when considering sequential computing , registers are universal in the sense that they allow to solve any problem that can be solved @xcite . [ [ register - in - message - passing - systems ] ] register in message - passing systems + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in a message - passing system , the computing entities communicate only by sending and receiving messages transmitted through a communication network . hence , in such a system , a register is not a communication object given for free , but constitutes a communication abstraction which must be built with the help of the communication network and the local memories of the processes . several types of registers can be defined according to which processes are allowed to read or write it , and the quality ( semantics ) of the value returned by each read operation . we consider here registers which are single - writer multi - reader ( swmr ) , and atomic . atomicity means that ( a ) each read or write operation appears as if it had been executed instantaneously at a single point of the time line , between is start event and its end event , ( b ) no two operations appear at the same point of the time line , and ( c ) a read returns the value written by the closest preceding write operation ( or the initial value of the register if there is no preceding write ) @xcite . algorithms building multi - writer multi - reader ( mwmr ) atomic registers from single - writer single - reader ( swsr ) registers with a weaker semantics ( safe or regular registers ) are described in several textbooks ( e.g. , @xcite ) . many distributed algorithms have been proposed , which build a register on top of a message - passing system , be it failure - free or failure - prone . in the failure - prone case , the addressed failure models are the process crash failure model , or the byzantine process failure model ( see , the textbooks @xcite ) . the most famous of these algorithms was proposed by h. attiya , a. bar - noy , and d. dolev in @xcite . this algorithm , which is usually called abd according to the names its authors , considers an @xmath0-process asynchronous system in which up to @xmath1 processes may crash ( it is also shown in @xcite that @xmath1 is an upper bound of the number of process crashes which can be tolerated ) . this simple and elegant algorithm , relies on ( a ) quorums @xcite , and ( b ) a simple broadcast / reply communication pattern . abd uses this pattern once in a write operation , and twice in a read operation implementing an swmr register . [ [ fast - operation ] ] fast operation + + + + + + + + + + + + + + to our knowledge , the notion of a _ fast implementation _ of an atomic register operation , in failure - prone asynchronous message - passing systems , was introduced in @xcite for process crash failures , and in @xcite for byzantine process failures . these papers consider a three - component model , namely there are three different types of processes : a set of writers @xmath3 , a set of readers @xmath4 , and a set of servers @xmath5 which implements the register . moreover , a client ( a writer or a reader ) can communicate only with the servers , and the servers do not communicate among themselves . in these papers , _ fast _ means that a read or write operation must entail exactly one communication round - trip delay between a client ( the writer or a reader ) and the servers . when considering the process crash failure model ( the one we are interested in in this paper ) , it is shown in @xcite that , when @xmath6 , the condition @xmath7 is necessary and sufficient to have fast read and write operations ( as defined above ) , which implement an atomic register . it is also shown in @xcite that there is no fast implementation of an mwmr atomic register if @xmath8 . [ [ content - of - the - paper ] ] content of the paper + + + + + + + + + + + + + + + + + + + + the work described in @xcite is mainly on the limits of the three - component model ( writers , readers , and servers constitute three independent sets of processes ) in the presence of process crash failures , or byzantine process failures . these limits are captured by predicates involving the set of writers ( @xmath3 ) , the set of readers ( @xmath4 ) , the set of servers ( @xmath9 , and the maximal number of servers that can be faulty ( @xmath2 ) . both the underlying model used in this paper and its aim are different from this previous work . while keeping the spirit ( basic principles and simplicity ) of abd , our aim is to design a _ time - efficient _ implementation of an atomic register in the classical model used in many articles and textbooks ( see , e.g. , @xcite ) . this model , where any process can communicate with any process , can be seen as a peer - to - peer model in which each process is both a client ( it can invoke operations ) and a server ( it manages a local copy of the register that is built ) . ) , we obtain a two - component model with one writer and reader - server processes . in this model , the necessary and sufficient condition @xmath7 can never be satisfied , which means that , it is impossible to design a fast implementation of a swmr atomic register in such a two - component model . ] adopting the usual distributed computing assumption that ( a ) local processing times are negligible and assumed consequently to have zero duration , and ( b ) only communication takes time , this paper focuses on the communication time needed to complete a read or write operation . for this reason the term _ time - efficiency _ is defined here in terms on message transfer delays , namely , the cost of a read or write operation is measured by the number of `` consecutive '' message transfer delays they require to terminate . let us notice that this includes transfer delays due to causally related messages ( for example round trip delays generated by request / acknowledgment messages ) , but also ( as we will see in the proposed algorithm ) message transfer delays which occur sequentially without being necessarily causally related . let us notice that this notion of a time - efficient operation does not involve the model parameter @xmath2 . in order to give a precise meaning to the notion of a `` time - efficient implementation '' of a register operation , this paper considers two distinct ways to measure the duration of read and write operations , each based on a specific additional synchrony assumption . one is the `` bounded delay '' assumption , the other one the `` round - based synchrony '' assumption . more precisely , these assumptions and the associated time - efficiency of the proposed algorithm are the following . * _ bounded delay _ assumption . + let us assume that every message takes at most @xmath10 time units to be transmitted from its sender to any of its receivers . in such a context , the algorithm presented in the paper has the following time - efficiency properties . * * a write operation takes at most @xmath11 time units . * * a read operation which is write - latency - free takes at most @xmath11 time units . ( the notion of write - latency - freedom is defined in section [ sec : fast - definition ] . intuitively , it captures the fact that the behavior of the read does not depend on a concurrent or faulty write operation , which is the usual case in read - dominated applications . ) otherwise , it takes at most @xmath12 time units , except in the case where the read operation is concurrent with a write operation and the writer crashes during this write , where it can take up to @xmath13 time units . ( let us remark that a process can experience at most once the @xmath13 read operation scenario . ) * _ round - based synchrony _ assumption . + here , the underlying communication system is assumed to be round - based synchronous @xcite . in such a system , the processes progress by executing consecutive synchronous rounds . in every round , according to its code , a process possibly sends a message to a subset of processes , then receives all the messages sent to it during the current round , and finally executes local computation . at the end of a round , all processes are directed to simultaneously progress to the next round . in such a synchronous system , everything appears as if all messages take the very same time to go from their sender to theirs receivers , namely the duration @xmath14 associated with a round . when executed in such a context , the proposed algorithm has the following time - efficiency properties . * * the duration of a write operation is @xmath15 time units . * * the duration of a read operation is @xmath15 time units , except possibly in the specific scenario where the writer crashes while executing the write operation concurrently with the read , in which case the duration of the read can be @xmath16 time units ( as previously , let us remark that a process can experience at most once the @xmath16 read operation scenario . ) hence , while it remains correct in the presence of any asynchronous message pattern ( e.g. , when each message takes one more time unit than any previous message ) , the proposed algorithm is particularly time - efficient when `` good '' scenarios occur . those are the ones defined by the previous synchrony patterns where the duration of a read or a write operation corresponds to a single round - trip delay . moreover , in the other synchronous scenarios , where a read operation is concurrent with a write , the maximal duration of the read operation is precisely quantified . a concurrent write adds uncertainty whose resolution by a read operation requires one more message transfer delay ( two in the case of the @xmath10 synchrony assumption , if the concurrent write crashes ) . [ [ roadmap ] ] roadmap + + + + + + + the paper consists of [ sec : conclusion ] sections . section [ sec : model - definitions ] presents the system model . section [ sec : fast - definition ] defines the atomic register abstraction , and the notion of a time - efficient implementation . then , section [ sec : algorithm ] presents an asynchronous algorithm providing an implementation of an atomic register with time - efficient operations , as previously defined . section [ sec : proof ] proves its properties . finally , section [ sec : conclusion ] concludes the paper . [ [ processes ] ] processes + + + + + + + + + the computing model is composed of a set of @xmath0 sequential processes denoted @xmath17 , ... , @xmath18 . each process is asynchronous which means that it proceeds at its own speed , which can be arbitrary and remains always unknown to the other processes . a process may halt prematurely ( crash failure ) , but executes correctly its local algorithm until it possibly crashes . the model parameter @xmath2 denotes the maximal number of processes that may crash in a run . a process that crashes in a run is said to be _ faulty_. otherwise , it is _ correct _ or _ non - faulty_. [ [ communication ] ] communication + + + + + + + + + + + + + the processes cooperate by sending and receiving messages through bi - directional channels . the communication network is a complete network , which means that any process @xmath19 can directly send a message to any process @xmath20 ( including itself ) . each channel is reliable ( no loss , corruption , nor creation of messages ) , not necessarily first - in / first - out , and asynchronous ( while the transit time of each message is finite , there is no upper bound on message transit times ) . a process @xmath19 invokes the operation `` @xmath21 tag(@xmath22 ) to @xmath20 '' to send @xmath20 the message tagged tag and carrying the value @xmath22 . it receives a message tagged tag by invoking the operation `` @xmath23 tag ( ) '' . the macro - operation `` @xmath24 tag(@xmath22 ) '' is a shortcut for `` * for each * @xmath25 @xmath21 tag(@xmath22 ) to @xmath20 * end for * '' . ( the sending order is arbitrary , which means that , if the sender crashes while executing this statement , an arbitrary possibly empty subset of processes will receive the message . ) let us notice that , due to process and message asynchrony , no process can know if an other process crashed or is only very slow . [ [ notation ] ] notation + + + + + + + + in the following , the previous computation model , restricted to the case where @xmath1 , is denoted @xmath26 $ ] ( crash asynchronous message - passing ) . it is important to notice that , in this model , all processes are a priori `` equal '' . as we will see , this allows each process to be at the same time a `` client '' and a `` server '' . in this sense , and as noticed in the introduction , this model is the `` fully connected peer - to - peer '' model ( whose structure is different from other computing models such as the client / server model , where processes are partitioned into clients and servers , playing different roles ) . a _ concurrent object _ is an object that can be accessed by several processes ( possibly simultaneously ) . an swmr _ atomic _ register ( say @xmath27 ) is a concurrent object which provides exactly one process ( called the writer ) with an operation denoted @xmath28 , and all processes with an operation denoted @xmath29 . when the writer invokes @xmath30 it defines @xmath31 as being the new value of @xmath27 . an swmr atomic register ( we also say the register is _ linearizable _ @xcite ) is defined by the following set of properties @xcite . * liveness . an invocation of an operation by a correct process terminates . * consistency ( safety ) . all the operations invoked by the processes , except possibly for each faulty process the last operation it invoked , appear as if they have been executed sequentially and this sequence of operations is such that : * * each read returns the value written by the closest write that precedes it ( or the initial value of @xmath27 if there is no preceding write ) , * * if an operation @xmath32 terminated before an operation @xmath33 started , then @xmath32 appears before @xmath33 in the sequence . this set of properties states that , from an external observer point of view , the object appears as if it was accessed sequentially by the processes , this sequence ( a ) respecting the real time access order , and ( ii ) belonging to the sequential specification of a read / write register . the notion of a time - efficient operation is not related to its correctness , but is a property of its implementation . it is sometimes called _ non - functional _ property . in the present case , it captures the time efficiency of operations . as indicated in the introduction , we consider here two synchrony assumptions to define what we mean by time - efficient operation implementation . as we have seen , both are based on the duration of read and write operations , in terms of message transfer delays . let us remember that , in both cases , it is assumed that the local processing times needed to implement these high level read and write operations are negligible . let us assume an underlying communication system where message transfer delays are upper bounded by @xmath10 . [ [ write - latency - free - read - operation - and - interfering - write ] ] write - latency - free read operation and interfering write + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + intuitively , a read operation is _ write - latency - free _ if its execution does `` not interleave '' with the execution of a write operation . more precisely , let @xmath34 be the starting time of a read operation . this read operation is _ write - latency - free _ if ( a ) it is not concurrent with a write operation , and ( b ) the closest preceding write did not crash and started at a time @xmath35 . let @xmath36 be a read operation , which started at time @xmath34 . let @xmath37 be the closest write preceding @xmath36 . if @xmath37 started at time @xmath38 , it is said to be _ interfering _ with @xmath36 . [ [ bounded - delay - based - definition ] ] _ bounded delay_-based definition + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + an implementation of a read / write register is _ time - efficient _ ( from a bounded delay point of view ) if it satisfies the following properties . * a write operation takes at most @xmath11 time units . * a read operation which is write - latency - free takes at most @xmath11 time units . * a read operation which is not write - latency - free takes at most * * @xmath12 time units if the writer does not crash while executing the interfering write , * * @xmath13 time units if the writer crashes while executing the interfering write ( this scenario can appear at most once for each process ) . let us assume that the underlying communication system is round - based synchronous , where each message transfer delay is equal to @xmath14 . when considering this underlying synchrony assumption , it is assumed that a process sends or broadcasts at most one message per round , and this is done at the beginning of a round . an implementation of a read / write register is _ time - efficient _ ( from the round - based synchrony point of view ) if it satisfies the following properties . * the duration of a write operation is @xmath15 time units . * the duration of a read operation is @xmath15 time units , except possibly in the `` at most once '' scenario where the writer crashes while executing the write operation concurrently with the read , in which case the duration of the read can be @xmath16 time units . [ [ what - does - the - proposed - algorithm ] ] what does the proposed algorithm + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + as we will see , the proposed algorithm , designed for the asynchronous system model @xmath39 $ ] , provides an swmr atomic register implementation which is time - efficient for both its `` bounded delay''-based definition , and its `` round synchrony''-based definition . the design of the algorithm , described in figure [ algo : fast - basic - version ] , is voluntarily formulated to be as close as possible to abd . for the reader aware of abd , this will help its understanding . [ [ local - variables ] ] local variables + + + + + + + + + + + + + + + each process @xmath19 manages the following local variables . * @xmath40 contains the value of the constructed register @xmath27 , as currently known by @xmath19 . it is initialized to the initial value of @xmath27 ( e.g. , the default value @xmath41 ) . * @xmath42 is the sequence number associated with the value in @xmath40 . * @xmath43 is the sequence number of the last read operation invoked by @xmath19 . * @xmath44 is a synchronization local variable . it contains the sequence number of the most recent value of @xmath27 that , to @xmath19 s knowledge , is known by at least @xmath45 processes . this variable ( which is new with respect to other algorithms ) is at the heart of the time - efficient implementation of the read operation . * @xmath46 is the value of @xmath27 whose sequence number is @xmath44 . [ [ client - side - operation - sf - write - invoked - by - the - writer ] ] client side : operation @xmath47 invoked by the writer + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + when the writer @xmath19 invokes @xmath30 , it increases @xmath42 , updates @xmath40 , and broadcasts the message write@xmath48 ( line [ f - swmr-01 ] ) . then , it waits until it has received an acknowledgment message from @xmath45 processes ( line [ f - swmr-02 ] ) . when this occurs , the operation terminates ( line [ f - swmr-03 ] ) . let us notice that the acknowledgment message is a copy of the very same message as the one it broadcast . [ [ server - side - reception - of - a - message - sc - writewsnv ] ] server side : reception of a message @xmath49 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + when a process @xmath19 receives such a message , and this message carries a more recent value than the one currently stored in @xmath40 , @xmath19 updates accordingly @xmath42 and @xmath40 ( line [ f - swmr-07 ] ) . moreover , if this message is the first message carrying the sequence number @xmath50 , @xmath19 forwards to all the processes the message write@xmath51 it has received ( line [ f - swmr-08 ] ) . this broadcast has two aims : to be an acknowledgment for the writer , and to inform the other processes that @xmath19 `` knows '' this value . when @xmath19 receives a message write@xmath51 for the first time . ] moreover , when @xmath19 has received the message write@xmath51 from @xmath45 different processes , and @xmath44 is smaller than @xmath50 , it updates its local synchronization variable @xmath44 and accordingly assigns @xmath31 to @xmath46 ( lines [ f - swmr-09]-[f - swmr-11 ] ) . [ [ server - side - reception - of - a - message - readrsn ] ] server side : reception of a message read@xmath52 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + when a process @xmath19 receives such a message from a process @xmath20 , it sends by return to @xmath20 the message state@xmath53 , thereby informing it on the freshness of the last value of @xmath27 it knows ( line [ f - swmr-12 ] ) . the parameter @xmath54 allows the sender @xmath20 to associate the messages state@xmath55 it will receive with the corresponding request identified by @xmath54 . [ [ client - side - operation - sf - read - invoked - by - a - process - p_i ] ] client side : operation @xmath56 invoked by a process @xmath19 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + when a process invokes @xmath29 , it first broadcasts the message read@xmath57 with a new sequence number . then , it waits until `` some '' predicate is satisfied ( line [ f - swmr-05 ] ) , and finally returns the current value of @xmath46 . let us notice that the value @xmath46 that is returned is the one whose sequence number is @xmath44 . the waiting predicate is the heart of the algorithm . its first part states that @xmath19 must have received a message state@xmath55 from @xmath45 processes . its second part , namely @xmath58 , states that the value in @xmath19 s local variable @xmath46 is as recent or more recent than the value associated with the greatest write sequence number @xmath50 received by @xmath19 in a message state@xmath55 . combined with the broadcast of messages write@xmath59 issued by each process at line [ f - swmr-08 ] , this waiting predicate ensures both the correctness of the returned value ( atomicity ) , and the fact that the read implementation is time - efficient . the properties proved in this section are independent of the message transfer delays ( provided they are finite ) . [ lemma : termination ] if the writer is correct , all its write invocations terminate . if a reader is correct , all its read invocations terminate . let us first consider the writer process . as by assumption it is correct , it broadcasts the message write@xmath60 ( line [ f - swmr-01 ] ) . each correct process broadcasts write@xmath60 when it receives it for the first time ( line [ f - swmr-08 ] ) . as there are at least @xmath45 correct processes , the writer eventually receives write@xmath60 from these processes , and stops waiting at line [ f - swmr-02 ] . let us now consider a correct reader process @xmath19 . it follows from the same reasoning as before that the reader receives the message state@xmath55 from at least @xmath45 processes ( lines [ f - swmr-05 ] and [ f - swmr-12 ] ) . hence , it remains to prove that the second part of the waiting predicate , namely @xmath61 ( line [ f - swmr-05 ] ) becomes eventually true , where @xmath62 is the greatest write sequence number received by @xmath19 in a message state@xmath55 . let @xmath20 be the sender of this message . the following list of items is such that item @xmath63 @xmath64 item @xmath65 , from which follows that @xmath61 ( line [ f - swmr-05 ] ) is eventually satisfied . 1 . @xmath20 updated @xmath66 to @xmath62 ( line [ f - swmr-07 ] ) before sending state@xmath67 ( line [ f - swmr-12 ] ) . hence , @xmath20 received previously the message write@xmath68 , and broadcast it the first time it received it ( line [ f - swmr-08 ] ) . it follows that any correct process receives the message write@xmath68 ( at least from @xmath20 ) , and broadcasts it the first time it receives it ( line [ f - swmr-08 ] ) . 4 . consequently , @xmath19 eventually receives the message write@xmath68 from @xmath45 processes . when this occurs , it updates @xmath44 ( line [ f - swmr-10 ] ) , which is then @xmath69 , which concludes the proof of the termination of a read operation . [ lemma : atomicity ] the register @xmath27 is atomic . let @xmath70 $ ] be a read operation issued by a process @xmath19 which returns the value with sequence number @xmath63 , and @xmath71 $ ] be the write operation which writes the value with sequence number @xmath72 . the proof of the lemma is the consequence of the three following claims . * claim 1 . if @xmath70 $ ] terminates before @xmath71 $ ] starts , then @xmath73 . * if @xmath74 $ ] terminates before @xmath75 $ ] starts , then @xmath76 . * if @xmath70 $ ] terminates before @xmath77 $ ] starts , then @xmath76 . claim 1 states that no process can read from the future . claim 2 states that no process can read overwritten values . claim 3 states that there is no new / old read inversions @xcite . + proof of claim 1 . + this claim follows from the following simple observation . when the writer executes @xmath71 $ ] , it first increases its local variable @xmath50 which becomes greater than any sequence number associated with its previous write operations ( line [ f - swmr-01 ] ) . hence if @xmath70 $ ] terminates before @xmath71 $ ] starts , we necessarily have @xmath73 . + proof of claim 2 . + it follows from line [ f - swmr-02 ] and lines [ f - swmr-07]-[f - swmr-08 ] that , when @xmath74 $ ] terminates , there is a set @xmath78 of at least @xmath45 processes @xmath79 such that @xmath80 . on another side , due to lines [ f - swmr-04]-[f - swmr-05 ] and line [ f - swmr-12 ] , @xmath75 $ ] obtains a message state@xmath81 from a set @xmath82 of at least @xmath45 processes . as @xmath83 , @xmath84 , and @xmath85 , it follows that @xmath86 is not empty . there is consequently a process @xmath87 , such that that @xmath80 . hence , @xmath79 sent to @xmath19 the message state@xmath88 , where @xmath89 . due to ( a ) the definition of @xmath90 , ( b ) the predicate @xmath91 ( line [ f - swmr-05 ] ) , and ( c ) the value of @xmath92 , it follows that @xmath93 when @xmath75 $ ] stops waiting at line [ f - swmr-05 ] . as , @xmath89 , it follows @xmath94 , which proves the claim . + proof of claim 3 . + when @xmath70 $ ] stops waiting at line [ f - swmr-05 ] , it returns the value @xmath46 associated with the sequence number @xmath95 . process @xmath19 previously received the message write@xmath96 from a set @xmath97 of at least @xmath45 processes . the same occurs for @xmath20 , which , before returning , received the message write@xmath98 from a set @xmath99 of at least @xmath45 processes . as @xmath100 , @xmath101 , and @xmath85 , it follows that @xmath102 is not empty . hence , there is a process @xmath79 which sent state@xmath103 to @xmath19 , and later sent state@xmath104 to @xmath20 . as @xmath105 never decreases , it follows that @xmath76 , which completes the proof of the lemma . [ theo : main-1 ] algorithm _ [ algo : fast - basic - version ] _ implements an _ swmr _ atomic register in @xmath39 $ ] . the proof follows from lemma [ lemma : termination ] ( termination ) and lemma [ lemma : atomicity ] ( atomicity ) . as already indicated , this underlying synchrony assumption considers that every message takes at most @xmath10 time units . moreover , let us remind that a read ( which started at time @xmath34 ) is write - latency - free if it is not concurrent with a write , and the last preceding write did not crash and started at time @xmath106 . [ lemma : fast - writer ] a write operation takes at most @xmath11 time units . the case of the writer is trivial . the message write@xmath81 broadcast by the writer takes at most @xmath10 time units , as do the acknowledgment messageswrite@xmath81 sent by each process to the writer . in this case @xmath11 correspond to a causality - related maximal round - trip delay ( the reception of a message triggers the sending of an associated acknowledgment ) . [ [ when - the - writer - does - not - crash - while - executing - a - write - operation ] ] when the writer does not crash while executing a write operation + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the cases where the writer does not crash while executing a write operation are captured by the next two lemmas . [ lemma : fast - read - write - correct - writer-2 ] a write - latency - free read operation takes at most @xmath11 time units . let @xmath19 be a process that issues a write - latency - free read operation , and @xmath34 be its starting time . moreover , let @xmath107 the starting time of the last preceding write . as the read is write latency - free , we have @xmath108 . moreover , as messages take at most @xmath10 time units , and the writer did not crash when executing the write , each non - crashed process @xmath79 received the message write@xmath96 ( sent by the preceding write at time @xmath109 ) , broadcast it ( line [ f - swmr-08 ] ) , and updated its local variables such that we have @xmath110 ( lines [ f - swmr-07]-[f - swmr-11 ] ) at ime @xmath109 . hence , all the messages state@xmath81 received by the reader @xmath19 carry the write sequence number @xmath63 . moreover , due to the broadcast of line [ f - swmr-08 ] executed by each correct process , we have @xmath95 at some time @xmath111 . it follows that the predicate of line [ f - swmr-05 ] is satisfied at @xmath19 within @xmath11 time units after it invoked the read operation . [ lemma : fast - read - write - correct - writer-3 ] a read operation which is not write - latency - free , and during which the writer does not crash during the interfering write operation , takes at most @xmath12 . let us consider a read operation that starts at time @xmath34 , concurrent with a write operation that starts at time @xmath107 and during which the writer does not crash . from the read operation point of view , the worst case occurs when the read operation is invoked just after time @xmath112 , let us say at time @xmath113 . as a message state@xmath55 is sent by return when a message read@xmath52 is received , the messages state@xmath55 received by @xmath19 by time @xmath114 can be such that some carry the sequence number @xmath63 ( due to last previous write ) while others carry the sequence number @xmath115 ( due to the concurrent write ) are sent by the processes that received read@xmath52 before @xmath107 , while the messages state@xmath116 are sent by the processes that received read@xmath52 between @xmath107 and @xmath117 . ] . hence , @xmath118 or @xmath119 ( predicate of line [ f - swmr-05 ] ) . if @xmath118 , we also have @xmath95 and @xmath19 terminates its read . if @xmath119 , @xmath19 must wait until @xmath120 , which occurs at the latest at @xmath121 ( when @xmath19 receives the last message of the @xmath45 messages write@xmath98 which makes true the predicates of lines [ f - swmr-09]-[f - swmr-10 ] , thereby allowing the predicate of line [ f - swmr-05 ] to be satisfied ) . when this occurs , @xmath19 terminates its read operation . as @xmath122 , @xmath19 returns at the latest @xmath123 time units after it invoked the read operation . [ [ when - the - writer - crashes - while - executing - a - write - operation ] ] when the writer crashes while executing a write operation + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the problem raised by the crash of the writer while executing the write operation is when it crashes while broadcasting the message write@xmath96 ( line [ f - swmr-01 ] ) : some processes receive this message by @xmath10 time units , while other processes do not . this issue is solved by the propagation of the message write@xmath96 by the non - crashed processes that receive it ( line [ f - swmr-08 ] ) . this means that , in the worst case ( as in synchronous systems ) , the message write@xmath96 must be forwarded by @xmath124 processes before being received by all correct processes . this worst scenario may entail a cost of @xmath125 time units . figure [ algo - modified ] presents a simple modification of algorithm [ algo : fast - basic - version ] , which allows a fast implementation of read operations whose executions are concurrent with a write operation during which the writer crashes . the modifications are underlined . when a process @xmath19 receives a message read@xmath81 , it now returns a message state@xmath81 containing an additional field , namely the current value of @xmath40 , its local copy of @xmath27 ( line [ f - swmr-12 ] ) . when a process @xmath19 receives from a process @xmath20 a message state@xmath126 , it uses it in the waiting predicate of line [ f - swmr-05 ] , but executes before the lines [ f - swmr-07]-[f - swmr-11 ] , as if this message was write@xmath51 . according to the values of the predicates of lines [ f - swmr-07 ] , [ f - swmr-09 ] , and [ f - swmr-10 ] , this allows @xmath19 to expedite the update of its local variables @xmath42 , @xmath40 , @xmath44 , and @xmath46 , thereby favoring fast termination . the reader can check that these modifications do not alter the proofs of lemma [ lemma : termination ] ( termination ) and lemma [ lemma : atomicity ] ( atomicity ) . hence , the proof of theorem [ theo : main-1 ] is still correct . [ lemma : faulty - writer ] a read operation which is not write - latency - free , and during which the writer crashes during the interfering write operation , takes at most @xmath13 time units . let @xmath34 be the time at which the read operation starts . as in the proof of lemma [ lemma : fast - read - write - correct - writer-2 ] , the messages state@xmath127 received @xmath19 by time @xmath114 can be such that some carry the sequence number @xmath128 ( due to last previous write ) while some others carry the sequence number @xmath129 ( due to the concurrent write during which the writer crashes ) . if all these messages carry @xmath128 , the read terminates by time @xmath114 . if at least one of these messages is state@xmath130 , we have @xmath131 , and @xmath19 waits until the predicate @xmath132 becomes true ( line [ f - swmr-05 ] ) . when it received state@xmath130 , if not yet done , @xmath19 broadcast the message write@xmath130 , ( line [ f - swmr-08 ] of figure [ algo - modified ] ) , which is received by the other processes within @xmath10 time units . if not yet done , this entails the broadcast by each correct process of the same message write@xmath130 . hence , at most @xmath10 time units later , @xmath19 has received the message write@xmath116 from @xmath45 processes , which entails the update of @xmath44 to @xmath65 . consequently the predicate of line [ f - swmr-05 ] becomes satisfied , and @xmath19 terminates its read operation . when counting the number of consecutive communication steps , we have : the message read@xmath52 by @xmath19 , followed by a message state@xmath130 sent by some process and received by @xmath19 , followed by the message write@xmath116 broadcast by @xmath19 , followed by the message write@xmath116 broadcast by each non - crashed process ( if not yet done ) . hence , when the writer crashes during a concurrent read , the read returns within at most @xmath133 time units . [ theo : main-2 ] algorithm _ [ algo : fast - basic - version ] _ modified as indicated in figure _ [ algo - modified ] _ implements in @xmath39 $ ] an _ swmr _ atomic register with time - efficient operations ( where the time - efficiency notion is based on the _ bounded delay _ assumption ) . the proof follows from theorem [ theo : main-1 ] ( termination and atomicity ) , lemma [ lemma : fast - writer ] , lemma [ lemma : fast - read - write - correct - writer-2 ] , lemma [ lemma : fast - read - write - correct - writer-3 ] , and lemma [ lemma : faulty - writer ] ( time - efficiency ) . as already indicated , this notion of a time - efficient implementation assumes an underlying round - based synchronous communication system , where the duration of a round ( duration of all message transfer delays ) is @xmath14 . [ lemma : writer ] the duration of write operation is @xmath15 . the proof follows directly from the observation that the write operation terminates after a round - trip delay , whose duration is @xmath15 . [ lemma : reader ] the duration of a read operation is @xmath15 time units if the writer does not crash while executing a write operation concurrent with the read . otherwise , it can be @xmath16 . considering a read operation that starts at time @xmath34 , let us assume that the writer does not crash while concurrently executing a write operation . at time @xmath134 all processes receives the message read@xmath52 sent by the reader ( line [ f - swmr-04 ] ) , and answer with a message state@xmath55 ( line [ f - swmr-12 ] ) . due the round - based synchrony assumption , all these messages carry the same sequence number @xmath63 , which is equal to both their local variable @xmath42 and @xmath44 . it follows that at time @xmath135 , the predicate of line [ f - swmr-05 ] is satisfied at the reader , which consequently returns from the read operation . if the writer crashes while concurrently executing a write operation , it is possible that during some time ( a round duration ) , some processes know the sequence number @xmath63 , while other processes know only @xmath136 . but this synchrony break in the knowledge of the last sequence number is mended during the next round thanks to the message write@xmath137 sent by the processes which are aware of @xmath63 ( see figure [ algo - modified ] ) . after this additional round , the read terminates ( as previously ) in two rounds . hence , the read returns at the latest at time @xmath138 . [ theo : main-3 ] algorithm _ [ algo : fast - basic - version ] _ modified as indicated in figure _ [ algo - modified ] _ implements in @xmath39 $ ] an _ swmr _ atomic register with time - efficient operations ( where the time - efficiency notion is based on the _ round - based synchrony _ assumption ) . the proof follows from theorem [ theo : main-1 ] ( termination and atomicity ) , lemma [ lemma : writer ] , and lemma [ lemma : reader ] ( time - efficiency ) . this work has presented a new distributed algorithm implementing an atomic read / write register on top of an asynchronous @xmath0-process message - passing system in which up to @xmath1 processes may crash . when designing it , the constraints we imposed on this algorithm were ( a ) from an efficiency point of view : provide time - efficient implementations for read and write operations , ( b ) and from a design principle point of view : remain `` as close as possible '' to the flagship abd algorithm introduced by attiya , bar - noy and dolev @xcite . + the `` time - efficiency '' property of the proposed algorithm has been analyzed according to two synchrony assumptions on the underlying system . * the first assumption considers an upper bound @xmath10 on message transfer delays . under such an assumption , any write operation takes then at most @xmath11 time units , and a read operation takes at most @xmath11 time units when executed in good circumstances ( i.e. , when there is no write operation concurrent with the read operation ) . hence , the inherent cost of an operation is a round - trip delay , always for a write and in favorable circumstances for a read . a read operation concurrent with a write operation during which the writer does not crash , may require an additional cost of @xmath10 , which means that it takes at most @xmath12 time units . finally , if the writer crashes during a write concurrent with a read , the read may take at most @xmath13 time units . this shows clearly the incremental cost imposed by the adversaries ( concurrency of write operations , and failure of the writer ) . * the second assumption investigated for a `` time - efficient implementation '' is the one provided by a round - based synchronous system , where message transfer delays ( denoted @xmath14 ) are assumed to be the same for all messages . it has been shown that , under this assumption , the duration of a write is @xmath15 , and the duration of a read is @xmath15 , or exceptionally @xmath16 when the writer crashes while concurrently executing a write operation . it is important to remind that the proposed algorithm remains correct in the presence of any asynchrony pattern . its time - efficiency features are particularly interesting when the system has long synchrony periods . differently from the proposed algorithm , the abd algorithm does not display different behaviors in different concurrency and failure patterns . in abd , the duration of all write operations is upper bounded by @xmath11 time units ( or equal to @xmath15 ) , and the duration of all read operations is upper bounded by @xmath13 time units ( or equal to @xmath139 ) . the trade - off between abd and our algorithm lies the message complexity , which is @xmath140 in abd for both read and write operations , while it is @xmath141 for a write operation and @xmath140 for a read operation in the proposed algorithm . hence our algorithm is particularly interesting for registers used in read - dominated applications . moreover , it helps us better understand the impact of the adversary pair `` writer concurrency + writer failure '' on the efficiency of the read operations . this work has been partially supported by the franco - german dfg - anr project 40300781 discmat devoted to connections between mathematics and distributed computing , and the french anr project displexity devoted to the study of computability and complexity in distributed computing . guerraoui r. , levy r. , and chakraborty a. , how fast can a distributed atomic read be ? 23rd acm symposium on principles of distributed computing ( podc04 ) _ , acm press , pp . 236 - 245 , 2004 .	the atomic register is certainly the most basic object of computing science . its implementation on top of an @xmath0-process asynchronous message - passing system has received a lot of attention . it has been shown that @xmath1 ( where @xmath2 is the maximal number of processes that may crash ) is a necessary and sufficient requirement to build an atomic register on top of a crash - prone asynchronous message - passing system . considering such a context , this paper visits the notion of a fast implementation of an atomic register , and presents a new time - efficient asynchronous algorithm . its time - efficiency is measured according to two different underlying synchrony assumptions . whatever this assumption , a write operation always costs a round - trip delay , while a read operation costs always a round - trip delay in favorable circumstances ( intuitively , when it is not concurrent with a write ) . when designing this algorithm , the design spirit was to be as close as possible to the one of the famous abd algorithm ( proposed by attiya , bar - noy , and dolev ) . + + * keywords * : asynchronous message - passing system , atomic read / write register , concurrency , fast operation , process crash failure , synchronous behavior , time - efficient operation .
neutrino experiments have provided a convincing evidence that neutrinos oscillate among different flavors . the experimental data collected over more than a decade imply that at least two neutrinos have nonzero mass and that there is mixing in the lepton sector , in analogy to the cabibbo - kobayashi - maskawa ( ckm ) mixing in the quark sector . yet neutrinos have surprised us since their properties are quite different from those of charged fermions . table 1 reports our present knowledge of neutrino mass and mixing parameters obtained by the analysis of the global @xmath0 oscillation data . @xcite although neutrino oscillation experiments are not sensitive to the absolute neutrino mass scale , direct kinematical searches and cosmological bounds set @xmath1 ev . thus neutrinos are at least six orders of magnitude lighter than the other fermions in the standard model ( sm ) . their mass hierarchies among generations also turn out to be different from those of charged leptons and quarks . in particular , there is a mild neutrino hierarchy @xmath2 . by contrast , for the charged - lepton masses , @xmath3 and @xmath4 , and the quark mass relations @xmath5 are verified . the mixing pattern in the lepton sector also looks quite distinct from its analogue in the quark sector . indeed , while the ckm matrix exhibits a small mixing pattern , the pontecorvo - maki - nakagawa - sakata ( pmns ) lepton mixing matrix has two large mixing angles : the solar @xmath6 and atmospheric @xmath7 . the third angle @xmath8 is mainly constrained by the reactor data to be small , @xmath9 and , from the combined global data , there is a hint for a nonzero @xmath8 . recent data from t2k @xcite and minos @xcite experiments also indicate a relatively large value for @xmath8 . at 90% c.l . , the t2k data are consistent with @xmath10 for normal ( inverted ) hierarchy in the absence of dirac cp violation . the minos collaboration reports the best - fit values @xmath11 . if these results are confirmed , they will have strong impact on neutrino physics and , particularly , will open the window for the possibility of detecting leptonic cp violation . another puzzling aspect of neutrinos is their nature . in the sm neutrinos are strictly massless . this is so because no right - handed ( rh ) neutrino fields are introduced , thus preventing a dirac mass term for them . no majorana mass term of the form @xmath12 can be either generated since such term would violate @xmath13 , which is an exact symmetry in the sm , conserved at the quantum level . if rh neutrino fields are added to the sm particle content , neutrinos could then be either dirac or majorana particles . yet , theoretically , the latter appear to be more natural . indeed , a dirac neutrino mass would require extremely small yukawa couplings @xmath14 . on the other hand , a majorana mass term can be interpreted as the lowest - order effective operator beyond the sm , namely , the dimension - five weinberg operator @xmath15 , where @xmath16 is the sm higgs doublet , @xmath17 stands for the lepton doublet , and @xmath18 is a mass scale associated to some new physics . @xcite then , after electroweak symmetry breaking , neutrinos would acquire a majorana mass @xmath19 , which can be easily suppressed if the scale @xmath18 is much higher than the electroweak scale @xmath20 , without the need of unnaturally small yukawa couplings . this approach is commonly referred to as the seesaw mechanism for neutrino masses . the seesaw is a simple and attractive mechanism for the realization of the weinberg operator in gauge theories . in this framework , the effective operator is induced by the exchange of mediator particles with a mass scale @xmath18 . depending on the su(2 ) representation of the mediators , several seesaw realizations are conceivable for neutrino mass generation . the most common are the type - i ( singlet rh neutrinos ) , type - ii ( triplet scalars ) and type - iii ( triplet fermions ) seesaw mechanisms . such representations are commonly present in grand unified theories . for instance , rh neutrinos nicely fit into the * 16 * spinorial representation of so(10 ) . introducing in the sm right - handed neutrinos @xmath21 with heavy masses @xmath22 is one of the simplest possibilities to generate light neutrino masses . to be specific , let us consider the case of three rh neutrinos , i.e. one for each fermion generation . the lagrangian reads @xmath23 where @xmath24 ; @xmath25 , @xmath26 is the @xmath27 dirac - neutrino yukawa coupling matrix and @xmath28 is the @xmath27 symmetric rh neutrino mass matrix . integrating out the heavy majorana fields in eq . , and after the electroweak symmetry breaking , the effective neutrino mass matrix is given by the seesaw relation @xmath29 therefore , for sufficiently large majorana masses @xmath22 , the required light neutrino masses @xmath30 can be naturally reproduced . in the basis where the charged - lepton yukawa couplings are diagonal , the matrix @xmath31 is diagonalized by the @xmath27 unitary pmns leptonic mixing matrix @xmath32 , @xmath33 light neutrino masses can be generated from other high - scale seesaw mechanisms in a similar manner . as a bonus , seesaw models also give the possibility of implementing leptogenesis , @xcite i.e. the generation of a lepton asymmetry in early universe due to the out - of - equilibrium and cp - violating decays of the seesaw mediators . this asymmetry will be subsequently converted into a baryon asymmetry , thus offering a dynamical explanation for the absence of any primordial antimatter in the observable universe . the neutrino oscillation data in table 1 hint at certain symmetry properties in the leptonic mixing . a remarkable fact is that the @xmath34 mass eigenstate is quite equally mixed with all three flavors @xmath35 ( trimaximal mixing ) , whereas @xmath36 is almost equally mixed with @xmath37 and @xmath38 ( bimaximal mixing ) and contains a very small @xmath39 component . this peculiar flavor structure has led many authors to look for underlying flavor symmetries compatible with the data . @xcite discrete groups such as @xmath40 , @xmath41 , @xmath42 , @xmath43 , @xmath44 , @xmath45 , @xmath46 , @xmath47 , and @xmath48 have been widely used in several proposals . @xcite in this context , the permutation group @xmath40 has been quite popular since it is particularly suitable to reproduce the tribimaximal mixing @xcite @xmath49 which is consistent at @xmath50 level with the experimental data . an attractive feature of mass - independent mixing schemes such as @xmath51 is that they lead to a predictive neutrino mass matrix , which contains just a few parameters that can be directly related to low - energy neutrino observables ( e.g. the neutrino mass - squared differences , the absolute neutrino mass scale , and the effective mass parameter in neutrinoless double beta decays ) . it is nevertheless fair to say that constructing neutrino mass models based on flavor symmetries is not an easy task . additional discrete and/or continuous symmetries are typically required to guarantee the correct vacuum alignment and mass hierarchies , thus making these models quite intricate . in the light of the recent t2k results , models that lead to tribimaximal mixing appear to be disfavored . in general , deviations from @xmath52 in these models can not bring @xmath8 into agreement with data without spoiling the predictions for the solar and atmospheric mixing angles . several alternative solutions have been recently put forward to explain a relatively large value of @xmath8 in the framework of discrete flavor symmetries . @xcite it is also remarkable that the imposition of certain flavor symmetries in the lepton sector of the theory may lead to severe constraints on the leptonic cp asymmetries relevant for leptogenesis in the framework of seesaw models . in particular , in type - i and type - iii seesaw flavor models that lead to an exact mass - independent leptonic mixing the leptogenesis cp asymmetries are zero in leading order . @xcite by contrast , in a type - ii seesaw framework the leptonic cp asymmetries are in general nonzero and leptogenesis is viable . no primordial antimatter is found in our observable universe . from the analysis of the wilkinson microwave anisotropy probe ( wmap ) seven - year data combined with baryon acoustic oscillations it is inferred that @xcite @xmath53 where @xmath54 and @xmath55 are the number densities of baryons , antibaryons and photons , respectively . remarkably , this result is also consistent with the one obtained from the concordance of the light elements and big bang nucleosynthesis . the explanation of such a tiny but nonzero number poses a challenge to both particle physics and cosmology . wmap measurements have also made it clear that the current state of the universe is very close to a critical density and that the primordial density perturbations that seeded large - scale structure formation are nearly scale invariant and gaussian , which is consistent with the inflationary paradigm . since any primordial asymmetry would have been exponentially wiped out during the inflationary period , one then expects the baryon asymmetry to be generated by some dynamical mechanism after inflation . the necessary ingredients for such a dynamical recipe have been formulated long ago by sakharov : b - violation , c and cp - violation , and departure from thermal equilibrium . @xcite it is noteworthy that the sm contains the three sakharov ingredients . yet not all of them are available in a sufficient amount . baryon number is violated by the electroweak sphaleron processes , which are fast and unsuppressed in early universe . the c symmetry is maximally violated by the weak interactions , and cp is violated by the ckm phase , but not enough to generate the required asymmetry @xmath56 . finally , at the electroweak phase transition departure from thermal equilibrium takes place . however , a successful baryogenesis requires a strongly first order phase transition , which can only occur if the higgs mass is less than 60 gev , i.e. in a mass range that is already excluded by the experimental electroweak data . thus the explanation of the cosmological baryon asymmetry requires new physics beyond the sm . among the viable dynamical scenarios to explain the matter - antimatter asymmetry , leptogenesis is undoubtedly one of the simplest , most attractive and well - motivated mechanisms . @xcite new heavy particles are introduced in the theory in such a way that the interactions relevant for leptogenesis are simultaneously responsible for the seesaw neutrino masses . the three sakharov conditions are naturally fulfilled in this framework : ( i ) the seesaw mechanism requires lepton number violation and nonperturbative @xmath57-violating sphaleron processes will partially reprocess the lepton asymmetry into a baryon asymmetry ; ( ii ) neutrino complex yukawa couplings provide the necessary source of cp violation ; ( iii ) departure from thermal equilibrium is accomplished by the out - of - equilibrium decays of the new heavy particles at temperatures above the electroweak scale . as an illustrative example , let us consider again the case of the sm extended with three rh neutrinos . the relevant interactions are described by the type - i seesaw lagrangian of eq . . working in the mass eigenbasis of the charged leptons @xmath58 , and in the mass eigenbasis @xmath59 of the rh neutrinos @xmath21 , the cp asymmetry @xmath60 in the lepton flavor @xmath61 produced in the @xmath59 decays is given by @xmath62}\nonumber \\ & = \frac{1}{8\pi\mathbf{h}_{ii}^{\nu}}\sum_{j \neq i } \bigl\ { \text{im}\left[\mathbf{y}^{\nu\ast}_{\alpha i } \mathbf{h}^\nu_{ij } \mathbf{y}^\nu_{\alpha j}\right]\,\bigl(f(x ) + g(x)\bigr ) + \text{im}\left[\mathbf{y}^{\nu\ast}_{\alpha i}\mathbf{h}^\nu_{ji } \mathbf{y}^\nu_{\alpha j}\right ] g^\prime(x)\bigr\ } , \end{aligned}\ ] ] where @xmath63 , @xmath64 , @xmath65 , \ , g(x ) = \frac{\sqrt{x}\,(1-x)}{(x-1)^2+\gamma_{n_j}^2/m_i^2}=g^\prime(x)\sqrt{x},\end{aligned}\ ] ] and @xmath66 is the total tree - level decay rate of @xmath59 . the cp asymmetry @xmath60 is obtained from the interference of the tree - level and one - loop diagrams of fig . [ fig1 ] . summing over the lepton flavors , one obtains the unflavored asymmetry @xmath67 \,\bigl(f(x ) + g(x)\bigr).\end{aligned}\ ] ] if the universe reheats to a thermal bath composed of particles with gauge interactions after inflation , the final baryon - to - photon number ratio @xmath56 can be estimated as the product of three suppression factors : ( the leptonic cp asymmetry @xmath68 in @xmath59-decays ) @xmath69 ( an efficiency factor @xmath70 due to washout processes in scattering , decays and inverse decays ) @xmath69 ( a reduction factor due to chemical equilibrium , charge conservation and the redistribution of the asymmetry among different particle species ) . in particular , the final efficiency factor @xmath70 is computed by solving numerically the relevant boltzmann equations , which describe the out - of - equilibrium dynamics of the processes responsible for leptogenesis . simple analytical estimates can be obtained in some specific regimes . @xcite traditionally , the unflavored regime has been considered , assuming also a heavy majorana neutrino mass hierarchy @xmath71 ( the so - called @xmath72-dominated scenario ) . in this oversimplified setup , one can show that the cp asymmetry has an upper bound , @xmath73 . to reproduce the observed value given in eq . , two constraints must then be satisfied : a lower bound on @xmath74 and the reheating temperature @xmath75 independent of the initial conditions , @xmath76 gev , and an upper bound on the light neutrino mass scale , @xmath77 ev . @xcite another notable feature of the unflavored asymmetry is that it does not depend directly on the parameters of the pmns mixing matrix @xmath32 . this raises the question of whether there is a direct link between leptogenesis and low - energy neutrino observables . as it turns out , any link can only be established in a model - dependent way , and imposing certain conditions on the yukawa coupling matrix @xmath26 . @xcite going beyond the simple @xmath72-dominated approximation , it is possible to relax the bounds on @xmath74 and @xmath75 . for instance , if @xmath78 , i.e. when the rh neutrinos are quasi - degenerate in mass , the flavored @xmath60 and unflavored @xmath68 asymmetries are resonantly enhanced by the self - energy corrections depicted in fig.[fig1 ] , @xmath79\ , \text{im}\left[\mathbf{y}^{\nu\ast}_{\alpha i } \mathbf{y}^\nu_{\alpha j}\right]}{\mathbf{h}^\nu_{ii}\mathbf{h}^\nu_{jj } } , \quad \epsilon_{i } \simeq -\frac{1}{2 } \sum_{j \neq i } \frac{\text{im}\left[(\mathbf{h}_{ij}^{\nu})^2\right ] } { \mathbf{h}^\nu_{ii}\mathbf{h}^\nu_{jj}}.\end{aligned}\ ] ] although theoretically challenging , it is possible to construct models in which the heavy majorana neutrino mass splitting is naturally as small as the decay width at the leptogenesis scale . in the radiative resonant leptogenesis scenario , the required splitting can be generated by renormalization group running effects , assuming that the heavy majorana neutrinos are exactly degenerate at the gut scale . @xcite recently , it has also been emphasized that lepton flavor effects can play a significant role in leptogenesis . @xcite when the interactions mediated by the charged lepton yukawa couplings are in thermal equilibrium , the flavored leptonic asymmetries and the boltzmann equations for individual flavor asymmetries must be taken into account . since interactions involving the @xmath80 ( @xmath81 ) yukawa couplings are in equilibrium for @xmath82 gev , the corresponding lepton doublets are distinguishable mass eigenstates below these temperatures and should be properly introduced into the dynamics . in the flavored regime , the arguments leading to the upper bound on @xmath68 do not apply , and the leptogenesis efficiency factors can be larger than in the unflavored case . furthermore , the parameters in the mixing matrix @xmath32 directly affect the final baryon asymmetry so that it becomes possible to have successful leptogenesis just from low - energy leptonic cp violation.@xcite the possibility of linking neutrino physics to the cosmological matter - antimatter asymmetry obviously raises the question about the testability of the seesaw and leptogenesis mechanisms . since heavy states and/or very small yukawa couplings are typically present , achieving this goal is not straightforward . there is however a hope that colliders such as the lhc will shed some light on the majorana nature of neutrinos and , in some cases , probe the seesaw as well . smoking gun signatures would be the detection of neutrinoless double beta decay and the production of same - sign lepton pairs at colliders . @xcite testing leptogenesis seems more difficult , and the success of this task will depend on the energy scale at which leptogenesis takes place and the type of seesaw related to it . @xcite in any case , the upcoming years are expected to be another exciting golden era for neutrino physics , and future experiments may provide important clues to some of the still unanswered questions . this work was supported by _ fundao para a cincia e a tecnologia _ ( fct , portugal ) under the projects cern / fp/116328/2010 , ptdc / fis/098188/2008 , and cftp - fct unit 777 , which are partially funded through pocti ( feder ) . m. c. gonzalez - garcia , m. maltoni and j. salvado , jhep * 1004 * ( 2010 ) 056 . t. schwetz , m. trtola and j. w. f. valle , new j. phys . * 13 * ( 2011 ) 063004 ; see also addendum arxiv:1108.1376 . g. l. fogli , e. lisi , a. marrone , a. palazzo and a. m. rotunno , arxiv:1106.6028 . k. abe _ et al . _ [ t2k collaboration ] , phys . * 107 * ( 2011 ) 041801 . p. adamson _ et al . _ [ minos collaboration ] , arxiv:1108.0015 . s. weinberg , phys . d * 22 * ( 1980 ) 1694 . p. f. harrison , d. h. perkins and w. g. scott , phys . b * 530 * ( 2002 ) 167 . e. ma and d. wegman , arxiv:1106.4269 . d. meloni , arxiv:1107.0221 . r. d. a. toorop , f. feruglio and c. hagedorn , arxiv:1107.3486 . e. e. jenkins and a. v. manohar , phys . b * 668 * ( 2008 ) 210 . e. bertuzzo , p. di bari , f. feruglio and e. nardi , jhep * 0911 * ( 2009 ) 036 . d. aristizabal sierra _ et al . _ , b * 827 * ( 2010 ) 34 . r. gonzlez felipe and h. serdio , phys . d * 81 * ( 2010 ) 053008 . i. d. m. varzielas , r. gonzlez felipe and h. serdio , phys . d * 83 * ( 2011 ) 033007 . e. komatsu _ et al . _ [ wmap collaboration ] , astrophys . j. suppl . * 192 * ( 2011 ) 18 . a. d. sakharov , pisma zh . fiz . * 5 * ( 1967 ) 32 . for a review see e.g. s. davidson , e. nardi , y. nir , phys . * 466 * ( 2008 ) 105 . g. f. giudice _ et al . _ , b * 685 * ( 2004 ) 89 . w. buchmller , p. di bari , m. plmacher , annals phys . * 315 * ( 2005 ) 305 . g. c. branco _ et al . _ , nucl . b * 617 * ( 2001 ) 475 . g. c. branco , r. gonzlez felipe , f. r. joaquim and m. n. rebelo , nucl . b * 640 * ( 2002 ) 202 . g. c. branco _ et al . _ , phys . d * 67 * ( 2003 ) 073025 . r. gonzlez felipe , f. r. joaquim and b. m. nobre , phys . d * 70 * ( 2004 ) 085009 . g. c. branco , r. gonzlez felipe , f. r. joaquim and b. m. nobre , phys . b * 633 * ( 2006 ) 336 . g. c. branco , r. gonzlez felipe , m. n. rebelo and h. serdio , phys . d * 79 * ( 2009 ) 093008 . a. abada _ et al . _ , jcap * 0604 * ( 2006 ) 004 . e. nardi , y. nir , e. roulet , j. racker , jhep * 0601 * ( 2006 ) 164 . a. abada _ et al . _ , jhep * 0609 * ( 2006 ) 010 . a. pilaftsis , j. phys . * 171 * ( 2009 ) 012017 . s. pascoli , s. t. petcov , a. riotto , phys . * d75 * ( 2007 ) 083511 . g. c. branco , r. gonzlez felipe and f. r. joaquim , phys . b * 645 * ( 2007 ) 432 . g. senjanovi , arxiv:1012.4104 ; arxiv:1107.5322 .	the discovery of neutrino oscillations provides a solid evidence for nonzero neutrino masses and leptonic mixing . the fact that neutrino masses are so tiny constitutes a puzzling problem in particle physics . from the theoretical viewpoint , the smallness of neutrino masses can be elegantly explained through the seesaw mechanism . another challenging issue for particle physics and cosmology is the explanation of the matter - antimatter asymmetry observed in nature . among the viable mechanisms , leptogenesis is a simple and well - motivated framework . in this talk we briefly review these aspects , making emphasis on the possibility of linking neutrino physics to the cosmological baryon asymmetry originated from leptogenesis .
the study of microscopic structure and dynamics of traveling wave solutions in multi - species one - dimensional stochastic systems has attracted people s attention in this field considerably in recent years @xcite-@xcite . for instance , the microscopic dynamics of shocks are studied for three families of single - species one - dimensional reaction - diffusion systems with open boundaries and nearest - neighbors interactions which include the partially asymmetric simple exclusion process ( pasep ) , the branching - coalescing random walk ( bcrw ) and the asymmetric kawasaki - glauber process ( akgp ) @xcite . it has been shown that in all three systems the time evolution of a product shock measure with a single shock front is equivalent to that of a simple random walker on a finite lattice with homogeneous hopping rates in the bulk and special reflection rates at the boundaries , provided that some constraints on the microscopic reaction rates are fulfilled . the steady - states of these three systems can be essentially written as a linear superposition of such product shock measures . on the other hand , the steady - states of these systems can be obtained using the matrix - product formulation @xcite in which the steady - state weights are written in terms of the product of non - commuting operators which satisfy a quadratic algebra ( for a recent review of this approach see @xcite ) . surprisingly , it has been found that the conditions under which these operators have two - dimensional matrix representations are exactly those for a product shock measure to have a simple random walk dynamics in these systems @xcite . + it has been shown in @xcite that the existence of a two - dimensional representation for the quadratic algebra of a multi - species one - dimensional reaction - diffusion system with open boundaries and nearest - neighbors interactions implies that the steady - state of the system can be written in terms of a linear superposition of product shock measures . it seems that the same is true for the systems whose quadratic algebras have higher - dimensional matrix representations . for instance , for the totally asymmetric simple exclusion process with open boundaries it is known that the quadratic algebra has an infinite - dimensional matrix representation . it has been shown that this is associated with the fact that the steady - state of this system can be written as a linear superposition of product shock measures with infinite number of shocks @xcite . + the microscopic dynamics of shock fronts has also been studied in driven - diffusive systems with more than a single species of particles and also in these systems with next - nearest - neighbor interactions @xcite-@xcite . + despite of these efforts it is still generally unclear whether the existence of a finite - dimensional matrix representation of the quadratic algebra is related to the fact that the steady - state can be expressed as a superposition of product shock measures . in this paper we consider a generalized one - dimensional single - species coagulation - decoagulation model with reflecting boundaries as a new example . we believe that the study of this exactly solvable model provides us with another piece of evidence for the existence of such relation and definitely sheds more light on the unknown aspects of this problem . we will show that a product shock measure with two shock fronts which have simple random walk dynamics can evolve in this system provided that the microscopic reaction rates satisfy some constraints . this will enable us to construct the steady - state of the system simply by considering a linear superposition of such measures ; however , we will not follow this approach here . instead we will show that under the _ same constraints _ the quadratic algebra of this model has a four - dimensional representation . + our paper is organized as follows : in section ( 2 ) we will explain the mathematical preliminaries and introduce the model . in section ( 3 ) we will study the dynamics of a product shock measure with two shock fronts and investigate the conditions ( by imposing some constraints on the microscopic reaction rates of the model ) under which it has a simple time evolution similar to that of two random walkers moving on a one - dimensional lattice while reflecting from the boundaries . the fourth section will be devoted to the investigation of whether a finite - dimensional representation exists for the quadratic algebra of the model under the same conditions . in the last section we will discuss the summery of results . we define @xmath0 as the probability vector of a markovian interacting particle system at the time @xmath1 . the time evolution of this vector is governed by the master equation which can be written as a schrdinger equation in imaginary time as follows : @xmath2 in which @xmath3 is an stochastic hamiltonian . the matrix elements of the hamiltonian are the transition rates between different configurations . for a single - species system with nearest - neighbors interactions defined on a one - dimensional lattice of length @xmath4 with reflecting boundaries the hamiltonian @xmath3 is of the following form : @xmath5 in which : @xmath6 where @xmath7 is a @xmath8 identity matrix and @xmath9 is a @xmath10 matrix for the bulk interactions . in the basis @xmath11 the hamiltonian @xmath9 can be written as : @xmath12 in which we have defined the basis vectors : @xmath13 associated with the presence of a particle and a hole in each lattice site respectively . requiring the conservation of probability , the sum of the elements in each column should be zero ; therefore , one has @xmath14 for @xmath15 . + in our generalized coagulation - decoagulation model the non - zero reaction rates belong to the following reactions : @xmath16 in which @xmath17 and @xmath18 stand for the presence of a particle and a hole in each lattice site respectively . this model has already been studied in related literatures in a couple of special cases . for instance consider the case : @xmath19 the complete spectrum of the hamiltonian in this case has been obtained exactly @xcite . the steady - state of the system has also been studied using both the method of empty intervals @xcite and the matrix - product approach @xcite . it turns out that the phase diagram of the model in this case has two different phases : a low - density phase and a high - density phase which are separated by a coexistence line . the particle concentration on the lattice in both phases have exponential behaviors with three different length scales while on the coexistence line it changes linearly in the bulk of the lattice and grows exponentially near one of the boundaries . by using the matrix product approach it has been shown that for the special tuning of the parameters given in ( [ r2 ] ) the quadratic algebra of the model has a finite - dimensional matrix representation @xcite . since the system has three different length scales and that these length scales are determined by the eigenvalues of these matrices @xcite , the authors in @xcite have found that the minimum dimensionality of these matrices should be four . + the most general case ( [ r1 ] ) has already been studied on a finite lattice with open boundaries in which the particles can enter into the system or leave it from the boundaries @xcite and also on a lattice with periodic boundary conditions @xcite . note that the microscopic reaction rates in our generalized model are exactly those of the bcrw studied in @xcite . as we mentioned before , it is known that traveling wave solutions exist for the bcrw with open boundaries under some constraints on the microscopic reaction rates . in @xcite the author has shown that in the open boundaries case the steady - state can be written as a matrix - product state with two - dimensional matrix representation given that the same constraints are satisfied . the authors in @xcite have studied the dynamical phase transition in this system when considered on an infinite lattice . in what follows we consider the most general reaction rates ( [ r1 ] ) and study the time evolution of a product shock measure with two shock fronts defined as : @xmath20 on a lattice of length @xmath4 with reflecting boundaries . we define two auxiliary sites @xmath21 and @xmath22 for convenience . in figure 1 a simple sketch of a double - shock structure is given . we investigate the conditions under which the time evolution equations for ( [ p ] ) are simply those of two random walkers at sites @xmath23 and @xmath24 moving on a finite lattice . it turns out that by imposing two constraints on the microscopic reaction rates one finds the appropriate answer . these constraints are found to be : @xmath25 provided that the height of the shocks is : @xmath26 one should note that these constraints are exactly those obtained in @xcite for the model with periodic boundary conditions . and @xmath24.,height=113 ] under these constraints the time evolution equations for @xmath27 take the following form : @xmath28 in which the bulk hopping rates for the shock fronts ( random walkers ) are : @xmath29 and also the reflection rate from the boundaries is found to be : @xmath30 note that the original model has six independent parameters while by imposing the constraints ( [ con ] ) only four independent parameters are remained . + the last equation in ( [ te ] ) implies that an empty lattice is a trivial steady - state of the model . it should be mentioned that the random walkers repel each other so that their positions can not get closer than a single lattice site . this has already been observed in @xcite for the special choice of parameters given in ( [ r2 ] ) . for the present case , as in @xcite , one can simply define a new measure @xmath31 as : @xmath32 and investigate its dynamics to see that the shock positions never meet each other at two consecutive sites . the reason for this goes back to the fact that the dynamical rules ( [ r1 ] ) do not allow an empty system to be generated from a system with particles . in fact the system has two different steady - states . the trivial steady - state , as we mentioned above , is the one without any particles . the other one , which is a nontrivial steady - state , contains some particles . if we start with an empty lattice , it always remains empty . in contrast , if we start with a partially - filled lattice , it will reach to an steady - state which contains particles . it can be shown that the nontrivial steady - state of the system can be written as a linear superposition of @xmath27 s . this will be discussed and presented elsewhere . in the next section we are going to calculate the steady - state of the system using the matrix - product approach . the question is that whether or not a finite - dimensional matrix representation exists for the quadratic algebra of this model under the constraints ( [ con ] ) . according to the matrix product formalism the stationary probability distribution @xmath33 of any configuration @xmath34 is assumed to be of the form : @xmath35 in which the occupation number @xmath36 is defined as follows : @xmath37 if the site @xmath38 is empty and @xmath39 if it is occupied by a particle . the factor @xmath40 in ( [ weigth ] ) is a normalization factor and called the partition function of the system and can be easily calculated using the normalization condition @xmath41 . for a system described by the hamiltonian of type ( [ h1 ] ) , the operators @xmath42 and @xmath43 , which stand for the presence of a particle and an empty site at each lattice site respectively , beside the vectors @xmath44 and @xmath45 should satisfy the following quadratic algebra : @xmath46= \left ( \begin{array}{c } \bar{e } \\ \bar{d } \end{array } \right ) \otimes \left ( \begin{array}{c } e \\ d \end{array } \right ) - \left ( \begin{array}{c } e \\ d \end{array } \right ) \otimes \left ( \begin{array}{c } \bar{e } \\ \bar{d } \end{array } \right ) , \\ \\ < w\| \left ( \begin{array}{c } \bar{e } \\ \bar{d } \end{array } \right)=0 , \quad \left(\begin{array}{c } \bar{e } \\ \bar{d } \end{array } \right ) \|v>=0 \end{array}\ ] ] in which the operators @xmath47 and @xmath48 are two auxiliary operators and @xmath9 is of the form ( [ h2 ] ) . using ( [ mpa ] ) one can simply calculate the corresponding quadratic algebra of our model : @xmath49 = [ e,\bar{e } ] = 0 , \\ \\ ( \omega_{32}+\omega_{42}+\omega_{24 } ) ec - ( \omega_{32}+\omega_{42}+\omega_{24}-\omega_{23 } ) e^{2 } - \omega_{24 } c^{2}+\\(\omega_{24}-\omega_{23})ce=\bar{e}c - e\bar{c } , \\ \\ ( \omega_{34}-\omega_{32})ec -(\omega_{23}+\omega_{43}+\omega_{34}-\omega_{32})e^{2 } - \omega_{34}c^{2}+\\ ( \omega_{34}+\omega_{23}+\omega_{43})ce= \bar{c}e - c\bar{e } , \\ \\ \langle w \vert\bar{e}=\langle w \vert \bar { c}=0 , \\ \\ \bar{e } \vert v \rangle = \bar{c}\vert v \rangle=0 \end{array}\ ] ] in which we have defined @xmath50 and @xmath51 . + for the special tuning of the parameters given in ( [ r2 ] ) the authors in @xcite have shown , for the first time , that the quadratic algebra ( [ finalalgebra ] ) has a four - dimensional matrix representation with non - diagonal @xmath47 and @xmath52 . in our generalized model it turns out that the quadratic algebra ( [ finalalgebra ] ) also has a four - dimensional matrix representation provided that the constraints ( [ con ] ) are fulfilled . note that these constraints are automatically fulfilled for ( [ r2 ] ) . + as long as the eigenvalues of the matrix @xmath53 are not equal and in an appropriate basis this matrix can be brought into a diagonal form . we have found that the matrices @xmath53 and @xmath43 besides the auxiliary matrices @xmath52 and @xmath47 have the following four - dimensional representation : @xmath54 in which @xmath55 and @xmath56 are functions of @xmath57 s and are presented in the appendix a. in this representation we have chosen the other parameters @xmath58 and @xmath59 in ( [ diagrepbulk ] ) as free parameters . on the other hand , the vectors @xmath45 and @xmath44 have the following matrix representations : @xmath60 in which @xmath61 and @xmath62 are free parameters , while @xmath63 and @xmath64 are complicated functions of @xmath57 s and are presented in the appendix a. nevertheless , if one looks for a nontrivial steady - state then the relation between @xmath61 and @xmath62 should be obtained by requiring that the probability of finding an empty lattice in long time limit is zero i.e. : @xmath65 by considering the constraint ( [ triv ] ) the partition function of the model can be easily calculated as follow : @xmath66 in which the coefficients @xmath67 s are function of @xmath57 s . the explicit form of these coefficients are presented in the appendix b. the phase diagram of the model can now be obtained by analyzing the thermodynamical behavior of the partition function @xmath40 ( i.e. in the limit @xmath68 ) which obviously depends on the values of the hopping rates of the shock positions i.e. @xmath69 and @xmath70 . one should note that the model has a mirror symmetry which means that it is invariant under the following transformations : @xmath71 defining two new variables @xmath72 and @xmath73 , it can be shown using ( [ ref ] ) that the case ( @xmath74 ) never happens because @xmath75 then has to be negative . and @xmath76 . the four regions i - iv are defined in the text . the coexistence line between the regions ii and iii is @xmath77 . , height=207 ] in terms of the two variables @xmath78 and @xmath76 the phase diagram of the model has four regions defined as : @xmath79 the phase diagram of the model is shown in figure 2 . one can easily check that under the symmetry ( [ sy ] ) the two variables @xmath78 and @xmath76 become @xmath80 and @xmath81 respectively . this implies that we need only to study two phases ( either i and ii or iii and iv ) since the information for the two other phases can be easily obtained using ( [ sy ] ) . in what follows we only consider the two phases iii and iv . in the phase iii the left random walker ( the left shock front ) has more tendency to move leftwards while the right random walker ( the right shock front ) has more tendency to move rightwards . in this phase the mean distance between the two random walkers is of order @xmath4 and therefore the density of particles in the bulk of the lattice increases ( this can be understood by looking at the definition of @xmath27 ) . we call this phase the high - density phase . in contrast , in the phase iv both random walkers have more tendency to move leftwards , and therefore , the mean distance between the random walkers ( the shock fronts ) decreases . in this phase the mean distance between the two random walkers is of order @xmath82 . since the mean density of particles decreases in this phase , it is called the low - density phase . since the matrix representation of the quadratic algebra is finite - dimensional , the density profile of the particles defined as : @xmath83 either has linear or exponential behaviors . note that in calculating ( [ prof ] ) the constraint ( [ triv ] ) should also be applied to its nominator . it turns out that ( [ prof ] ) has only exponential behaviors in both phases iii and iv . in the low - density ( high - density ) phase the mean density of the particles in the bulk of the lattice is zero ( @xmath75 ) while at the boundaries the density profile of the particles changes exponentially with two correlation lengths @xmath84 and @xmath85 ( @xmath84 and @xmath86 ) . + it is also interesting to study the coexistence line between the two phases iii and iv i.e. on the line @xmath87 for @xmath88 . one should note that in this case two eigenvalues of the matrix @xmath53 become equal , and thereby can not be diagonalized . on the coexistence line between the two phases iii and iv , we have found the following matrix representation for the operators in ( [ finalalgebra ] ) : @xmath89 and for the vectors : @xmath90 which is valid if the reaction rates @xmath57 s in ( [ r1 ] ) satisfy both ( [ con ] ) and the following constraint : @xmath91 as in the diagonal case some of the matrix elements are free . in ( [ nondiagrepbulk ] ) the free parameters are chosen to be @xmath92 and @xmath59 . the rest of the matrix elements are functions of @xmath57 s and are presented in the appendix a. the two elements @xmath61 and @xmath62 can be eliminated from the calculations by applying the constraints ( [ triv ] ) while the other elements @xmath63 and @xmath64 are presented in the appendix a. the partition function of the system , after applying the constraint ( [ triv ] ) , is obtained to be : @xmath93 where the coefficients @xmath94 s are complicated functions of the reaction rates @xmath57 s . these coefficients are presented in the appendix b. as can be seen , on the coexistence line the left random walker ( the left shock front ) has more tendency to move leftwards while the right random walker ( the right shock front ) performs an unbiased random walk . in this case we expect that the density of particles increases near the left boundary . the density profile of the particles , defined in ( [ prof ] ) , can be calculated on this line and it turns out that it is an exponentially increasing function near the left boundary with a characteristic length @xmath86 while it changes linearly in the bulk of the lattice . in this paper we considered a generalized coagulation - decoagulation model on a finite lattice with reflecting boundaries and studied the conditions under which the steady - state probability distribution of the system can be written as a product of non - commutative operators . we started with a product shock measure with two shock fronts and obtained its dynamics generated by the hamiltonian of the system . we found the necessary conditions on the microscopic reaction rates under which each of the shock fronts has a simple biased random walk dynamics . finally by studying the steady - state of this system using the matrix - product approach we showed that its quadratic algebra has a four - dimensional matrix representation under the same conditions . the phase diagram of the model and also the density profile of the particles in each phase and on the coexistence line were studied . we saw that the whole phase diagram of the model can be described by the hopping rates of the shock fronts . the system has three different length scales . these are given by the eigenvalues of the matrix @xmath53 which can be written in terms of the hopping rates of the shock fronts . + one can easily check that for the special tuning of the parameters ( [ r2 ] ) the constraints ( [ con ] ) are satisfied and that all of our calculations reduce to those obtained in @xcite . it is interesting to investigate the solutions of ( [ te ] ) in long time limit and construct the steady - state of the system as a linear superposition of such solutions and then compare the results with those obtained from the matrix - product approach in present work . the results will be presented elsewhere . here we present the matrix elements for the diagonal matrices ( [ diagrepbulk ] ) and the vectors ( [ diagrepboun ] ) . the matrix elements of @xmath43 are obtained to be : @xmath95 the other matrix elements @xmath96 , @xmath97 and @xmath59 are free . for the matrix @xmath98 we have found : @xmath99 for the vector elements in ( [ diagrepboun ] ) we have found : @xmath100 on the other hand for the non - diagonal matrices ( [ nondiagrepbulk ] ) and the vectors ( [ nondiagrepboun ] ) we have found the following results . the matrix elements of @xmath43 are obtained to be : @xmath101 as in the diagonal case the other matrix elements @xmath96 , @xmath97 and @xmath59 are free . the matrix elements for @xmath98 are found to be : @xmath102 the non - free vector elements in this case are given by : @xmath103 the coefficients @xmath67 s for @xmath104 in ( [ partition1 ] ) are found to be : @xmath105 @xmath106 @xmath107 @xmath108 the coefficients @xmath94 s for @xmath104 in ( [ partition2 ] ) are also found to be : @xmath109 @xmath110 @xmath111 @xmath112 fhj would like to acknowledge the financial support provided by bu - ali sina university . 1 k. krebs , f. h. jafarpour , and g. m. schtz _ new journal of physics _ * 5 * 145.1 - 145.14 ( 2003 ) f. h. jafarpour and s. r. masharian _ _ p10013 ( 2007 ) v. belitsky and g. m. schtz _ el . j. prob . _ * 7 * paper no.11 1 ( 2002 ) f. h. jafarpour and s. r. masharian _ phys . * e 77 * 031115 ( 2008 ) m. arabsalmani and a. aghamohammadi _ eur . j. _ * b 55 * 439 ( 2007 ) m. arabsalmani and a. aghamohammadi _ phys . _ * e 74 * 011107 ( 2006 ) a. rkos and g. m. schtz _ j. stat . phys . _ * 117 * 55 ( 2004 ) f. h. jafarpour _ physica _ * a 358 * 413 ( 2005 ) f. h. jafarpour and s. r. masharian _ j. stat . . _ p03009 ( 2007 ) f. tabatabaei and g. m. schtz _ diffusion fundamentals _ * 4 * 5.1 - 5.38 ( 2006 ) m. paessens and g. m. schtz _ new journal of physics _ * 6 * 120 ( 2004 ) f. h. jafarpour _ physica a _ * 339 * 369 ( 2004 ) r. a. blythe and m. r. evans _ j. phys . a : math . _ * 40 * r333 ( 2007 ) h. hinrichsen , k. krebs and i. peschel _ z. phys . _ * b 100 * 105 ( 1996 )	the steady - state of a generalized coagulation - decoagulation model on a one - dimensional lattice with reflecting boundaries is studied using a matrix - product approach . it is shown that the quadratic algebra of the model has a four - dimensional representation provided that some constraints on the microscopic reaction rates are fulfilled . the dynamics of a product shock measure with two shock fronts , generated by the hamiltonian of this model , is also studied . it turns out that the shock fronts move on the lattice as two simple random walkers which repel each other provided that the same constraints on the microscopic reaction rates are satisfied .
the single top quark production processes are of great importance at hadron colliders since they allow a direct measurement of the coupling of the @xmath2-boson to the top quark _ i.e. _ the ckm matrix element @xmath4 . these processes can also be used to search for anomalous couplings of the top quark . with a mass in the neighborhood of 175 gev@xcite , the top quark is by far the heaviest of the known quarks . as a consequence , the electroweak decay of the top quark proceeds so rapidly that toponium bound states and @xmath5 mesons do not have time to form @xcite and the decay products of the top quark are correlated with its spin @xcite . therefore , if a top quark is produced with a significant spin correlation , this correlation will be translated into large angular correlations in such events . studies of these angular correlations in single top production can then be used as sensitive searchs for anomalous couplings of the top quark , _ i.e. _ physics beyond the standard model @xcite . traditionally in high energy physics processes , the discussions of spin - related observables take place in terms of the helicities of the fermions involved . however this description is most useful when the fermions are produced in the ultra - relativistic limit because in this limit the chirality eigenstates are identical to the helicity eigenstates . for fermions which are not ultra - relativistic , such as the top quark produced at the tevatron , one must deal with the fact that the chirality and helicity of a massive fermion may not be specified simultaneously . therefore , there is no _ a priori _ reason to believe that the helicity basis will give the best description of the spin of top quarks at the tevatron . in fact , it has recently been shown that the helicity basis does not lead to the largest values for various spin - related correlations in @xmath6 production at either the tevatron@xcite or an @xmath7 collider@xcite . thus , it is natural to ask : is there a better spin basis than helicity for the description of the spin correlations in _ single _ top production ? the answer to this question is yes : we will construct such a spin basis in this paper . we will concentrate on two important production mechanisms for single top quarks at the tevatron . both of these mechanisms produce the single top quark in a left handed chirality state through a virtual @xmath2 boson . therefore , significant spin correlations are expected even at the tevatron , where the top quark is produced well below the ultra - relativistic limit . the first of these ( fig . [ wstardiagram ] ) is the purely electroweak @xmath3 channel @xcite u \|d t\|b , [ wstarprocess ] while the second ( fig . [ wgfusiondiagram ] ) consists of the so - called @xmath2-gluon fusion ( @xmath8 fusion ) processes @xcite u g & & t \|b d ; \|d g & & t \|b \|u . [ wgfusionprocess ] the feasiblity of isolating single top quark production in a collider environment has been already been demonstrated for both the @xmath3 channel @xcite and the @xmath2-gluon fusion process @xcite . early on it was recognized by willenbrock and dicus @xcite that the @xmath8 fusion process is dominated by the configuration where the @xmath9 quark is nearly collinear with the incoming gluon , leading to a logarithmic factor @xmath10 in the total cross section . in the event that this factor is too large as a suitable measure , where @xmath11 is the strong coupling constant , the top quark mass @xmath12 , and the bottom quark mass @xmath13 . ] the perturbative calculation of the @xmath14 process becomes unreliable , and one should instead compute @xmath15 , with the large logarithm being absorbed into the @xmath16 parton distribution function . this latter approach has been employed by bordes _ @xcite in their effort to accurately compute the total cross section , including higher - order corrections . among their conclusions is the statement that to lowest order , for top quark masses up to a few hundred gev , the two pictures give comparable event descriptions and lead to similar cross sections @xcite . this statement is also true for the top quark spin correlation . therefore , we will frame our discussion of @xmath8 fusion in terms of the tree level description involving only the diagrams in fig . [ wgfusiondiagram ] . in sections ii and iii we discuss in detail the top quark spin correlations in single top quark production via the @xmath3 process and @xmath8 fusion process , respectively . finally we end with a discussion and conclusions . in the appendix we give an example in detail of how the top spin correlations lead to angular correlations in events . the example given is single top production in the @xmath1 channel . throughout this paper we will only consider processes which produce a top quark in the final state . the treatment of the charge - conjugated processes , where a top antiquark is produced , is similar . we begin with the simpler of the two production mechanisms for single top quarks at the tevatron , the electroweak process @xmath17 , which proceeds via a virtual @xmath2 boson ( see fig . [ wstardiagram ] ) . we represent the momentum of the each particle by its symbol , and write the amplitude in crossing symmetric form with all momenta outgoing . our results are easily derived using the spinor helicity method for massive fermions described in @xcite to treat the top spin . in particular , we decompose the top quark momentum into a sum of two massless auxiliary momenta , t_1 ( t + m_ts ) ; t_2 ( t - m_ts ) , where @xmath18 is the usual spin vector of the top quark . in the rest frame of the top quark , the spin of the top quark is in the same direction as the spatial part of @xmath19 . then , the matrix element squared for the production of a spin up top quark summed over color and all of the other spins is ( 0\|u d t_\|b)^2 = g_w^4 n_c^2 [ wstargenup ] while for a spin down top quark we have ( 0\|u d t_\|b)^2 = g_w^4 n_c^2 , [ wstargendn ] where @xmath20 is the weak coupling constant , @xmath21 and @xmath22 are the mass and width of the @xmath2 boson , @xmath23 is the number of colors , and @xmath24 is the cabibbo maskawa matrix element . throughout this paper we assume the standard model with three generations and suppress the ckm factor @xmath25 . the sum of ( [ wstargenup ] ) and ( [ wstargendn ] ) is obviously independent of the choice of the spin axis of the top quark , as is required . it is clear that the top quarks produced via the @xmath1 process are 100% polarized along the direction of the @xmath0-type quark , since ( [ wstargendn ] ) vanishes if we choose @xmath26 . consequently , the ideal basis for studying the @xmath27 spin is the one which uses the direction of the @xmath0-type quark as the spin axis . ( see the appendix for a discussion of this process keeping track of all of the correlations between production and decay . ) of course , in an actual experiment , we know only that one of the two initial state partons is a @xmath28 . however , the largest contribution to the total cross section comes from the case where the @xmath28 is donated by the antiproton . in fact , for the tevatron at 2 tev , we estimate that 98% of the cross section may be attributed to this configuration ( see table [ wstarpartons ] ) . this suggests that an excellent choice would be to decompose the top spin along the direction of the antiproton beam , independent of the actual identity of the parton supplied by that beam . we will refer to this choice as the `` antiproton '' basis . to aid in the comparison of the antiproton basis to the more traditional helicity decomposition , we present the matrix elements as a function of the angle @xmath29 between the direction of the top quark in the zero momentum frame ( zmf ) of the initial parton pair and the @xmath30 axis , and the speed @xmath31 of the top quark in the zmf . we orient our coordinate system such that the protons travel in the positive @xmath32 direction ; the antiprotons travel in the negative @xmath32 direction . because eqs . ( [ wstargenup ] ) and ( [ wstargendn ] ) are not symmetric under the interchange of @xmath33 and @xmath0 , the expressions we obtain in terms of these variables will depend upon which beam the @xmath28 quark comes from . in the following equations , the parton taken from the proton will always be written first , followed by the parton taken from the antiproton . we now turn to the actual matrix elements for the antiproton basis , where spin up means that in the rest frame of the top quark , its spin points in the same direction as the incoming antiproton beam is traveling in that frame . for the 98% of the time that the @xmath28 comes from the antiproton , we have ( u\|dt_\|b)^2 = g_w^4 n_c^2 ( 1+)(1 + ) , [ wstarpbarup ] where ^2 + ^2 and @xmath34 . the spin down amplitude vanishes in this case . on the other hand , when the @xmath28 is supplied by the proton instead , we obtain ( \|dut_\|b)^2 = g_w^4 n_c^2 [ wstarpbarpu ] for spin up , and ( \|dut_\|b)^2 = g_w^4 n_c^2 [ wstarpbarnd ] for spin down . thus , the presence of a @xmath28 sea in the proton introduces a small quantity of spin down top quarks into the sample . indeed , this contribution is dominated by the spin down component . however , the smallness of this contribution still results in a sample in which the top spin is aligned with the antiproton direction in the top quark rest frame 98% of the time . we now compare these results to the helicity basis , using the zmf as the frame in which we measure the helicity . we begin with the case where the @xmath28 quark comes from the antiproton , for which the matrix element squared is ( u\|dt_\|b)^2 = 12 ( 1-)(1-^2 ) [ wstarhelup ] for the production of spin up ( right - handed helicity ) top quarks and ( u\|dt_\|b)^2 = 12 ( 1+)(1+)^2 [ wstarheldn ] for the production of spin down top quarks . the expressions for the @xmath35 intial state may be obtained by making the replacement @xmath36 . the spin up amplitude is proportional to @xmath37 , causing it to vanish in the ultra - relativistic limit . at more moderate values of @xmath31 , such as are dominant at the tevatron , both spins are produced , with spin down ( left - handed helicity ) top quarks predominating . we find that in the over - all mixture at the tevatron , 83% of the top quarks have left - handed helicity . table [ wstarpurities ] summarizes the purities for the helicity , antiproton , and for completeness , proton bases . the proton basis is defined by choosing @xmath38 , _ i.e. _ the parton donated by the proton beam . ) and ( [ wstarpbarnd ] ) apply to the @xmath39 initial state once the replacement @xmath40 is made . likewise , eq . ( [ wstarpbarup ] ) represents the lone non - vanishing contribution from the @xmath35 initial state after making the same replacement . ] also included are the values of the spin asymmetries [ asym ] for each basis , as this is the coefficient which determines the magnitudes of the angular correlations . thus , the improvement from 83% left - handed helicity to 98% spin up in the antiproton basis translates into a factor of 1.45 increase in the size of the correlations . we plot the differential distributions in top quark @xmath41 for the helicity and antiproton bases as well as the total in fig . [ wstarplot ] . the dominant production mechanism for single 175 gev top quarks at the tevatron is the so - called @xmath8 fusion process . we consider the processes ( [ wgfusionprocess ] ) and hence the gauge - invariant set of diagrams shown in fig . [ wgfusiondiagram ] . once again we use the symbol for each particle to represent its momentum . for convenience , we employ the explicitly crossing - symmetric form in which all momenta are outgoing . because @xmath8 fusion is a @xmath42 process , the polarized production matrix elements squared for an arbitrary spin axis are too complicated to reproduce here@xcite . however , the sum over all spins and colors may be simply written as momentum is timelike , one should add the standard width term to the @xmath2 propagator . ] ( 0\|udgt\|b)^2 = g_w^4g_s^2n_c(n_c^2 - 1 ) ( u , d;t , b ) + z(d , u;b , t ) , where ( u , d;t , b ) = ( 2 dt ) \ { tu - u(b+g ) } . we present the relative contributions to the cross section from each of the partonic initial states for this process in table [ wgfusionpartons ] . as expected from the observation that the proton contains two @xmath33 quarks while the antiproton contains only one @xmath28 quark , the @xmath43 initial state gives the largest contribution ( 74% ) to the total . in terms of the helicity basis ( defined in the zero momentum frame of the incoming partons ) , we find that approximately 83% of the top quarks have negative helicity ( see table [ wgfusionpurities ] ) , leaving significant room for improvement . to illuminate our improved basis , we present the matrix element squared for the production of spin down top quarks in the basis where the spin axis is chosen to coincide with the @xmath0 quark direction : ( 0\|udgt_\|b)^2 = g_w^4 g_s^2 n_c(n_c^2 - 1 ) . besides being surprisingly simple , this result is significant in that it comes exclusively from the lower diagram in fig . [ wgfusiondiagram ] ; hence , there are no inverse powers of @xmath44 from the @xmath16-quark propagator . as is well - known @xcite , in the limit of vanishing @xmath16-quark mass , the @xmath8 fusion process develops a collinear singularity . for the physical ( non - zero ) value of the @xmath16 mass , this is reflected in the tendency for the @xmath16 quark to be produced at large pseudorapidity . thus , the majority of the total rate comes from the regions of phase space where @xmath44 is small : hence the spin down component ( no pole in @xmath44 ) is suppressed relative to the spin up component . in fact , for the @xmath43 and @xmath45 partonic initial states , we find that 97% of the tops are produced with spin up in this basis . since for the @xmath43 and @xmath45 initial states the @xmath0 quark becomes the spectator jet , we define the `` spectator '' basis by electing to use the direction of the spectator jet ( defined as the light jet appearing in the @xmath46 final state ) for the spin axis . although this picks the wrong spin axis direction for the @xmath47 and @xmath48 initial states , it is correct the majority of the time . we find that the overall composition consists of 96% spin up top quarks in this basis . for comparison , we give the results for the proton and antiproton bases in table [ wgfusionpurities ] . in terms of the spin asymmetry defined in eq . ( [ asym ] ) , we see that the spectator basis represents a factor of 1.36 improvement over the helicity basis . the differential distributions in top quark @xmath41 for the helicity and antiproton bases as well as the total appear in fig . [ wgfusionplot ] . in this paper we have found that the direction of the @xmath0-type quark provides the most effective spin axis for all single top production mechanisms . however , experimentally we do not know with certainty which physical object comprises the @xmath0-type quark in a given event . we have chosen the object which is most likely to be the @xmath0-type quark . in the case of the @xmath3 production mechanism , this means the direction of the antiproton beam , since it supplies the @xmath28 quark 98% of the time at the tevatron . using the antiproton as our basis , we find that the top quark is 98% spin up . as a result , the angular correlations with this choice of spin axis are 45% larger than those using the helicity basis . for @xmath8 fusion , the situation is potentially more complicated . nearly three - quarters of the cross section comes from the situation where the proton donates a @xmath33 quark : hence the @xmath0 quark appears as the spectator jet in the final state . in double - tagged events , identifying this jet is trivial ; in other cases , it may be necessary to assume that the jet with the largest pseudorapidity is the spectator jet . although a full simulation is beyond the scope of this paper , it is clear that this identification can be achieved with a small error rate because of the unique kinematics of this process . using the spectator jet as our basis , we find that the top quark is 95% spin up and that the angular correlations are 36% larger than the correlations using the helicity basis . we have demonstrated that the helicity basis is _ not _ the optimal basis for the discussion of angular correlations in single top quark production at the tevatron . instead , we have shown that the direction of the @xmath0-type quark provides a superior spin axis for _ all _ single top production mechanisms . the fermi national accelerator laboratory is operated by universities research association , inc . , under contract de - ac02 - 76cho3000 with the u.s . department of energy . high energy physics research at the university of michigan is supported in part by the u.s . department of energy , under contract de - fg02 - 95er40899 . gm would like to thank tony gherghetta for useful discussions related to this work . for the @xmath1 production process of single top quarks it is instructive to study the full matrix element including both the production and decay of the the top quarks to see how the top spin correlation translates into angular correlations in the events . consider the production of a top quark via u \|d t \|b and its subsequent semi - leptonic decay t \| b. using the symbol of the particle to represent its momentum , the full matrix element squared for this process including all correlations between production and decay , summed over all colors and spins , is given by ( u \|d t \|b \| b \|b)^2 = & & 2 n_c^2 g_w^8 ( 2u \|b ) ( 2b ) \{2(t \|d)(t \| ) - t^2 ( \|d \| ) } + & & [ ( 2ud - m_w^2)^2 + ( m_w_w)^2]^-1 ^-1 + & & [ ( 2\| -m_w^2)^2 + ( m_w_w)^2]^-1 [ wstarall ] if we use the narrow width approximation for the top quark , then the quantity in the curly brackets in eq . ( [ wstarall ] ) evaluated in the top quark rest frame is equal to m_t^2 e_\|d e_\| ( 1+_\|d\| ) where @xmath50 is the energy of the _ i_th particle and @xmath51 is the angle between the @xmath28-quark and the charged lepton in this frame . the @xmath52 is precisely the correlation expected if the top quark spin is along the direction of the @xmath28-quark momentum in the top quark rest frame . this is confirmation of eqs . ( [ wstargenup ] ) and ( [ wstargendn ] ) and discussion that follows in section [ wstarsection ] . f. abe , _ et . 74 * , 2626 ( 1995 ) ; s. abachi , _ et . 74 * , 2632 ( 1995 ) ; j. lys for the cdf collaboration , talk at the _ 28th international conference on high energy physics , _ warsaw , poland , 2531 july 1996 , fermilab conf96/409e ; e.w . varnes for the d collaboration , talk at the _ 1996 meeting of the divison of particles and fields , _ minneapolis , minnesota , 1015 august 1996 , fermilab conf96/243-e . .fractional cross sections for single top production in the @xmath3 channel at the tevatron at 2.0 tev , decomposed according to the parton content of the initial state . we use the mrs(r1 ) structure functions @xcite evaluated at the scale @xmath53 . we obtain a total cross section of approximately 0.33 pb . [ wstarpartons ] [ cols="^,^,^,^,>,^,^",options="header " , ] ccccdcc & & basis & spin content & @xmath54 & & + & & helicity & 83% @xmath55(l ) & @xmath560.67 & & + & & proton & 68% @xmath57 & 0.37 & & + & & antiproton & 54% @xmath55 & @xmath560.07 & & + & & spectator & 96% @xmath57 & 0.91 & &	we show in single top quark production that the spin of the top quark is correlated with the direction of the @xmath0-type quark in the event . for single top production in the @xmath1 channel , the @xmath0-type quark comes dominantly from the antiproton at the tevatron , whereas for the @xmath2-gluon fusion channel the spectator jet is the @xmath0-type quark the majority of the time at this machine . our results are that 98% of the top quarks from the @xmath3 process have their spins in the antiproton direction , and 96% of the top quarks in the @xmath2-gluon fusion process have their spins in the spectator jet direction . we also compare with the more traditional , but less effective , helicity basis . the direction of the top quark spin is reflected in angular correlations in its decay products .
there is considerable interest in determining whether jovian planets on wide orbits represents a continuum that extends to brown dwarf masses , or whether there is a strong cutoff in the number of companions as a function of mass ( e.g. , * ? ? ? this relates to possible formation pathways for substellar companions : either companions form within a circumstellar disk and reach a mass above the deuterium burning limit ( e.g. , @xcite ) or via cloud fragmentation , as in binary systems with a high mass ratio ( @xmath10 , @xcite ) . population statistics from direct imaging provide essential observational parameters to test formation history . from numerous surveys , only a handful of imaged substellar companions are @xmath11100 au from their host stars . in particular , this separation regime has shown a lack of brown dwarfs with @xmath12 ( the `` brown - dwarf desert '' , e.g. , @xcite ) around stars with m@xmath13m@xmath14 . however , this parameter space has recently begun to be populated by direct imaging ( e.g. , * ? ? ? * ; * ? ? ? since the contrast of a substellar object is more favorable for imaging with youth , direct imaging surveys tend to target sources with evidence of a relatively young age ( @xmath15 myr ) . the presence of a debris disk , leftover of planet formation , is an insight for a younger - than - field age since the dust luminosity is known to decrease with time ( see @xcite ) . however , the dust can persist to longer time if small planetesimals are dynamically perturbed by orbiting companion , hence other youth indicators are necessary to constrain the age of a star . an accurate estimate of the age of the star is necessary since companion properties are often derived from evolutionary models . membership in nearby young associations ( e.g. , * ? ? ? * ) provides the tightest constraints on the age of a star ; otherwise , large variation in the derived companion masses exists ( e.g. , * ? ? ? * ) . among the brown dwarf - mass companions that have been discovered around stars with infrared excess , none have been previously seen inside the inner hole of a resolved disk , which offers the opportunity to study dynamical interactions . using the gemini planet imager ( gpi , @xcite ) , we report the discovery of a companion to the debris disk host hr 2562 . hr 2562b has a projected separation within the `` brown - dwarf desert '' , and within a possible cleared inner hole . we derive properties for this companion based on spectra and colors , and discuss the potential roll it plays in maintaining and shaping the debris disk . image is from february 2016 and demonstrates two possible solutions for the inner edge of the disk ( 38 and 75 au with dashed and dotted - dashed lines respectively ) assuming inclination of @xmath16 and position angle of @xmath17 . ] image is from february 2016 and demonstrates two possible solutions for the inner edge of the disk ( 38 and 75 au with dashed and dotted - dashed lines respectively ) assuming inclination of @xmath16 and position angle of @xmath17 . ] image is from february 2016 and demonstrates two possible solutions for the inner edge of the disk ( 38 and 75 au with dashed and dotted - dashed lines respectively ) assuming inclination of @xmath16 and position angle of @xmath17 . ] image is from february 2016 and demonstrates two possible solutions for the inner edge of the disk ( 38 and 75 au with dashed and dotted - dashed lines respectively ) assuming inclination of @xmath16 and position angle of @xmath17 . ] hr 2562 is a f5v star with an estimated mass of 1.3 m@xmath8 ( @xcite ) . it has a distance of 33.63@xmath00.48 pc and proper motion of @xmath18 110 mas / yr @xcite . it was identified as having a debris disk with data from _ iras _ and _ spitzer _ by @xcite . @xcite identify the source as active based on the ca ii h and k lines , while @xcite identify it as an x - ray source . several groups have also computed metallicity estimates , which range from [ m / h ] = [ -0.05,+0.08 ] @xcite , with possible evidence for peculiar individual abundances @xcite . currently there are several disparate age estimates for hr 2562 in the literature . an estimate of 300@xmath0120 myr was made by @xcite , who used a combination of space motions and evolutionary model - derived ages to suggest that the source is part of a nearby local association subgroup called b3 . in their identification of the source using _ iras _ , @xcite use space motions , a lithium non - detection , and x - ray luminosity to give a rough age estimate of @xmath18300 myr . conversely , analysis of data from the geneva - copenhagen survey , in which metallicities and temperatures are used to derive ages from models , gives an estimated age of 0.9 - 1.6 gyr ( @xcite ) . in their assessment of age based on ca ii h and k lines , @xcite also derive an age of @xmath18900 myr . most recently , @xcite used photometric modeling with atmosphere models to derive an age range of 300@xmath19 myr . the banyan ii group / field membership estimation code @xcite gives low probabilities for the star to belong to any known young nearby group , but suggests the star is younger than field stars , giving a wide range of possible ages between @xmath1820 myr to @xmath181 gyr . while the age of the star remains uncertain , sufficient evidence of moderate youth led to the inclusion of hr 2562 in the sample for the gemini planet imager exoplanet survey ( gpies ) . for the purposes of this paper , we adopt a nominal age range of 300900 myr . hr 2562 was observed in january 2016 . gpies observations are taken in angular differential imaging ( adi , @xcite ) @xmath4 band spectroscopic mode . a candidate companion was identified in this initial data set . follow - up observations were made within a month in the @xmath20 , @xmath21 , and @xmath3 bands . sky frames were also obtained right after the @xmath21 sequence . table 1 gives the log of these observations . weather conditions were median with dimm seeing around @xmath22 when available . another ( longer ) @xmath21 sequence was acquired to provide higher a signal - to - noise ratio for the companion and was used for spectroscopy . all data were acquired with the @xmath4-band apodizer providing a near - ir constant star - to - satellite - spot ratio of @xmath23 mag @xcite . astrometric calibrator observations were obtained in january and february 2016 , and were analyzed as in @xcite . these observations showed no change in the ifs calibration as measured in previous gpies observations . therefore , as in @xcite , pixel scale and position angle offset of @xmath24 mas / px and @xmath25 deg were used . data were processed using the gpi data reduction pipeline version @xmath26 ( @xcite , and references therein ) to obtain calibrated @xmath27 datacubes . further processing to suppress the star point spread function was performed as described in @xcite and @xcite , using four independent pipelines and several adi algorithms : cadi @xcite , tloci @xcite , and klip / pyklip @xcite . the post - processed cubes were then combined to create broad - band images , examples of which are shown in figure [ fig : images ] . from each of these pipelines , positions and contrast - per - slice and associated errors were extracted following the forward modeled and minimization techniques described in @xcite , and @xcite . the final astrometric errors were combined in quadrature from the errors on the measurements ( @xmath28 px depending on the dataset ) , astrometric calibration , and star center ( @xmath29 px ) . for the spectroscopy , the errors on the measurements and on the star - to - spot ratio were similarly combined . final values for astrometry and contrasts for the companion are derived by averaging the results of each pipeline . table 1 gives the astrometry and photometry derived from each observation set . the spectra of the companion were then obtained after normalization with a calibrated template f5v spectrum from the pickles library @xcite , the 2mass @xmath30 magnitudes of the host star and the gpi response functions . figure [ fig : spectrum ] shows the final spectrum from each bandpass . lcccccccc [ tab : log ] 2016 jan 25 & @xmath4 & 45 & 37 & 20.2 & 619@xmath03 & 297.56@xmath00.35 & 11.7@xmath00.1 & 14.2@xmath00.1 + 2016 jan 28 & @xmath20 & 65 & 23 & 11.9 & 618@xmath05 & 297.40@xmath00.25 & 10.6@xmath00.1 & 13.0@xmath00.1 + 2016 jan 28 & @xmath21 & 75 & 24 & 11.3 & 618@xmath04 & 297.76@xmath00.37 & 10.4@xmath00.1 & 12.8@xmath00.1 + 2016 feb 25 & @xmath21 & 75 & 47 & 25.7 & 619@xmath02 & 297.50@xmath00.25 & 10.4@xmath00.1 & 12.8@xmath00.1 + 2016 feb 28 & @xmath3 & 35 & 54 & 26.6 & 620@xmath03 & 297.90@xmath00.25 & 12.6@xmath00.1 & 15.3@xmath00.1 + at the distance of hr 2562 , the projected separation of the companion is 20.3@xmath00.3 au ( @xmath31 ) . we assessed whether the candidate companion is comoving with the star . we compared all relative astrometric measurements with the predictions of the location of an infinitely far background object . the companion being an infinitely far background object is ruled out to 5@xmath32 @xcite . additionally , as in @xcite , we estimate likely orbit tracks for bound objects . the results of this assessment are shown in figure [ fig : astrometry ] . the astrometry falls clearly in the realm of bound orbits , with a preferred range of semi - major axes of @xmath181542 au . archival hubble space telescope nicmos data from 2007 was found for hr 2562 ( pi rhee , i d 11157 ) . these data were processed as part of the archival legacy investigation of circumstellar debris disks ( alice , @xcite ) . although the contrast achieved in these images is insufficient to detect the companion at its present location , we searched for point sources at the location the companion would have been if it was a background object ( @xmath181.3 ) . no source is detected at this location . as we demonstrate in section [ spec ] , the companion photometry and spectroscopy are inconsistent with an infinitely far background star . a more likely source of contamination might instead be a non - stationary foreground or slightly background l / t dwarf . following the methodology in @xcite , we determine the probability of finding a l5-t5 type object in the gpi field of view by combining l and t dwarf space densities and absolute magnitudes @xcite . we calculate a false alarm probability of 7@xmath33 within the gpi field of view for l5-t5 dwarfs . since the companion was found after observing 203 targets , the final false alarm probability is 1.4@xmath34 . therefore , the chance alignment of a non - stationary , unbound brown dwarf is unlikely and the bound status is more probable . photometric and spectral comparisons were made to assess the likely properties of hr 2562b . first , we used several libraries of spectra and routines to determine the best - matching spectral type for hr 2562b . we used the splat toolkit , which makes use of the spex prism library , to determine which spectral types match hr 2562b @xcite . we find that the best - matching source comparing all three spectral bands simultaneously is wise j174102.78 - 464225.5 ( wise 1741 - 4642 ) , a recently discovered peculiar l7@xmath02 type brown dwarf with an estimated age of 10 - 100 myr @xcite . classifying the hr 2562b using spectral standards in splat returns a spectral type of l7@xmath00.5 . in a separate analysis , we used a @xmath35 goodness of fit test with data from the spex prism library supplemented by spectra from @xcite . the @xmath35 fits were produced by normalizing the empirical spectra to hr 2562b with a constant based on the flux and uncertainties of both spectra , following @xcite . the @xmath35 values were then calculated between the empirical spectrum and hr 2562b and assessed as a function of spectral type . in this analysis , we considered all three bands simultaneously and separate fits to each of the bands individually . we find that sources with spectral types between l3.5 and t2 provide the best fits to the data , depending on the band . the best fit to the j - band is a t2 type object , while l3.5 and l4.5 sources best match @xmath4 and @xmath5 , respectively . the simultaneous fit to all bands returns the same best fit as splat , wise 1741 - 4642 . in our analysis of @xmath35 as a function of spectral type , we find that while earlier spectral types are preferred at h and k , mid - to - late l types have nearly equivalent @xmath35 . we also find that two other young l / t transition objects , vhs j125601.92 - 125723.9b ( vhs 1256b , @xcite ) and pso j318.5338 - 22.8603 ( pso 318 , @xcite ) are reasonable matches to the spectra in individual bands . when considering the bands simultaneously , it is clear that the overall flux in each wavelength is not perfectly matched by any other object , including the best - fit wise 1741 - 4642 . however , brown dwarfs can have similar spectral features despite varying flux in different wavelength bands ( i.e. , a range of colors ) , as shown , for example , by @xcite and @xcite . to investigate this , @xcite created band - normalized templates from optically - classified field l0-l8 and l0@xmath36-l4@xmath36 objects and found that objects with a range of @xmath37 colors as large as 0.60 can match band - normalized templates with average @xmath380.9 . we therefore adopt a spectral type of l7@xmath03 for hr 2562b . figure [ fig : spectrum ] shows the spectrum of hr 2562b , along with the spectra of the best fitting objects in each band and other similar young objects . we then compare the colors of hr 2562b with other objects in color magnitude diagrams ( see figure [ fig : cmd ] ) . the source clusters with the sequence of young , red , l - type objects in the @xmath39 vs @xmath37 and @xmath40 vs @xmath41 diagrams . its @xmath41 is similar to vhs 1256b and pso 318 . it is somewhat bluer in @xmath37 , though still near pso 318 in absolute magnitude at @xmath39 . its consistency with young l type objects is an additional suggestion of youth ( e.g. * ? with the mlt sequence of 180 field objects ( light grey symbols ) from @xcite . young brown dwarfs or directly imaged companions are also shown ( black symbols ) , and a few peculiar objects ( specific symbols ) from @xcite , gagn et al . ( in prep . ) . ( bottom right ) : the j band absolute magnitude as a function of spectral type . the grey line shows the field sequence with the @xmath42 spread marked by the grey shaded area following @xcite . ] ) with the mlt sequence of 180 field objects ( light grey symbols ) from @xcite . young brown dwarfs or directly imaged companions are also shown ( black symbols ) , and a few peculiar objects ( specific symbols ) from @xcite , gagn et al . ( in prep . ) . ( bottom right ) : the j band absolute magnitude as a function of spectral type . the grey line shows the field sequence with the @xmath42 spread marked by the grey shaded area following @xcite . ] ) with the mlt sequence of 180 field objects ( light grey symbols ) from @xcite . young brown dwarfs or directly imaged companions are also shown ( black symbols ) , and a few peculiar objects ( specific symbols ) from @xcite , gagn et al . ( in prep . ) . ( bottom right ) : the j band absolute magnitude as a function of spectral type . the grey line shows the field sequence with the @xmath42 spread marked by the grey shaded area following @xcite . ] ) with the mlt sequence of 180 field objects ( light grey symbols ) from @xcite . young brown dwarfs or directly imaged companions are also shown ( black symbols ) , and a few peculiar objects ( specific symbols ) from @xcite , gagn et al . ( in prep . ) . ( bottom right ) : the j band absolute magnitude as a function of spectral type . the grey line shows the field sequence with the @xmath42 spread marked by the grey shaded area following @xcite . ] we can also compare absolute photometry to other brown dwarfs in both the field and young moving groups . the variation of j band magnitude with spectral type has been commonly used as a method of distinguishing between field and potentially young objects ( e.g , * ? ? ? * ) , with younger l - type objects tending to be fainter at @xmath3 band than older objects of the same spectral type . this is thought to be a natural consequence of changes in surface gravity with age impacting atmospheric properties such as clouds ( e.g. , @xcite ) . figure [ fig : cmd ] shows the @xmath3 band magnitude variation with spectral type . at a spectral type of l7 , hr 2562b clusters strongly with members of nearby moving groups below the field population . following the methods of @xcite , we constructed a spectral energy distribution ( sed ) using the spectra and photometry to determine an empirical @xmath6(l@xmath7/l@xmath8 ) of @xmath434.62@xmath00.12 . using this value and an age range of 300900 myr , we then use the evolutionary models from @xcite to estimate the physical properties of hr 2562b ( solar metallicity , hybrid cloud ) . we find a mass of 30@xmath015 m@xmath9 , a radius of 1.11@xmath00.11 r@xmath44 , a @xmath45 of 4.70@xmath00.32 , and a temperature of 1200@xmath0100 k. this temperature is somewhat lower than field l7 type objects , but is consistent with estimates for young objects of this spectral type . the identification of a brown dwarf in a system with a debris disk presents interesting opportunities for constraining system properties . from _ pacs , the disk is marginally resolved . @xcite use this data to derive an average dust radius of 112.1@xmath08.4 au , with evidence for an inner hole of radius between @xmath1818 - 70 au . the average outer radius is found to be 187 au . interestingly , they find that the disk has a high inclination of 78.0@xmath06.3@xmath46 and a position angle on sky of 120.1@xmath03.2@xmath46 . given the separation of hr 2562b of @xmath1820 au and average position angle of @xmath18298@xmath46 , the source appears to be interior to the hole in the disk and coplanar with the disk to within the uncertainties ( see figure [ fig : images ] ) . this is the first example of a brown dwarf - mass companion within the inner hole of a debris disk . the existence of a companion interior to the disk provides the exciting possibility to independently constrain the properties of hr 2562 system . @xcite discuss the possibility of a self - stirring mechanism generating the resolved disk material for a source as old as hr 2562a . in this mechanism , secular perturbations from a companion generate enhanced collisions of smaller planetesimals , leading to the generation of dust at wide separations @xcite . using equation 6 in @xcite and assuming their nominal disk parameters , we find that a mass of only 13m@xmath9 would be required to generate collisions out to @xmath18187 au in 900 myr assuming an eccentricity of 0.01 . if we use 30 m@xmath47 for hr 2562b , we derive a crossing time of @xmath18385 myr , consistent with the lower end of our adopted age range . while the uncertainties in the outer radius of the disk and mass of the planet allow for crossing times @xmath481 gyr , the nominal parameters are consistent with the scenario that hr 2562b is responsible for generating the observed disk . a more complicated question is whether the inner hole can also be used to place an upper limit on the mass of the companion . though @xcite derive an average inner radius of 38 @xmath0 20 au from _ herschel _ images , the uncertainty on this value is fairly large . @xcite also performed a separate sed fit to available photometry and derived an inner radius of 64 @xmath0 6 au . since these two values are not consistent , we performed a quick analysis in which we simultaneously fit the sed and the _ herschel _ pacs image using mcfost @xcite . we used the geometric parameters from @xcite , a flat surface density profile , and a minimum grain size of @xmath18micron . we find that the sed and image together are best reproduced using an inner radius of @xmath1875 au . this radius is consistent with upper end of the uncertainty range and sed fit of @xcite . for completeness , we use both an inner radius of 38 au and our derived value of 75 au to roughly determine whether mass constraints are possible . using the dynamical stability criterion proposed by @xcite , we estimate the mass of the hr 2562b assuming it is responsible for clearing the inner hole and that it has an eccentricity of zero . we find that if the inner hole is 38 au , the upper limit on the mass of the companion is @xmath1820 m@xmath9 . the difference between an age of 300 myr and 900 myr is negligible given our other assumptions . if instead we use an inner radius of 75 au , the upper limit on the mass is @xmath180.24 m@xmath8 , well beyond the highest estimates from evolutionary models . if the inner radius is indeed @xmath4875 au , it might suggest an elevated eccentricity for hr 2562b . future observations that constrain the orbital parameters of the companion and high resolution images of the disk will offer insight into the potential history of interaction between the bodies in this system and provide meaningful mass limits . the hr 2562 system offers a relatively rare opportunity to probe the direct dynamical interaction of a substellar object with a kuiper belt analog . the overall system architecture may provide interesting clues to the formation of the companion . with a mass ratio of @xmath49 , hr 2562b is a new object in the growing list of substellar companions within @xmath50 au ( e.g. , * ? ? ? * ; * ? ? ? these are excellent candidates for formation via disk instability , which has been shown to naturally produce objects as massive as 42 m@xmath9 at separations @xmath51 au ( e.g. , * ? ? ? * ; * ? ? ? several challenges to this picture remain , however , particularly the proximity of observed objects to their host stars . at such close separations the fast cooling needed for the disk fragmentation into bound objects becomes difficult to realize , precluding in - situ formation of these brown dwarfs by the gravitational instability ( e.g. , * ? ? ? . however , it is plausible that the relatively massive hr 2562b formed beyond 50 - 70 au and migrated inward to its current location ( e.g. , * ? ? ? constraining the true mass and orbit of the companion is essential to determining its possible origin , which could offer evidence of `` planet '' formation above the deuterium burning limit . the authors thank richard gray for his clarifying points on spectral classification . we also thank adam burgasser and daniella bardalez - gagliuffi for helpful discussions . we also thank an an anonymous referee whose comments improved this manuscript . this research has benefited from the spex prism library and spex prism library analysis toolkit , maintained by adam burgasser at http://www.browndwarfs.org/spexprism , from the banyan ii web tool at http://www.astro.umontreal.ca/ gagne / banyanii.php?targetname = hr+2562&resolve = resolve , and from the simbad database , operated at cds , strasbourg , france . based on observations obtained at the gemini observatory , which is operated by the association of universities for research in astronomy , inc . , under a cooperative agreement with the national science foundation ( nsf ) on behalf of the gemini partnership : the nsf ( united states ) , the national research council ( canada ) , conicyt ( chile ) , the australian research council ( australia ) , ministrio da cincia , tecnologia e inovao ( brazil ) and ministerio de ciencia , tecnologa e innovacin productiva ( argentina ) . , r.d . and d.l . acknowledge support from the fonds de recherche du qubec . supported by nsf grants ast-1518332 ( r.j.d.r . , p.k . ) , ast-1411868 ( b.m . , k.w.d . ) , ast-141378 ( g.d . ) , ast-1211568 ( p.a.g . , e.l.r . ) , dge-1232825 ( a.z.g . ) , and ast-1313132 ( j.f.c . , e.l.r . ) . supported by nasa grants nnx15ad95g / nexss and nnx15ac89 g ( r.j.d.r . , t.m.e . ) , and nnx14aj80 g ( e.l.n . , m.p . ) . portions of this work were performed under the auspices of the u.s . department of energy by lawrence livermore national laboratory under contract de - ac52 - 07na27344 ( s.m.a . , d.p . ) .	we present the discovery of a brown dwarf companion to the debris disk host star hr 2562 . this object , discovered with the gemini planet imager ( gpi ) , has a projected separation of 20.3@xmath00.3 au ( @xmath1 ) from the star . with the high astrometric precision afforded by gpi , we have confirmed common proper motion of hr 2562b with the star with only a month time baseline between observations to more than @xmath2 . spectral data in @xmath3 , @xmath4 , and @xmath5 bands show morphological similarity to l / t transition objects . we assign a spectral type of l7@xmath03 to hr 2562b , and derive a luminosity of @xmath6(l@xmath7/l@xmath8)=-4.62@xmath00.12 , corresponding to a mass of 30@xmath015 m@xmath9 from evolutionary models at an estimated age of the system of 300900 myr . although the uncertainty in the age of the host star is significant , the spectra and photometry exhibit several indications of youth for hr 2562b . the source has a position angle consistent with an orbit in the same plane as the debris disk recently resolved with _ herschel_. additionally , it appears to be interior to the debris disk . though the extent of the inner hole is currently too uncertain to place limits on the mass of hr 2562b , future observations of the disk with higher spatial resolution may be able to provide mass constraints . this is the first brown dwarf - mass object found to reside in the inner hole of a debris disk , offering the opportunity to search for evidence of formation above the deuterium burning limit in a circumstellar disk .
q - balls are nontopological solitonic solutions of a self - interacting complex scalar field theory carrying a conserved global @xmath0 charge . introduced by coleman in 1985 @xcite , their properties have been extensively studied since then @xcite . if on the one hand , volkov and whnert @xcite showed , in the context of some theories with nonrenormalizable scalar potentials , that there exist particular q - ball configurations possessing nonvanishing angular momentum , now known as `` spinning q - balls '' , on the other hand , dvali , kusenko and shaposhnikov @xcite proved in the framework of supersymmetric extensions of the standard model that gauge - singlet combinations of squarks and sleptons corresponding to some flat direction of the supersymmetric potential can give rise to q - balls whose charge @xmath1 is some combination of baryon and lepton numbers . the attractive feature of theses `` supersymmetric q - balls '' is that they could represent the dark matter component of the universe @xcite ( for reviews on dark matter see , e.g. , ref . @xcite ) . motivated by this fact , experimental searches for q - balls are being carried out @xcite , although no compelling evidence for their existence has been reported so far . the aim of this paper is to present novel configurations of a charged scalar field describing nonspherically - symmetric supersymmetric q - balls with nonvanishing angular momentum . to our knowledge , _ spinning supersymmetric q - balls _ are the first example of analytical solution in field theory in minkowski spacetime representing a soliton possessing angular momentum . as we will see , spinning supersymmetric q - balls are excitations of spherically - symmetric supersymmetric q - balls , since their energy spectrum lies above the ground state represented by nonspinning q - balls . however , it is highly probable that these exited states could form during collisions between supersymmetric q - balls and , most importantly , during the process of fragmentation of the affleck - dine condensate , which is a plausible process that can lead to a copious production of q - balls at the end of inflation @xcite . the plan of the paper is as follows . in section ii we review the general properties of spherically - symmetric supersymmetric q - ball in order to make clearer the derivation and the study of nonspherically - symmetric supersymmetric q - balls configurations which will be tackled in section iii . in section iv we show that spinning supersymmetric q - balls are stable against small perturbations about their classical configurations . finally , in section v we draw our conclusions . let us consider a charged scalar field @xmath2 whose dynamics is described by lagrangian density @xmath3 being the theory invariant under a global @xmath0 transformation , there exists a conserved noether charge , @xmath4 , which we normalize as @xmath5 where a dot indicates a derivative with respect to time . the energy - momentum tensor associated to a given field configuration @xmath6 , reads @xmath7 so that the total energy is @xmath8 \!.\ ] ] in general , q - balls are solitonic solutions of the field equations carrying a definite value of the charge , let us say @xmath1 . an elegant way to construct such a type of solution @xcite is to introduce a lagrange multiplier @xmath9 associated to @xmath4 , and require that the physical configuration @xmath6 makes the functional @xmath10 \equiv e + \omega \left[q - \frac{1}{i } \int \ ! d^3x \left(\phi^ * \dot{\phi } - \phi \dot{\phi}^ * \right ) \right]\ ] ] stationary with respect to independent variations of @xmath2 and @xmath9 . noticing that the choice @xcite @xmath11 assures that the total energy @xmath12 is independent on the time , one finds that q - balls solutions have to satisfy the constraints @xmath13 the first two constraints lead to the equations of motion of the fields @xmath14 and @xmath15 , respectively , @xmath16 while the second one is equivalent to the requirement that the charge corresponding to the solution of the equation of motion is equal to @xmath1 : @xmath17 taking into account the equations of motion , the functional @xmath18 can be conveniently re - written as : @xmath19 + \omega q.\ ] ] a spherically - symmetric q - ball is defined as the solution @xmath20 of eq . satisfying , at fixed charge @xmath1 , the boundary conditions @xcite @xmath21 where @xmath22 . in particular , a ( spherically - symmetric ) supersymmetric q - ball is a q - ball configuration arising in a supersymmetric model of particle physics where supersymmetry is broken via low - energy gauge mediation @xcite . in this kind of model the coupling of the massive vector - like messenger fields to the gauge multiplets , with coupling constant @xmath23 , leads to the breaking of supersymmetry @xcite . the coupling itself gives rise to an effective potential for the flat direction @xmath24 whose lowest order ( two - loops ) contribution has been calculated in ref . @xcite : @xmath25}{[z^{-2}-x(1-x)]^2 } \ , .\ ] ] here , @xmath26 and @xmath27 , with @xmath28 the messenger mass scale . the value of the mass parameter @xmath29 is constrained as ( see , e.g. , ref . @xcite ) : @xmath30 where the gravitino mass , @xmath31 , is in the range @xmath32 @xcite . the asymptotic expressions of @xmath33 , for small and large @xmath34 are @xcite : @xmath35 a widely used approximation in constructing q - ball solutions , whose validity has been ascertained in ref . @xcite , consists in replacing the full potential @xmath33 with its asymptotic expansions ( [ potentialapprox ] ) in which a plateau plays the role of the logarithmic rise for large values of @xmath34 . more precisely , the approximate supersymmetric potential has the form @xmath36 where @xmath37 is the soft breaking mass and is of order @xmath38 @xcite . within this approximation , it has been shown that the potential @xmath39 allows spherically - symmetric q - ball solutions as the nonperturbative ground state of the model @xcite . the profile of the supersymmetric q - ball is easily found from eqs . and : @xmath40 where @xmath41 and @xmath42 are constants of integration , @xmath43 is the zeroth - order spherical bessel function of first kind , @xmath44 is the zeroth - order spherical bessel function of second kind , @xmath45 is the zeroth - order modified spherical bessel function of first kind , and @xmath46 is the zeroth - order modified spherical bessel function of second kind @xcite . here , we have introduced the `` thickness '' of the q - ball , @xmath47 we can now define the `` radius '' of the q - ball , @xmath48 , as the solution of the equation @xmath49 . in the limit @xmath50 we have @xmath51 , from which it follows that @xmath52 ( we will se in the following that the condition @xmath50 will correspond to have large values of the charge @xmath1 . ) if the thickness of the q - ball is much smaller than its radius ( we will see , below , that indeed large charges @xmath1 implies that @xmath53 ) , we can write @xmath54 which is the solution found in ref . @xcite . inserting the above solution in eq . and minimizing with respect to @xmath9 [ see last equation in eq . ] , we find the parameter @xmath9 as a function of the charge @xmath1 : @xmath55 where we have introduced the `` critical charge '' @xmath56 , whose meaning will be clear in the following , as @xmath57 inserting eq . in eq . , we find the q - ball radius as a function of the charge : @xmath58 from the above equation and taking into account eqs . and , we find that for large charges , @xmath59 , it results @xmath60 , and this justifies our approximation to neglect the thickness of the q - ball in computing its profile . inserting eq . in eq . , and observing that , at fixed charge @xmath1 , the energy coincides with the functional @xmath18 , we find @xmath12 as a function of the charge : @xmath61 finally , inserting eq . in eq . and taking into account eq . , we find the value of @xmath41 as a function of the charge : @xmath62 the above relation clarifies the meaning of the critical charge : the q - ball solution we found in the limit @xmath50 , corresponds indeed to the case of large charges compared to @xmath56 . if the energy @xmath12 of the q - ball at fixed charge @xmath1 is less then @xmath63 , the soliton decays into @xmath1 quanta of the field ( the perturbative spectrum of the theory ) , each of them with mass @xmath64 . instead , if @xmath65 the q - ball is said to be classically stable , and then represents the ground state of the theory . using eq . , we find classical stability , @xmath66 , for @xmath67 , with @xmath68 in general , spherically - symmetric q - balls have zero angular momentum . in fact , the total angular momentum for a scalar field configuration is given by @xmath69 where @xmath70 is the total angular momentum tensor @xcite @xmath71 with @xmath72 being the energy - momentum tensor given by eq . . now , using spherical coordinates , @xmath73 , we obtain @xmath74 \ ! , \\ \label{j2 } & & \!\!\!\!\!\!\!\!\!\ ! } = \int \ ! d^3x \ ! \left [ -\cos(2\theta ) \cos \varphi \ , t^0_\theta + \cot \ ! \theta \sin \varphi \ , t^0_\varphi \right ] \ ! , \\ \label{j3 } & & \!\!\!\!\!\!\!\!\!\ ! j^{3 } = - \int \ ! d^3x \ , t^0_\varphi \ , , \end{aligned}\ ] ] so that for a spherically - symmetric q - ball , @xmath75 , we get @xmath76 . on the other hand , for a nonspherically - symmetric q - ball ( if it ever exists ) , @xmath77 , we could have in principle a nonvanishing angular momentum ( for supersymmetric q - balls this will be indeed the case ) . in particular , if one makes use of the `` axially - symmetric ansatz '' , @xmath78 @xmath79 being a real constant , one easily finds @xmath80 where @xmath4 is given by eq . . since single - valuedness of the scalar field requires @xmath81 , the constant @xmath79 must be an integer . therefore , for this particular configuration , the third component of the angular momentum is quantized and proportional to the charge . it is useful for the following discussion to observe that if a q - ball configuration is such that the field @xmath2 is given by @xmath82 with @xmath83 a real function , then ( as it easy to verify ) it results @xmath84 , so that the nonspherically - symmetric q - ball is indeed a spinning q - ball with total angular momentum directed along the @xmath34-axis . we now return to the supersymmetric case to find nonspherically - symmetric supersymmetric q - ball configurations . we start by writing the equation of motion for the field @xmath15 [ eq . ] in spherical coordinates : @xmath85 using the technique of separation of variables , @xmath86 we easily find [ using the approximate form of the supersymmetric potential , eq . ] the solution of eq . : @xmath87 where @xmath88 are the usual spherical harmonics of degree @xmath89 and order @xmath79 , with @xmath90 being the associated legendre polynomials of degree @xmath89 and order @xmath79 @xcite , and @xmath91 here , @xmath92 , @xmath93 , @xmath42 , @xmath94 are constants of integration , the q - ball thickness @xmath95 is the same as in eq . , @xmath96 and @xmath97 are the spherical bessel function of order @xmath89 of first and second kind respectively , and @xmath98 and @xmath99 are the modified spherical bessel function of order @xmath89 of first and second kind respectively @xcite . we can now define the `` radius '' of the q - ball , @xmath100 , as the solution of the equation @xmath101 . in the limit @xmath102 we have @xmath103 , from which it follows that @xmath104 where @xmath105 represents the first zero of the bessel function of order @xmath106 of first kind , @xmath107 @xcite . is an increasing function of @xmath106 with @xmath108 and @xmath109 , @xmath110 , @xmath111 , @xmath112 , @xmath113 @xcite , etc . moreover , the asymptotic expansion of @xmath105 , as @xmath114 , is : @xmath115 , where @xmath116 is the first negative zero of the airy function @xmath117 @xcite . ] ( we will se in the following that the condition @xmath102 will correspond to have large values of the charge @xmath1 . ) if the thickness of the q - ball is much smaller than its radius ( we will see , below , that indeed large charges @xmath1 implies that @xmath118 ) , we can write @xmath119 inserting the above solution in eq . and minimizing with respect to @xmath9 , we find the parameter @xmath9 as a function of the charge @xmath1 : @xmath120 where , from now on , quantities with the subscript `` 0 '' refer to the case of spherically - symmetric supersymmetric q - balls analyzed in section ii . inserting eq . in eqs . and we find , respectively , the q - ball radius and energy as a function of the charge : @xmath121 and @xmath122 while , inserting eq . in eq . and taking into account eq . , we obtain @xmath123 where we have introduced the `` critical charge '' @xmath124 as @xmath125 ^ 4 \ , q^{\rm ( cr)}_0.\ ] ] therefore , the q - ball solution we found in the limit @xmath102 , corresponds indeed to the case of large charges compared to @xmath124 . also , from eq . and taking into account eqs . and , we find that for large charges , @xmath126 , it results @xmath127 , and this justifies our approximation to neglect the thickness of the q - ball in computing its profile . finally , using eq . , we find classical stability , @xmath128 , for @xmath129 , with @xmath130 taking into account that the general form of a supersymmetric q - ball solution [ given by eqs . , - , and ] is of the form with @xmath83 a real function , and taking into account the discussion at the beginning of this section , we conclude that the nonspherically - symmetric q - ball solutions we found describe q - balls with total angular momentum directed along the @xmath34-axis and equal to @xmath131 . moreover , observing that the angular part of the q - ball solution is proportional to the spherical harmonic @xmath132 , we deduce that @xmath89 determines its parity @xmath133 : @xmath134 however , not all values of @xmath89 are admitted since , at fixed charge @xmath1 , a q - ball with definite angular momentum @xmath135 ( or , which is the same , with definite value of @xmath79 ) is such that its energy is minimum . looking at eq . and taking into account that @xmath136 is an increasing function of @xmath89 , we deduce that at fixed @xmath1 and @xmath79 , two values of @xmath89 are allowed : for even ( odd ) @xmath137 , @xmath138 if the parity of the q - ball solution is positive ( negative ) and @xmath139 if the parity of the q - ball solution is negative ( positive ) . accordingly , the energy spectrum of allowed states of a spinning supersymmetric q - ball looks like that in fig . 1 . . also shown are the corresponding values of the radius @xmath100 , the angular momentum @xmath135 , and the parity @xmath133 of the state.,scaledwidth=45.0% ] in fig . 2 , we plot the spinning supersymmetric q - ball s profile , @xmath140 , as a function of @xmath141 and @xmath142 for different values of @xmath89 and @xmath79 , at fixed charge @xmath143 . we observe that , using well - known properties of spherical harmonics , the profiles for negative values of @xmath79 coincide with the corresponding positive ones if @xmath137 is even , while they get an extra minus sign if @xmath137 is odd . , for @xmath143 . from upper to lower panel : @xmath144 , @xmath145 , @xmath146.,scaledwidth=45.0% ] , for @xmath143 . from upper to lower panel : @xmath144 , @xmath145 , @xmath146.,scaledwidth=45.0% ] , for @xmath143 . from upper to lower panel : @xmath144 , @xmath145 , @xmath146.,scaledwidth=45.0% ] before concluding , we would like to show that spinning supersymmetric q - balls are stable against small perturbations about their classical configurations . we will closely follow an analysis performed in ref . @xcite ( see also references therein ) regarding the stability of q - balls arising in some theories with nonrenormalizable scalar potentials . writing @xmath147 and varying the functional @xmath18 [ defined by eq . ] with respect to @xmath2 , we find the equation of motion for the field @xmath148 : @xmath149 where @xmath150 is the linear differential operator @xmath151 since eq . is linear in @xmath24 [ due to the form of the potential @xmath152 , see eq . ] , the evolution of small perturbations @xmath153 about the background spinning supersymmetric q - ball configurations is described by a similar equation : @xmath154 the solutions @xmath155 of the above equation are easily found : @xmath156 where @xmath157 , @xmath158 and @xmath159 being the eigenvectors and eigenvalues of @xmath150 : @xmath160 from eq . , we get that the background solution is unstable [ i.e. @xmath161 grows unboundedly with time ] if there exists a @xmath159 such that @xmath162 < 0 $ ] . however , this is not the case since @xmath163 is real . in fact , writing eq . as @xmath164 and remembering that the eigenvalues of the operator @xmath165 are strictly positive real numbers , we obtain @xmath166 where in the last inequality we used eq . . the above equation shows that @xmath163 is a real quantity , as anticipated . we have succeeded in obtaining , analytically , nontopological solitonic solutions with nonvanishing angular momentum in ( 3 + 1)-dimensional minkowski spacetime in the theory of a self - interacting complex scalar field carrying a conserved global @xmath0 charge . this kind of solitons ( known as spinning q - balls ) naturally emerge in a particular class of supersymmetric extensions of the standard model of particle physics where supersymmetry is spontaneously broken at low energy . the scalar field is in this case a gauge - singlet combination of squarks and sleptons corresponding to some flat direction of the supersymmetric potential , while the conserved global charge is some combination of baryon and lepton numbers . in this class of models an effective potential for the flat directions arises due to the breaking of supersymmetry . we have shown that such a type of potential admits , as the nonperturbative ground state of the theory , axisymmetric q - balls whose angular momentum is directed along the axis of symmetry . working in the limit of large charges , we have found that the state of a _ spinning supersymmetric q - ball _ can be labeled by the triple @xmath167 , where @xmath1 is the conserved @xmath0 charge , @xmath89 is positive integer that can take the values @xmath168 and @xmath169 and defines the parity of the state , @xmath170 , while @xmath79 is a integer which gives the projection of the angular momentum on the axis of symmetry through @xmath171 . moreover , we have found the expressions for the energy and radius of spinning supersymmetric q - balls , which fully determine their astrophysical and cosmological properties . it turns out that they do not explicitly depend on @xmath79 and , at fixed charge , are increasing functions of @xmath89 . this indicates that spinning supersymmetric q - balls are indeed excitations of spherically - symmetric supersymmetric q - balls . they are classically stable due to conservation of angular momentum and parity and stable against small perturbations about their classical configurations , and could form during collisions between supersymmetric q - balls and/or during the process of fragmentation of the affleck - dine condensate at the end of inflation . a. g. cohen , _ et al . _ , nucl . phys . b * 272 * , 301 ( 1986 ) ; a. m. safian , s. r. coleman and m. axenides , _ ibid . _ * 297 * ( 1988 ) 498 ; k. m. lee , _ et al . _ , d * 39 * , 1665 ( 1989 ) ; a. kusenko , m. e. shaposhnikov and p. g. tinyakov , pisma zh . fiz . * 67 * , 229 ( 1998 ) ; [ jetp lett . * 67 * , 247 ( 1998 ) ] ; k. enqvist and j. mcdonald , phys . b * 425 * , 309 ( 1998 ) ; m. axenides , _ et al . _ , _ ibid . _ * 447 * ( 1999 ) 67 ; t. multamaki and i. vilja , nucl . b * 574 * , 130 ( 2000 ) ; r. battye and p. sutcliffe , _ ibid . _ * 590 * , 329 ( 2000 ) ; s. kasuya and m. kawasaki , phys . rev . lett . * 85 * , 2677 ( 2000 ) ; s. theodorakis , phys . rev . d * 61 * , 047701 ( 2000 ) ; m. axenides , _ et al . _ , _ ibid . _ * 61 * , 085006 ( 2000 ) ; s. kasuya and m. kawasaki , _ ibid . _ * 61 * , 041301 ( 2000 ) ; _ ibid . _ * 62 * , 023512 ( 2000 ) ; _ ibid . _ * 64 * , 123515 ( 2001 ) ; k. n. anagnostopoulos , _ et al . _ , _ ibid . _ * 64 * ( 2001 ) 125006 ; f. paccetti correia and m. g. schmidt , eur . j. c * 21 * , 181 ( 2001 ) ; a. kusenko and p. j. steinhardt , phys . rev . lett . * 87 * , 141301 ( 2001 ) ; n. graham , phys . b * 513 * , 112 ( 2001 ) ; t. multamaki and i. vilja , _ ibid . _ * 535 * , 170 ( 2002 ) ; m. fujii and k. hamaguchi , _ ibid . _ * 525 * , 143 ( 2002 ) ; k. enqvist , _ et al . _ , _ ibid . _ * 526 * , 9 ( 2002 ) ; m. postma , phys . rev . d * 65 * , 085035 ( 2002 ) ; m. kawasaki , f. takahashi and m. yamaguchi , _ ibid . _ * 66 * , 043516 ( 2002 ) ; a. kusenko , l. loveridge and m. shaposhnikov , _ ibid . _ * 72 * , 025015 ( 2005 ) ; b. kleihaus , j. kunz and m. list , _ ibid . _ * 72 * , 064002 ( 2005 ) ; t. a. ioannidou , a. kouiroukidis and n. d. vlachos , j. math . phys . * 46 * , 042306 ( 2005 ) ; s. kasuya and f. takahashi , jcap * 0711 * , 019 ( 2007 ) ; s. clark , arxiv:0706.1429 [ hep - th ] ; a. kusenko and a. mazumdar , phys . lett . * 101 * , 211301 ( 2008 ) ; i. m. shoemaker and a. kusenko , phys . rev . d * 78 * , 075014 ( 2008 ) ; y. brihaye and b. hartmann , nonlinearity * 21 * , 1937 ( 2008 ) [ arxiv:0711.1969 [ hep - th ] ] ; phys . rev . d * 79 * , 064013 ( 2009 ) ; a. kusenko , a. mazumdar and t. multamaki , arxiv:0902.2197 [ astro-ph.co ] ; y. brihaye , _ et al . _ , arxiv:0903.5419 [ gr - qc ] . for a review on q - balls as dark matter see : a. kusenko , hep - ph/0009089 , _ invited talk at 3rd international conference on dark matter in astro and particle physics ( dark 2000 ) , heidelberg , germany , 10 - 16 jul 2000 . published in * heidelberg 2000 , dark matter in astro- and particle physics * 306 - 315_. f. cappella , r. cerulli and a. incicchitti , eur . j. c * 4 * , 14 ( 2002 ) ; y. takenaga _ et al . _ [ super - kamiokande collaboration ] , phys . b * 647 * , 18 ( 2007 ) ; a. pohl and h. wissing , to appear in the proceedings of workshop on exotic physics with neutrino telescopes , uppsala , sweden , 20 - 22 sep 2006 , astro - ph/0701333 ; s. cecchini _ et al . _ [ slim collaboration ] , eur . phys . j. c * 57 * , 525 ( 2008 ) [ arxiv:0805.1797 [ hep - ex ] ] ; z. sahnoun , arxiv:0812.3248 [ hep - ex ] .	we construct nontopological solitonic solutions in ( 3 + 1)-dimensional minkowski spacetime carrying a conserved global @xmath0 charge and nonvanishing angular momentum in a supersymmetric extension of the standard model with low - energy , gauge - mediated symmetry breaking .
the perturbative particle - vibration - coupling ( pvc ) model for odd - particle - number nuclei emerges naturally from the self - consistent green s function theory @xcite . it describes the polarization of the nucleus when one particle is added or removed @xcite , and its results can , in principle , be directly compared against experimental data . as used in nuclear physics , the perturbative pvc method employs one - particle or one - hole states ( or one - quasiparticle states ) coupled with the rpa or qrpa excitations of an even - even reference nucleus , and the residual nucleon - nucleon interaction that mixes these states at second - order perturbation theory . numerous pvc calculations of increasing level of sophistication have already been performed , see , e.g. , refs . @xcite and excellent recent reviews thereof available in refs . @xcite . an alternative to describing odd nuclei within the perturbative pvc calculations are the energy - density - functional ( edf ) methods , see , e.g. , refs . @xcite , which use blocking of single - particle ( s.p . ) or quasiparticle orbitals . to distinguish these methods from the full pvc approach , in the present study we call them mean - field polarizations or polarization corrections . the advantage of blocked mean - field calculations is that they are non - perturbative and variational . as it turns out , the effects obtained within the blocked mean - field methods are substantially different and , in general , weaker than those obtained from the perturbative pvc @xcite . this discrepancy between models , even when using exactly the same interactions and model spaces , needs to be solved , and this is the main purpose of the present work . the link between the mean - field and perturbative methods has been proposed long time ago @xcite . here , we identify several approximations that are required to firmly establish such a link , and we also extend the derivations to edfs based on density - dependent interactions and to those that include pairing correlations . since the determination of mean - field polarizations requires breaking symmetries , no numerical comparison of the two approaches , such as given here , is available up to now . as required by a thorough comparison , both in the mean - field and ( q)rpa calculations we use full self - consistency and exactly the same particle - hole edfs , pairing interactions , and model spaces . the paper is built around two main chapters presenting theoretical derivations in sec . [ sec2 ] and appendix [ sec5 ] , and numerical results in sec . [ sec3 ] . for theory , we present results pertaining to the hartree - fock ( hf ) approximation ( sec . [ sec2.2 ] ) , density functionals ( sec . [ sec2.4 ] ) , and pairing correlations treated within the hartree - fock - bogoliubov ( hfb ) framework ( sec . [ sec2.3 ] ) . then , results of calculations are discussed for the density - independent ( sec . [ sec3.1 ] ) and density - dependent ( sec . [ sec3.2 ] ) particle - hole interactions , and for paired systems ( sec . [ sec3.3 ] ) . conclusions are given in sec . in this section we revisit the classic problem @xcite of the polarization effect exerted by an odd particle on a mean - field state . to put further discussion in perspective , we study the problem in the hf approximation , and we assume that the mean field is obtained by the hf averaging @xcite of a given known two - body _ density - independent _ interaction that has antisymmetrized matrix elements @xmath0 . let @xmath1 and @xmath2 denote , respectively , the self - consistent density matrix and hf hamiltonian for a system of @xmath3 fermions : @xmath4=a@xmath5 . similarly , let @xmath6 denote the self - consistent density - matrices corresponding to the hf solution for the @xmath7-particle system : @xmath8 , @xmath9 . we use the notation , in which the upper and lower signs correspond to adding or subtracting a particle . without pairing correlations , even and odd systems are described in exactly the same way , so without any loss of generality , we assume that @xmath3 is even . self - consistent hf total energies of the @xmath3- and @xmath7-particle systems are given by @xcite , [ eq:50 ] @xmath10 here , @xmath11 represents the matrix of one - body kinetic energy . in what follows _ we always neglect _ the so - called center - of - mass correction to the kinetic energy @xcite . these corrections are explicitly @xmath3-dependent and thus give trivial so - called mass polarization corrections @xcite to energy differences @xmath12 . although they can always be added , they would obscure the analysis of standard polarization corrections , which are due to two - body interactions , and which are the main focus of the present study . suppose now that @xmath13 ( @xmath14 ) is the density matrix of a s.p . state @xmath15 . we may now ask what the relations are between the density matrices @xmath1 , @xmath6 , and @xmath13 . of course , we can always define a corrective density matrix @xmath16 ( @xmath17 ) such that , by definition , @xmath18 however , a perturbative treatment can only be obtained in the case when @xmath16 is small small in the sense that when the energy of the odd systems @xmath19 , eq . ( [ eq:50b ] ) , is calculated for the density matrix in eq . ( [ eq:1 ] ) , only terms up to second order in @xmath16 are important . note that , by definition , the three density matrices are hermitian and projective , @xmath20 also note that @xmath16 does depend on the polarizing orbital @xmath15 ; nevertheless , we do not mark it with superscript @xmath15 . this is to avoid a confusion of understanding @xmath16 as a correction to the orbital itself ; indeed , this correction certainly corresponds to a modification of _ all _ orbitals of the system . the corrective density matrix @xmath16 can be small only when the orbital @xmath15 , and the states in even and odd nuclei , are chosen in a specific way . we may then have four interesting cases to consider . in the first case , let us assume that we initially solve the self - consistent equations of the even @xmath3-particle system , and @xmath21 is one of the unoccupied hf eigenstates therein ( a particle state ) , that is , @xmath22 , @xmath23 . we may now put a particle in this orbital and solve the self - consistent equations of the @xmath24-particle system . in this sense , the @xmath24-particle system becomes polarized by an addition of a particle to the @xmath3-particle system . note that by this procedure _ all _ hf single - states of the @xmath3-particle system become modified , including the added orbital @xmath21 . the second case is obtained by a similar procedure , where instead we arrive at a polarized @xmath25-particle system . for that , we pick @xmath21 as one of the occupied hf eigenstates of the @xmath3-particle system ( a hole state ) , that is , @xmath22 , @xmath26 . by removing a particle from this state and solving the self - consistent equations of the @xmath25-particle system , we now obtain the polarization correction corresponding to a hole . note that for the two choices discussed up to now , the density matrices @xmath27 and @xmath28 , are projective , that is , @xmath29 the two remaining interesting cases correspond to inverse polarizations , namely , we may initially solve the self - consistent equations of the @xmath7-particle systems , and then pick @xmath15 either as an unoccupied orbital in the @xmath25-particle system , @xmath30 , @xmath31 or as an occupied orbital in the @xmath24-particle system , @xmath32 , @xmath33 . of course , in both cases , the self - consistent equations solved for the @xmath3-particle system give the same solutions as before , however , now orbitals @xmath21 correspond to the @xmath7-particle systems , and thus density matrices @xmath34 and @xmath35 , are projective , that is , @xmath36 we see that equations we are going to derive for the corrective density @xmath16 do depend on the choices made for the orbital @xmath21 . for the direct polarizations , that is , when @xmath21 is a self - consistent orbital in the even system , we square both sides of eq . ( [ eq:1 ] ) , and from eq . ( [ eq:50c ] ) we obtain @xmath37 for the inverse polarizations , that is , when @xmath21 are self - consistent orbitals in the odd systems , we rewrite eq . ( [ eq:1 ] ) in the form , @xmath38 and then square both sides , which from eq . ( [ eq:51 ] ) gives , @xmath39 equations ( [ eq : j4 ] ) and ( [ eq : j4a ] ) allow us to derive specific properties of @xmath16 that , however , are different for direct and inverse polarizations . assuming that we can split @xmath16 into terms of first , second , and higher ( neglected ) orders , that is , @xmath40 we now separately discuss direct and inverse polarizations . in what follows , we refer to the expansion in eq . ( [ eq:52 ] ) as rpa expansion , and we strive to discuss what an acceptable magnitude of @xmath13 is , for which such an expansion is meaningful . beginning with the inverse polarizations , eq . ( [ eq : j4a ] ) gives @xmath41 we can now discuss properties of the particle - hole ( ph ) , particle - particle ( pp ) , and hole - hole ( hh ) matrix elements of @xmath16 , where the standard hole and particle states correspond to the occupied and unoccupied states , respectively , _ in the even @xmath3-particle system _ , that is , @xmath42 equation ( [ eq:53a ] ) does not put any constraint on the ph matrix elements of @xmath43 , and it requires that its pp and hh matrix elements vanish identically . therefore , the leading - order ( second - order ) pp and hh matrix elements of @xmath16 are determined by eq . ( [ eq:53b ] ) , and they solely depend on the leading - order ( first - order ) ph matrix elements thereof , that is , [ eq:57 ] @xmath44 we see that the standard ph structure of @xmath16 , pertaining to the @xmath3-particle system , appears for the _ inverse polarizations_. however , as we derived above , in this case the polarizing orbital @xmath21 must be calculated in the _ odd _ system . let us next discuss direct polarizations , for which eq . ( [ eq : j4 ] ) holds . the rpa expansion ( [ eq:52 ] ) then gives @xmath45 properties of the particle - hole ( ph ) , particle - particle ( pp ) , and hole - hole ( hh ) matrix elements of @xmath16 , now pertain to nonstandard hole and particle states , which are the occupied and unoccupied states , respectively , _ in the odd @xmath7-particle system _ , that is , @xmath46 we then have unconstrained matrix elements @xmath47 and [ eq:56 ] @xmath48 in summary , for _ direct polarizations _ , for which the polarizing orbital @xmath21 is calculated in the _ even _ system , we obtain the nonstandard ph structure ( [ eq:56 ] ) of @xmath16 . however , for _ inverse polarizations _ , for which the polarizing orbital @xmath21 is calculated in the _ odd _ system , we obtain the standard ph structure ( [ eq:57 ] ) of @xmath16 . from these considerations , it appears that a rigorous analysis of the hf polarization effects , based on the elements solely determined in the even system , does not exist , and one must make further simplifying assumptions . the easiest way out is to neglect the differences between the polarizing orbitals calculated in the even and odd systems and use equations pertaining to inverse polarizations along with @xmath21 determined in the even system . in what follows , we use this strategy . equations for the polarization corrections to the s.p.energies can be derived by comparing the self - consistent energies in even and odd systems . inserting the odd - system density matrices ( [ eq:1 ] ) into the odd - system energy ( [ eq:50b ] ) , we obtain : @xmath49 we now use the following facts and definitions : @xmath50 equation ( [ eq:55a ] ) is the standard definition of the hf mean field in the @xmath3-particle system and @xmath51 ( [ eq:55b ] ) is its diagonal matrix element in the self - consistent basis . equation ( [ eq:55c ] ) is a simple consequence of the antisymmetry of the two - body matrix elements and represents the fact that in the hf approximation there is no self interaction ( si ) . equations ( [ eq:55d ] ) and ( [ eq:55e ] ) define the mean - field potentials generated by the polarizing orbital @xmath21 and correction @xmath16 , respectively . in terms of these definitions , the odd - system energy can be written as , @xmath52 up to now , expression ( [ eq:5b ] ) is exact . to simplify it , we can use the rpa expansion ( [ eq:52 ] ) and thus conditions ( [ eq:57 ] ) , and neglect terms beyond second order . in the basis of particle and hole states , the mean - field hamiltonian @xmath53 is by definition diagonal ; therefore , owing to eqs . ( [ eq:57 ] ) , the third term on the right - hand side is of the second order in @xmath43 . similarly , the fifth term is obviously of the second order too . however , unless we assume that @xmath54 is small ( of the first rpa order ) , the fourth term may contain subleading second - order terms , which we can not treat . therefore , to have a consistent second - order expression for the energy of the @xmath55 system , we must make the assumption of @xmath54 being small as compared to @xmath2 . this assumption can also be understood as @xmath13 being small as compared to @xmath1 , that is , the system being appropriately heavy . in fact , such an assumption can partially be tested by keeping the leading - order ( second - order ) pp@xmath56 and hh@xmath56 matrix elements of the fourth term , which depend on the leading - order ( first - order ) matrix elements of @xmath16 . then , we obtain the following approximate expression , @xmath57 this can be summarized in the form of polarization corrections to energies of odd states @xmath58 , @xmath59 or polarization corrections to s.p . energies @xmath60 , @xmath61 for @xmath62 where @xmath16 and @xmath63 represent vectors of ph matrix elements , @xmath64 and @xmath65 , respectively , that is , [ eq : e1 ] @xmath66 and matrices @xmath67 and @xmath68 , [ eq : e ] @xmath69 build the rpa matrix @xmath70 . we see that the second - order terms depending on @xmath54 , which we have kept in eq . ( [ eq:5cc ] ) , lead to modified matrix elements @xmath71 , as compared to the standard rpa matrix @xmath72 . in this formulation , the rpa equations do depend on the polarizing orbital @xmath15 . in sec . [ sec3 ] , we perform numerical calculations with and without these terms , and we check that they play a minor role and can be safely omitted , thus supporting the validity of the assumption about the smallness of @xmath54 . equation for the correction @xmath16 can be derived from the fact that the density matrix of eq . ( [ eq:1 ] ) is a self - consistent solution of the hf equations in the @xmath7-particle system , @xmath73 as previously , we neglect differences between the orbitals @xmath21 calculated in even and odd systems , that is , we have @xmath74 . moreover , since @xmath75 is the self - consistent solution of the @xmath3-particle system , we have @xmath76 , which gives @xmath77 in the leading ( first ) order , the last term , quadratic in the density @xmath16 , can be dropped , and we also drop the second - order matrix elements @xmath78 and @xmath79 . then , the ph and hp matrix elements of the above equation read , @xmath80 \label{eq:10a } 0 & = & \bigl ( e_h - e_p \bigr ) \delta\rho_{hp } \mp h^\lambda_{hp } \pm \sum_{h ' } h^\lambda_{hh ' } \delta\rho_{h'p } \nonumber \\ & & \mp \sum_{p ' } \delta\rho_{hp ' } h^\lambda_{p'p } -\delta h_{hp } \nonumber\\ & = & -(a'^ * \delta\rho^)_{ph } - ( b^ \delta\rho)_{ph } \mp h^{\lambda*}_{ph},\end{aligned}\ ] ] and in the matrix notation they can be written as , @xmath81 here again we see that the matrix elements of @xmath63 must be at least of the same rpa order ( the first - order ) as are those of @xmath16 . condition ( [ eq:10b ] ) is exactly equal to the condition that the total energy of the odd system ( [ eq:5da ] ) is stationary with respect to correction @xmath16 . in other words , vanishing variation of @xmath58 , eq . ( [ eq:5db ] ) , with respect to @xmath16 gives eq . ( [ eq:10b ] ) . then , at the stationary point , the correction to the total energy reads @xmath82 that is , for a positive - definite rpa matrix , the correction to the total energy is always negative , irrespective of adding or subtracting a particle . for the corrections to s.p . energies we have @xmath83 that is , particle states move down and hole states move up . in view of eq . ( [ eq:10b ] ) , corrections ( [ eq:5 g ] ) can also be written in two other equivalent forms : @xmath84 in eq . ( [ eq:111a ] ) , the inverse of the rpa matrix can be calculated either through its eigenvectors or through the rpa amplitudes . in the second case , we use the rpa equations and completeness relations @xcite , @xmath85 @xmath86 where @xmath87 is a diagonal matrix of positive rpa eigenvalues . this gives @xmath88 the inverse of the rpa matrix exists if all eigenvalues are non - zero , and has the form , @xmath89 finally , in terms of the rpa amplitudes and energies , corrections ( [ eq:111a ] ) then become equal to @xcite , @xmath90 let us now rederive the polarization corrections of sec . [ sec2.2 ] for the case of the total energy given by a minimization of an edf , and not of the hf average of a hamiltonian . differences between these two cases can be of dual origin . first , a quasilocal edf built as the most general quadratic function of densities deviates from a hf average of a zero - range momentum - dependent interaction unless its coupling constants obey a specific set of linear conditions , see , e.g. , refs . @xcite . for the skyrme edf , these conditions can be formulated as a linear dependence of the time - odd coupling constants on the time - even ones , and a linear dependence between the isovector and isoscalar spin - orbit coupling constants @xcite . in this work we only consider edfs of this type . the second reason for differences arises because of so - called density - dependent interactions , which also lead to edfs that are not equal to hf averages of hamiltonians . focusing on this second case , we now consider edfs determined by the hf averaging of antisymmetrized density - dependent matrix elements @xmath91 $ ] . then , the total energies read , [ eq:60 ] @xmath92\rho^{a } ) , \\ \label{eq:60b } e^{a\pm1 } & = & { \rm tr}(t\rho^{a\pm1 } ) \nonumber \\ & & + { \tfrac{1}{2}}{{{\rm tr}}}_1{{{\rm tr}}}_2(\rho^{a\pm1}\bar{v}[\rho^{a\pm1}]\rho^{a\pm1 } ) .\end{aligned}\ ] ] we see that both energies , apart from the standard quadratic dependencies on densities , cf . ( [ eq:50 ] ) , do depend on densities through the density dependence of interactions . these latter dependence precludes comparing energies of even and odd systems , unless we make the simplifying assumption that @xmath93 $ ] and @xmath94 $ ] can be connected by a second - order expansion in @xmath95 . from eq . ( [ eq:1 ] ) we see again that this requires @xmath13 to be of the same ( first ) rpa order as @xmath96 . under this assumption , we have @xmath97 \simeq \bar{v}_{i'k'ik}[\rho^{a } ] \pm \sum_{mn}\frac{\partial \bar{v}_{i'k'ik}}{\partial\rho_{mn } } \left(\rho^\lambda_{mn}\pm\delta{\rho}_{mn}\right ) \nonumber \\ & & \!\!+{\tfrac{1}{2}}\!\!\!\ ! \sum_{m'n'mn}\!\!\frac{\partial^2 \bar{v}_{i'k'ik}}{\partial\rho_{mn}\partial\rho_{m'n ' } } \left(\rho^\lambda_{mn } \pm\delta{\rho}_{mn}\right ) \left(\rho^\lambda_{m'n'}\pm\delta{\rho}_{m'n'}\right ) , \nonumber \\\end{aligned}\ ] ] where all partial derivatives must be evaluated at @xmath98 . we can now insert eqs . ( [ eq:1 ] ) and ( [ eq:61 ] ) into the odd - system energy ( [ eq:60b ] ) and obtain up to the second order in @xmath99 , @xmath100 where the mean - field hamiltonian @xmath101 , @xmath102 and effective two - body matrix elements , @xmath103 and @xmath104 , contain rearrangement terms , [ eq:64 ] @xmath105 the redefined two - body matrix elements allow us to write the odd - system energy in the form analogous to eq . ( [ eq:5b ] ) , @xmath106 but with the following redefinitions , @xmath107 we see that the first order rearrangement terms ( [ eq:64a ] ) become fully absorbed in the s.p . energies , which are now , as usual , the eigenvalues of mean fields @xmath108 . moreover , both the polarizing fields @xmath109 and rpa matrices @xmath3 and @xmath68 , see eqs . ( [ eq : e1 ] ) and ( [ eq : e ] ) , must now be determined by using the second - order rearrangement terms ( [ eq:64b ] ) . therefore , owing to the fact that the effective two - body matrix elements ( [ eq:64a ] ) are not antisymmetric , the si term ( [ eq:66c ] ) , @xmath110 is non - zero , and explicitly appears in eq . ( [ eq:65 ] ) . this leads to corrections to s.p . energies now having the form , @xmath111 where , based on the analogy with eq . ( [ eq:117 ] ) , the first term can be called self - interaction - free ( sif ) polarization correction , @xmath112 the second - order mean fields @xmath113 ( [ eq:66d ] ) and @xmath114 ( [ eq:66e ] ) are simply related to the linearized first - order mean fields , that is , @xmath115 these expressions can be explicitly verified directly from definitions ( [ eq:64 ] ) . they are extremely useful in practical applications , because : ( i ) the second - order mean - fields ( [ eq:67d ] ) that define the polarization vertex ( [ eq:68a ] ) can be determined without explicitly calculating the second derivatives of matrix elements , ( ii ) the amplitude mean - fields ( [ eq:67e ] ) are the only objects that one has to calculate when using the iterative methods to solve the rpa equations @xcite , and ( iii ) exactly the same piece of code can be used to calculate both mean fields ( [ eq:67d ] ) and ( [ eq:67e ] ) . the si term ( [ eq:88 ] ) , where a particle interacts with the mean field generated by itself , is unphysical , because in reality each nucleon should interact with the other nucleons only . as discussed in sec . [ sec2.2.2 ] , for an edf generated by hamiltonian , no si appears . on the other hand , edfs generated by density - dependent interactions do produce the si . an edf has a one - body si if it gives non - zero energy for a single nucleon state . this was discussed in ref . @xcite , where it has been shown how the one - body si of a skyrme edf can be removed by introducing extra constraints between the skyrme coupling constants . in general , the si results from the violation of the antisymmetry of effective matrix elements ( [ eq:64b ] ) . for density functionals used in electronic structure calculations , an si correction was originally introduced by perdew and zunger @xcite , and numerous variations and improvements of the method have been later constructed . a short review of the various si - correction methods used in electronic structure calculations can be found in ref . @xcite . in nuclear physics context , the si problem in connection with density - dependent skyrme interactions was early on discussed in ref . @xcite and more recently in ref . @xcite . in refs . @xcite , the si problem and ways to correct it was discussed in detail , together with the related concept of self pairing , see also sec . [ sec2.3 ] . within the skyrme edf approach without pairing , the si results from the zero - range density - dependent interaction and from the coulomb exchange , which is treated in the slater approximation . since in the polarization correction ( [ eq:68 ] ) the si term appears explicitly , one can simply remove it from this expression and thus obtain sif result ( [ eq:68a ] ) . however , we stress here that the self - consistent calculations performed in odd nuclei do contain the si term , and lead to the polarization correction ( [ eq:68 ] ) with the si term included . for a paired system corresponding to the average number of particles @xmath3 , one diagonalizes the quasiparticle hamiltonian @xmath116 , which in the standard double - dimension representation @xcite reads , @xmath117 where @xmath118 and where @xmath119 and @xmath120 are the ph and pp mean potentials , respectively . in what follows , for clarity we write only one fermi energy @xmath15 generalizations to separate neutron and proton fermi energies being obvious . the eigenequation for @xmath116 defines one - quasiparticle states @xmath121 and one - quasiparticle energies @xmath122 , @xmath123 where positive ( negative ) indices @xmath124 ( @xmath125 ) correspond to positive ( negative ) quasiparticle energies @xmath126 @xmath127 of quasiparticle ( quasihole ) states . then , the basic dynamical quantity describing the system is the generalized density matrix , @xmath128 , @xmath129 which projects states on the space of occupied quasihole states , @xmath130 and depends on the density matrix @xmath1 and pairing tensor @xmath131 . when the quasiparticle and quasihole states are arranged as columns of matrix @xmath132 in doubled dimensions , they form the matrix of the bogoliubov transformation @xcite , @xmath133 in terms of which we have @xmath134 and @xmath135 . the generalized density matrix of an odd system @xmath136 is obtained by the blocking procedure @xcite , whereupon one occupied quasihole state for @xmath137 is replaced by its empty quasiparticle partner for @xmath138 . then , eqs . ( [ h - dens])([eq:70 ] ) are solved self - consistently again , and the qrpa polarization correction @xmath139 is defined in analogy with eq . ( [ eq:1 ] ) as , @xmath140 where the generalized density matrix @xmath141 describes the blocked orbital [ see eq . ( [ eq:70 ] ) ] , @xmath142 this gives the density matrix and pairing tensor of the odd system in the form , [ eq : kappdel2 ] @xmath143 with explicit contributions coming from the blocked orbital given by , [ eq : kappdel3 ] @xmath144 density matrices and pairing tensors , self - consistently calculated in even and odd systems , determine the corresponding total energies with pairing as , [ eq:71 ] @xmath145 where @xmath146 and @xmath147 . let us now discuss the case of the ph and pp potentials being determined by averaging the corresponding ph and pp two - body antisymmetric interaction matrix elements , that is , [ eq:72 ] @xmath148 with the analogous equations defining the potentials in the odd system , @xmath149 and @xmath150 . in most nuclear - physics applications , the ph and pp matrix elements are different , which means that the total energies do not , strictly speaking , correspond to an average value of a hamiltonian . to discuss the structure of the resulting expressions , we first analyze the situation of these matrix elements being density - independent , that is , with the rearrangement terms ignored . then , the potentials depend linearly on densities , and each term in the generalized density matrix @xmath136 ( [ eq:73 ] ) gives the corresponding term in the generalized potential @xmath151 , @xmath152 by inserting eqs . ( [ eq:73 ] ) and ( [ eq:74a ] ) into expression for the total energy of the odd system ( [ eq:71b ] ) , we easily obtain the analogue of eq . ( [ eq:5b ] ) with pairing , that is , @xmath153,\end{aligned}\ ] ] where , to keep notation consistent with the unpaired case of sec . [ sec2.2.2 ] , we have denoted @xmath154 and @xmath155 . at this point , we have to recall that the average number of particles in the blocked state , @xmath156 for @xmath157 is not necessarily equal to @xmath55 . therefore , to calculate the average energy of the @xmath55 system , one has to introduce a linear correction @xcite , such that @xmath158 where by definition @xmath159 is the fermi energy of the odd system . we then see that , under the assumption of self - consistent fermi energies in the even and odd systems being equal , @xmath160 , we have @xmath161.\end{aligned}\ ] ] from this point on , derivations proceed exactly as in the case of no pairing , sec . [ sec2.2 ] , so we only repeat principal definitions and results . we assume that the blocked quasiparticle wave function @xmath162 , which is determined in the odd system , is identical to that determined in the even system . only under such an assumption we have @xmath163 [ see eqs.([eq:79 ] ) and ( [ eq:82 ] ) ] and the analogue of eq . ( [ eq:53a ] ) holds , @xmath164 in eq . ( [ eq:77 ] ) we identify the si term , analogous to that derived without pairing ( [ eq:88 ] ) , namely , @xmath165 where @xmath166 and @xmath167 are the mean fields generated by the blocked quasiparticle , [ eq:92 ] @xmath168 thus the si term corresponds to the blocked quasiparticle @xmath169 that interacts with the generalized mean - field potential it has generated . by combining eqs . ( [ eq : kappdel3 ] ) and ( [ eq:92 ] ) , we can easily derive that @xmath170 where the antisymmetry of matrix elements @xmath171 and @xmath172 was used . we explicitly see that a non - zero value of @xmath173 can only appear when the pp and ph interactions , which define the edf with pairing , are not identical to one another . another assumption we have to make is that the mean field @xmath174 , generated by the blocked quasiparticle @xmath169 , is appropriately small of the first order in qrpa . then , the polarization corrections to paired energies of odd states @xmath58 and those to quasiparticle energies @xmath175 [ cf . ( [ eq:5da ] ) and ( [ eq:5dc ] ) ] , which are defined by @xmath176 can be expressed in the form @xmath177 where @xmath178 and @xmath179 represent vectors of quasiparticle - quasihole matrix elements of @xmath139 and @xmath174 , respectively , that is @xmath180 for @xmath181 , and @xmath3 and @xmath68 are the standard components of the qrpa matrix @xcite . equation for @xmath178 can easily be derived by following the steps presented in sec . [ sec2.2.3 ] it simply results from the requirement that the correction to the energy ( [ eq:81 ] ) is stationary , which gives , @xmath182 and @xmath183 for @xmath184 the qrpa sif polarization correction to quasiparticle energy explicitly reads @xmath185 where @xmath186 and @xmath187 are the standard qrpa amplitudes . before discussing numerical results in the next section , let us briefly touch upon the problem of conserved symmetries . no specific conserved symmetry was , in fact , assumed in the derivations presented so far . in appendix [ sec5 ] , we discuss in detail implications of conserving the spherical symmetry . for the conserved parity , the s.p . wave function @xmath188 of a polarizing orbital @xmath15 has a definite parity @xmath189 , that is , @xmath190 . therefore , the density matrix @xmath13 , which in the space coordinates reads @xmath191 , is _ always _ parity even , irrespective of the parity of the polarizing orbital : @xmath192 . with pairing correlations included , analogous arguments hold for quasiparticle wavefunction @xmath193 and generalized density matrix @xmath141 . since interactions are parity - invariant , the positive parity of @xmath13 or @xmath141 implies positive parity of mean fields @xmath54 or @xmath194 , and thus only positive - parity phonons contribute to the energy shifts in eqs . ( [ eq:68a ] ) or ( [ eq:86b ] ) . this means that , for conserved spherical and parity symmetries , all polarization corrections discussed in sec . [ sec2 ] relate to the rpa and qrpa channels and phonons @xmath195 with positive parity @xmath196 only . therefore , the discussed polarization corrections can not , and do not , involve any couplings to negative parity channels , including those to the very important @xmath197 channel . the latter can only be treated within the pvc methods @xcite , which will be discussed in the forthcoming study @xcite . the lack of couplings to negative - parity channels constitutes the main drawback of the mean - field methods in describing states in odd nuclei . we note here that the polarization corrections studied in this work are equivalent to the `` diagonal '' approximation to the pvc , whereupon the polarization vertex is limited to the same state as the one for which the pvc is evaluated . this shows again that the parity - conservation in the vertex excludes coupling to negative - parity phonons . such coupling is only possible when the `` non - diagonal '' pvc is calculated for intermediate states that involve opposite - parity s.p.or quasiparticle states . there is another hypothetical possibility of including the coupling to negative - parity phonons , namely , through a dynamical parity - breaking of the mean field . this would require performing generator - coordinate - like calculations based on mixing octupole - deformed states for odd nuclei . it is , however , unclear if such an approach can be equivalent to the pvc that includes negative - parity phonons . another drawback of the mean - field approach , clearly identified in secs . [ sec2.4 ] and [ sec2.3 ] , is the presence of the si terms in the mean - field binding energies of odd nuclei . based on the analyses performed within the rpa and qrpa methods , we could explicitly identify these terms , which allows for calculating them _ after variation_. the explicit identification may also allow us to remove them _ before variation _ , which will be the subject of subsequent studies . of course , although not explicitly `` visible '' , the si terms are also present in the mean - field binding energies of even nuclei , and in the future new functionals with these terms removed should also be studied . finally , we note that expression ( [ eq:90 ] ) for the si energy with pairing is valid only for density - independent interactions . however , expression ( [ eq:89 ] ) does take effects of density - dependent interactions into account , provided it is evaluated for mean - field potentials with rearrangement terms included , as derived in sec . [ sec2.2 ] , that is , @xmath198 all calculations presented in this section aim at comparing self - consistent results obtained by using the deformed solver ( v2.52k ) @xcite with rpa and qrpa solutions implemented in the spherical solver @xcite . we used the configuration space that includes all harmonic - oscillator shells up to @xmath199 we begin the presentation by showing examples of calculations performed for the case of an exact hf approximation , as discussed in sec . [ sec2.2 ] . to this end , we employed the density - independent skyrme interaction sv @xcite and we analyzed results only for neutrons , so as to avoid effects of density - dependent slater approximation for the coulomb exchange term . on the one hand , to treat the edf sv as fully generated by an interaction , we included in the functional all tensor terms , which were originally neglected @xcite . also the `` native '' time - odd terms of sv were included . on the other hand , as mentioned in sec . [ sec2.2 ] , we neglected the so - called center - of - mass correction to the kinetic energy . sn , determined by using the hf and rpa methods and skyrme edf sv @xcite , see text . lines connect the values obtained for different projections of the angular momentum @xmath201 ( from left to right ) . , scaledwidth=40.0% ] in fig . [ fig : three ] we test eq . ( [ eq:5dc ] ) , that is , we compare polarization corrections , @xmath202 obtained from the hf energies of odd and even systems , @xmath19 and @xmath203 , and hf s.p.energies , @xmath51 , with those determined form the rpa solutions , eq . ( [ eq:117 ] ) . apart from a few cases , the obtained agreement is nearly perfect . this result is particularly gratifying , because it confirms not only the analytical derivations presented in sec . [ sec2.2 ] and appendix , but also the validity of two completely independent numerical codes . at this point , we must discuss one important aspect of the hf calculations in odd nuclei . in principle , for any given value of @xmath204 , there may exist two solutions : one with prolate and another one with oblate shape . usually only the lowest one can be converged ; the other one , being excited , either does not converge or falls down to the lowest one . in our calculations , in full agreement with the standard nilsson diagram @xcite , we obtain converged prolate ( oblate ) solutions for low-@xmath204 ( high-@xmath204 ) particle states , and _ vise versa _ for the hole states . we note here that we did not constrain these solutions to axial symmetry ; nevertheless , stable triaxial solutions were never obtained . orbitals in @xmath200sn , determined by not including ( left bars ) and including ( right bars ) the orbital - dependent terms in the rpa matrices , see text . the order of orbitals is the same as shown in fig . [ fig : three ] . contributions coming from four rpa channels @xmath205 , 1@xmath206 , 2@xmath206 , and 3@xmath206 are shown separately ( note very different scales ) . , scaledwidth=40.0% ] next , we tested the assumption , discussed in sec . [ sec2.2 ] , related to the smallness of terms @xmath13 and @xmath54 with respect to the rpa expansion . in fig . [ fig : one ] , we compare polarization corrections determined by using the standard rpa matrices with those containing the orbital - dependent terms [ second line in eq . ( [ eq : ea ] ) ] . since both sets of results are almost identical , we conclude that in medium - heavy nuclei like @xmath200sn , the orbital - dependent terms can be safely ignored . this significantly simplifies the calculations , because a single common solution of the rpa equation can then be used to determine polarization corrections for all orbitals . orbitals in @xmath200sn , coming from different @xmath195 rpa channels , determined for the skyrme edf sv @xcite . the order of orbitals is the same as shown in fig . [ fig : three ] . contributions are ordered according to the value of @xmath207 , with the @xmath208 channels shown nearest the abscissa . , scaledwidth=40.0% ] but for the @xmath209 orbitals . , scaledwidth=40.0% ] in fig . [ fig : two ] , we show polarization corrections of the @xmath210 orbitals in @xmath200sn , split into contributions from different @xmath195 rpa channels . first we note that the geometric constraints in eq . ( [ eq:86e ] ) limit the polarizations of @xmath211 orbitals to channels with @xmath212 . as expected , the largest contributions come from the coupling to the quadrupole channel @xmath213 , however , the monopole @xmath208 and dipole @xmath214 channels also significantly contribute . for higher-@xmath211 orbitals , channels @xmath215 and @xmath216 show some effect , whereas , channels with @xmath217 can be safely neglected . for the @xmath209 orbitals shown in fig . [ fig : two_two ] , the convergence is slightly slower , but still all terms with @xmath218 contribute very little . we now proceed to discuss the problem of si energies in the edf calculations , presented in sec . [ sec2.4 ] . to this end , we repeated the self - consistent calculations presented in sec . [ sec3.1 ] , by employing the skyrme edf sly5 @xcite . this is a standard skyrme parametrization containing a strong density dependent term , for which we can study the si energies , as defined in eq . ( [ eq:88 ] ) . as before , the `` native '' time - odd terms of sly5 were included and the center - of - mass correction to the kinetic energy was neglected . , but for the skyrme edf sly5 @xcite . the rpa results correspond to the sif terms in eq . ( [ eq:68 ] ) , whereas rpa+si denotes both sif and si contributions combined . , scaledwidth=40.0% ] in fig . [ fig : four ] , we show the rpa ( sif ) contributions to polarization corrections ( [ eq:68 ] ) , and we compare the total polarization corrections calculated by using eq . ( [ eq:68 ] ) with the hf results ( [ eq:48 ] ) . the obtained agreement is very good , although not as perfect as that obtained in sec . [ sec3.1 ] for the skyrme edf sv . moreover , the rpa results obtained for the sv and sly5 functionals are significantly different from one another ; the latter ones being close to about @xmath2190.4mev for holes and particles , respectively . we also see that the sly5 results are much less @xmath204-dependent . ) , calculated in @xmath200sn for the skyrme edf sly5 . , scaledwidth=40.0% ] the most striking observation seen in fig . [ fig : four ] , also explicitly illustrated in fig . [ fig : five ] , is a strong cancellation between the sif and si contributions to the polarization corrections ( [ eq:68 ] ) . this cancellation makes the hf polarization corrections quite small , and gives the explanation to the long - standing problem of significant differences between the magnitudes of the hf and rpa values @xcite . indeed , it is the unphysical si contribution that renders the hf polarization corrections so small , see ref . @xcite for a set of comprehensive calculations across the mass chart . but for the contributions to the rpa sif polarization corrections of @xmath210 orbitals , determined for the skyrme edf sly5 . , scaledwidth=40.0% ] but for the @xmath209 orbitals . , scaledwidth=40.0% ] we conclude this section by showing , in figs . [ fig:7a ] and [ fig:7b ] , the rpa sif polarization corrections of the @xmath210 and @xmath209 orbitals , respectively , split into contributions coming from different @xmath195 rpa channels and calculated in @xmath200sn for the skyrme edf sly5 . these can be compared with the analogous ones shown in figs . [ fig : two ] and [ fig : two_two ] for the skyrme edf sv . we first see that for the @xmath210 orbitals , the convergence patterns obtained for both edfs are fairly similar . however , for the @xmath209 orbitals , contributions coming from the @xmath220 phonons turn out to be always quite large . for example , results obtained for the h@xmath221 orbital certainly require taking into account the @xmath222 phonons . to present numerical results pertaining to the description of polarization effects with pairing correlations included , sec . [ sec2.3 ] , we performed hfb and qrpa calculations for the tin isotopes @xmath223sn . as in sec . [ sec3.2 ] , we used the skyrme edf sly5 , whereas the pairing interaction was modelled by a contact volume pairing force with the strength of @xmath224mev@xmath225 and active pairing space restricted to states below 60mev . in fig . [ fig : six ] we aim at testing eq . ( [ eq:80 ] ) , where @xmath19 and @xmath203 are self - consistent hfb ground - state energies of odd and even nuclei , respectively , @xmath15 and @xmath226 are the hfb fermi energy and quasiparticle energy of the blocked state , and @xmath58 is the qrpa ( sif+si ) polarization correction ( [ eq:86 ] ) . for this comparison , we must decide whether to use the hfb results obtained for the lighter @xmath25 or heavier @xmath24 odd system . obviously , the former ( latter ) option must be used for predominantly hole - type ( particle - type ) quasiparticles . for quasiparticles near the fermi level , however , there is a certain degree of ambiguity , which we here arbitrarily resolve by checking whether the single - particle energy @xmath227 corresponding to the blocked quasiparticle state is below or above the fermi level @xmath15 . in practice , we determine @xmath227 by diagonalizing in the even nucleus the mean - field hamiltonian @xmath2 , which is a part of the hfb hamiltonian ( [ h - dens ] ) . in addition , to link results presented in this section to those presented before for magic nuclei without pairing , in figs . [ fig : six][fig:16 ] we plot results for hole states with flipped signs , that is , [ eq:80a ] @xmath228 [ cf . ( [ eq:48 ] ) ] . within such a convention , in fig . [ fig : six ] we show the qrpa sif+si ( symbols ) and hfb ( lines ) polarization corrections given by the left - hand and right - hand sides of eqs . ( [ eq:80a ] ) , respectively . we note that not all blocked quasiparticle states could be converged in all studied nuclei , and thus in the figure there is quite a number of missing hfb points . nevertheless , we conclude that the agreement between the qrpa and hfb results is satisfactory . by this we establish the equivalence of the two methods in determining the polarization corrections with pairing . but for @xmath229sn . , scaledwidth=40.0% ] but for @xmath230sn . , scaledwidth=40.0% ] in figs . [ fig:9a ] and [ fig:9b ] , we compare the qrpa sif ( [ eq:86b ] ) and si ( [ eq:89a ] ) contributions to the polarization corrections . similarly as in the case without pairing , shown in fig . [ fig : five ] , the sif and si terms always have opposite signs , and thus the si partially cancels the sif contribution . however , here the si terms are relatively smaller , and thus they to a lesser degree decrease the sif contributions , as compared to the results with no pairing . it is fairly difficult to pin down specific reasons for the qualitative differences between the si energies obtained with and without pairing correlations . it could be that the si energies related to density - dependence of the skyrme interaction ( [ eq:88 ] ) and those related to differences between the pp and ph channels ( [ eq:90 ] ) , partially cancel out . but for @xmath229sn . , scaledwidth=40.0% ] but for @xmath229sn . , scaledwidth=40.0% ] convergence of the qrpa polarization corrections as a function of the angular momentum @xmath207 of the qrpa phonons , shown in figs . [ fig:10a ] and [ fig:10b ] , is much faster than that without pairing , cf . figs . [ fig:7a ] and [ fig:7b ] . here , the 2@xmath206 channels clearly dominate . this can be interpreted as the result of the quadrupole collectivity being increased by the pairing correlations . in most cases , channels with @xmath217 can be safely neglected , with the exception of the @xmath231 channels that slightly contribute to the corrections of the @xmath232 quasiparticle states . , ( [ eq:102 ] ) , in tin isotopes . , scaledwidth=40.0% ] , but for the qrpa sif polarization corrections only . , scaledwidth=40.0% ] all results presented up to now pertain to single - reference hf(b ) and ( q)rpa calculations , that is , only one single orbital , with a fixed projection @xmath204 or @xmath233 , was occupied and was inducing polarization effects . as discussed previously , this required symmetry breaking in the hf(b ) solutions and required coupling of ( q)rpa phonons to odd particles in a symmetry - nonconserving way . however , a symmetry - conserving ( q)rpa coupling @xcite simply amounts to averaging the results obtained for different values of @xmath204 or @xmath233 , see eqs . ( [ eq:102 ] ) . in figs . [ fig:14][fig:16 ] , we present results for the averages obtained in this way . figures [ fig:14 ] and [ fig:15 ] summarize our results obtained for the qrpa polarization effects in tin isotopes . we see that the polarization corrections strongly depend on @xmath3 . this is mostly due to the fact that for the skyrme edf sly5 , the quadrupole collectivity in tin isotopes varies with @xmath3 , and peaks near @xmath234 where the nuclei are softest against the quadrupole deformation and the qrpa 2@xmath206 phonons are lowest in energy and have the largest strength . at @xmath234 , values of polarization corrections reach up to 0.6mev . note that when a given orbital crosses the fermi level , its plotted polarization correction jumps from negative to positive values , which is the result of the plotting convention explained in eqs . ( [ eq:80a ] ) . in fact , the qrpa polarization corrections to quasiparticle energies are always negative . for the sif contributions , cf . ( [ eq:86b ] ) and ( [ eq:86e ] ) , this fact is obvious , whereas for paired open - shell nuclei , smaller opposite - sign si contributions are unable to change this general rule . ( [ eq:102a ] ) plotted for quasiparticles closest to the fermi level in open - shell tin isotopes . , scaledwidth=40.0% ] in fig . [ fig:16 ] , for selected quasiparticles in open - shell tin isotopes , we show values of the average si corrections @xmath235 ( [ eq:102a ] ) . solid lines connect values obtained for quasiparticles at the fermi level , and dashed and dotted lines pertain to those just below and above the fermi level , respectively . it is interesting to see that for the quasiparticle at the fermi level , the si corrections become fairly small , not exceeding 50kev , whereas away from the fermi level they can be of the order of 200kev , see also figs . [ fig:9a ] and [ fig:9b ] . if this observation is confirmed or derived in a systematic way , we can hope that near ground states of odd open - shell nuclei , the effects of si energies might be small . this is important , because the odd - even mass staggering , where masses of odd nuclei enter , is routinely used to gauge the intensity of pairing correlations . in the present study , we investigated links between the mean - field and polarization - correction approaches to masses of odd nuclei . the former ones are rooted in the energy - density - functional methods and strive to describe odd systems by blocking odd particle or quasiparticle states . energies of odd nuclei are then obtained by minimization methods , in full analogy with those used in even nuclei , and by employing the same energy density functionals . the latter ones are based on the perturbative `` diagonal '' coupling between the odd particle and vibrational phonons calculated in even systems . following the classic analyses presented in refs . @xcite , we derived links between these two classes of approaches also in the case of density - dependent interactions and/or paired systems . this allowed us to show limitations of the polarization - correction methods as compared to the full `` non - diagonal '' particle - vibration - coupling that is rooted in the many - body perturbation theory . we performed numerical analyses by using the deformed mean - field code ( v2.52k ) @xcite , which is able to solve self - consistent equations in odd nuclei by breaking all symmetries . the polarization corrections were independently calculated by using the spherical code @xcite , which has the capability to solve efficiently the ( q)rpa equations . the comparison of results allowed us to identify the reason of discrepancies between the masses of odd nuclei calculated with these two approaches , which turns out to be the self - interaction energy , polluting the mean - field energies of odd nuclei . our derivations also allowed us to explicitly define and calculate the self - interaction energies , which then can be subtracted from the mean - field results leading to self - interaction - free masses . fruitful and inspiring discussions with gianluca col are gratefully acknowledged . this work was supported in part by the thexo jra within the eu - fp7-ia project ensar ( no . 262010 ) , academy of finland and university of jyvskyl within the fidipro programme , polish national science center under contract no.2012/07/b / st2/03907 , and bulgarian science fund under contract no . nupnet - sarfen dns7rp01/0003 . we acknowledge the csc - it center for science ltd , finland , for the allocation of computational resources . in this appendix , we specify the final equations obtained for the si energy ( [ eq:89 ] ) and polarization correction ( [ eq:86b ] ) to the case of spherical symmetry , for which the numerical analyses of this work were performed . first we note that the use of spherical symmetry does not mean that in the odd system the spherical symmetry is conserved . indeed , in the even system , the quasiparticles move in the spherical field and thus their wave functions are characterized by quantum numbers @xmath236 that comprise the principal quantum number @xmath237 , angular momentum @xmath238 , and its projection @xmath233 . quasiparticles having different projections @xmath233 are degenerate , and , therefore , any linear combination of them can be used as the blocked orbital . in this work , we make a simplifying assumption that the blocked quasiparticle corresponds to a specific single value of the projection @xmath233 . the general case could have been treated equally easy , and at the end of the appendix we discuss the meaning of it . an odd state , obtained by blocking a quasiparticle , becomes necessarily deformed . in the calculations performed with the deformed code , this is particularly well and explicitly visible , as the self - consistent solutions obtained in odd systems always acquire small but non - zero deformations . the aim of this appendix is to show in which way the deformation , and the dependence of final results on the values of @xmath233 , appears in the qrpa calculations that are performed in the spherical basis and by using the spherical code . similarly as for the angular - momentum - projected deformed states @xcite , we can write the deformed density matrix @xmath246 and pairing tensor @xmath247 as sums of those projected on good angular momentum @xmath207 and its projections on the laboratory axis @xmath248 and on the intrinsic axis @xmath249 , @xmath250 and @xmath251 , that is , where only the @xmath253 terms appear in the expansion @xcite . by using the wigner - eckart theorem @xcite , one can always express laboratory spherical - tensor matrices , @xmath250 and @xmath251 , corresponding to quantum numbers @xmath254 , as , at this point , we can use the fact that the spherical - basis properties of mean fields are exactly the same as those of densities , that is , eqs . ( [ eq:98])([eq:94 ] ) hold equally well for @xmath246 and @xmath247 replaced by @xmath166 and @xmath167 , respectively . then , by summing the clebsh - gordan coefficients again , traces in eq . ( [ eq:89 ] ) can be explicitly evaluated , which gives , @xmath257 similarly , we can evaluate the qrpa sif polarization correction of eq . ( [ eq:86b ] ) . since the spherical qrpa amplitudes @xmath258 and @xmath259 can be labeled with the good quantum numbers @xmath254 , we have from the wigner - eckart theorem , amplitudes @xmath261 and @xmath262 read @xmath263 by using the fact that the spherical - basis properties of quasiparticle matrix @xmath194 are the same as those of @xmath247 , see eqs . ( [ eq:98f ] ) , ( [ eq:98d ] ) , and ( [ eq:94b ] ) , we can derive that @xmath264 finally , we note that eqs . ( [ eq:98])([eq:86a ] ) hold for an arbitrary blocked quasiparticle . however , when the reduced matrix elements ( [ eq:94 ] ) are evaluated for a specific quasiparticle ( [ spherical ] ) that has a fixed value of projection @xmath233 , we obtain , in any channel @xmath207 , the results depend on @xmath233 only through the clebsh - gordan coefficient @xmath267 . it is , therefore , advantageous to define triple reduced matrix elements that do not depend on @xmath233 , we note that factors @xmath273 , which are present in eqs . ( [ eq:96 ] ) , allowed for reducing eq . ( [ eq:101f ] ) to terms with @xmath274 only . however , fields ( [ eq:117x ] ) are not restricted to @xmath275 and thus the qrpa sif corrections ( [ eq:86e ] ) must be summed up over @xmath245 and @xmath276 . from expressions ( [ eq:101faa ] ) we see that the polarization corrections calculated for orbitals with given values of @xmath233 are obtained by folding the @xmath207-dependent ( but @xmath233-independent ) contributions @xmath277 and @xmath278 with simple clebsh - gordan coefficients . values of these coefficients thus dictate how strongly a given channel @xmath207 contributes . moreover , owing to the summation properties of the clebsh - gordan coefficients , the average contributions read , as shown in this appendix , by blocking quasiparticles that have fixed values of projections @xmath233 , one obtains only the @xmath266 terms in densities and fields , that is , deformations of odd systems are axial . it is also clear that by blocking quasiparticles with mixed values of @xmath233 , one would have obtained non - zero reduced matrix elements for non - zero values of @xmath249 , and thus in odd systems , non - axial deformations would have appeared . numerical results presented in this study indicate , however , that axial solutions have systematically lower energies .	background : : : models based on using perturbative polarization corrections and mean - field blocking approximation give conflicting results for masses of odd nuclei . purpose : : : systematically investigate the polarization and mean - field models , implemented within self - consistent approaches that use identical interactions and model spaces , so as to find reasons for the conflicts between them . methods : : : for density - dependent interactions and with pairing correlations included , we derive and study links between the mean - field and polarization results obtained for energies of odd nuclei . we also identify and discuss differences between the polarization - correction and full particle - vibration - coupling ( pvc ) models . numerical calculations are performed for the mean - field ground - state properties of deformed odd nuclei and then compared to the polarization corrections determined by using the approach that conserves spherical symmetry . results : : : we have identified and numerically evaluated self - interaction ( si ) energies that are at the origin of different results obtained within the mean - field and polarization - correction approaches . conclusions : : : mean - field energies of odd nuclei are polluted by the si energies , and this makes them different from those obtained by using polarization - correction methods . a comparison of both approaches allows for the identification and determination of the si terms , which then can be calculated and removed from the mean - field results , giving the self - interaction - free energies . the simplest deformed mean - field approach that does not break parity symmetry is unable to reproduce full pvc effects .
the achievement of bose - einstein condensation ( bec ) @xcite in trapped dilute atomic gases has stimulated a large interest in the physics of ultracold gases , both bosonic and fermionic ones . related with the latter , similar cooling methods as those employed to reach bec have been recently employed to accomplish a degenerate fermi gas , i.e. a fermi gas with temperature @xmath3 below the fermi temperature ( @xmath1 ) @xcite . for @xmath4 the fermi pressure becomes noticeable , and as a consequence the fermi cloud becomes significantly broader than a bec under the same temperature . additionally , for temperatures below a critical one , @xmath5 , the system should undergo a bardeen - cooper - schrieffer ( bcs ) transition @xcite . when this occurs the system becomes superfluid , due to the cooper pairing of particles near the fermi surface . the accomplishment of bcs pairing and its observation have been considered in detail @xcite . however , the employed cooling schemes have not yet allowed for temperatures below @xmath5 , which has been predicted to be much smaller than @xmath1 . those cooling methods rely on collisional processes , which in the degenerate regime are strongly suppressed due to pauli blocking . as a consequence , the cooling efficiency decreases , and technical losses have prevented until now from reaching temperatures well below @xmath1 @xcite . recently , however , it has been proposed that the latter obstacle could be overcome by employing feshbach resonances @xcite or optical lattices @xcite in order to increase the ratio @xmath6 up to @xmath7 , opening the possibility to achieve bcs in a temperature range where the collisional processes are still efficient enough . in fact , very recently temperatures @xmath8 have been reported by ohara _ et al . _ @xcite , which are below the predicted value of @xmath9 . in this paper we show that laser cooling could provide an effective and realistic tool to cool one- and two - component fermi gases well below @xmath1 . the proposed method is realized in the _ festina lente _ ( fl ) regime @xcite , in which the spontaneous emission rate @xmath10 is smaller than the trap frequency @xmath11 . in this regime the heating introduced by photon reabsorptions can be suppressed . in the absence of technical losses the major limitation to the laser cooling in the _ festina lente _ regime is set up by the finite value of lamb - dicke parameter , defined as the square root of the ratio of the recoil energy to @xmath12 . in practice , when the spontaneous emission rate becomes much smaller than the trap frequency , the finite available time sets limit on achievable temperatures , due to always present loss mechanism . reaching fl regime is experimentally feasible , and has been in fact demonstrated in experiments by the group of weiss @xcite , in which a suppression of reabsorptions for decreasing @xmath13 has been observed for tightly bound atoms in an optical lattice . the laser cooling in the fl regime down to the quantum degeneracy has been already predicted to work for bosons @xcite and for polarized fermions @xcite . the latter case is particularly interesting , since evaporative cooling can not be employed to a single - component fermionic system , due to the absence of @xmath14-wave scattering , while less restrictive , sympathetic cooling techniques @xcite , are also eventually limited due to pauli blocking . the bcs transition in such a system could be realized by externally inducing interatomic interactions , for instance dipole - dipole ones @xcite . in the present paper we analyze the three - dimensional laser cooling of fermi gases down to temperatures @xmath15 , by performing monte carlo simulations of the quantum dynamics governed by the master equation ( me ) . for @xmath3 comparable to @xmath1 , the laser cooling of fermions is affected by the statistical inhibition of the spontaneous emission , which results in a decrease of the cooling efficiency . in a previous work @xcite we have analyzed the laser cooling of a polarized fermi gas . we have theoretically predicted that the above mentioned problem can be overcome by either dynamically adjusting the spontaneous emission rate in a raman cooling process , or by employing appropriately designed anharmonic traps . we first consider the single - component case . in particular , we analyze the statistics of the laser - cooled gases , comparing the results with those expected for a thermal distribution . we show that whereas the mean - number distribution can be well approximated by a corresponding thermal one , the fluctuations significantly differ from the thermal case . this could have significant effects in the bcs pairing of these gases after an interatomic interaction is induced in the way discussed above . in the second part of this work , we analyze the laser - cooling of a two - component fermi gas . for this case , the collisions between different components can not be neglected . we consider them within the formalism of quantum boltzmann master equation ( qbme ) . this formalism is shown to produced the expected thermal distribution in the absence of laser cooling . we analyze the case in which only one of the fermionic species is laser - cooled , whereas the other one is sympathetically cooled as a result of the collisions between different components . we show that the laser - cooling is significantly favored by the presence of the second component . we analyzed also the statistics of the atom distribution , showing that the fluctuations at the fermi surface are significantly enhanced compared to the thermal case . finally , we analyze the heating induced by background collisions , and show that the laser cooling can be employed to maintain a two - component fermi gas at a fixed temperature for a relatively long time without significant atom losses . the paper is organized as follows . in section ii , we discuss our cooling scheme , and obtain the me that describes the raman cooling in the fl regime . we also describe in this section the corresponding me for the two - component case . we present the resulting rate equations and discuss their physical properties . in section iii we present our numerical results concerning the laser cooling of a single component gas , characterizing the statistical properties of the laser - cooled gas . in section iv we present our results for the case of a two component fermi gas , in which one of the components is laser - cooled . we discuss the statistics for this case , as well as the heating induced by losses , and its control using laser - cooling . we finalize in section v with some conclusions . we consider @xmath16 fermionic atoms with an accessible electronic three level @xmath17 scheme , with levels @xmath18 , @xmath19 and @xmath20 . we assume that the ground state @xmath18 is coupled via a raman transition to @xmath19 ( which is assumed metastable ) . another laser couples @xmath19 to the upper state @xmath20 , which rapidly decays into @xmath18 . after the adiabatic elimination of @xmath20 , a two level system is obtained , with an effective rabi frequency @xmath21 , and an effective spontaneous emission rate @xmath10 . the latter can be controlled by modifying the coupling from @xmath19 to @xmath20 . the atoms are confined in a non - isotropic dipole trap with frequencies @xmath22 , @xmath23 , different for the ground and the excited states , and non - commensurable one with another . the latter assumption simplifies enormously the dynamics of the spontaneous emission processes in the fl limit . when deriving the quantum me , it is important to cancel the off - diagonal terms of the density matrix , in order to reduce the me to a rate equation , which is possible to simulate numerically with monte carlo methods . the anisotropy of the trap does not need to be very large , and for practical purposes can be rather small . its effects in the rate equations is not relevant at all , and consequently in our simulation we consider the simple case of isotropic trap . the cooling process consists of sequences of raman pulses of appropriate frequencies , adjusted is such a way that they induce the transition of atoms to the lower motional states of the trap . we assume weak optical excitation , so that no significant population in @xmath19 is present . this allows to adiabatically eliminate @xmath19 , and consequently to consider only the density matrix @xmath24 describing all atoms in @xmath18 , and being diagonal in the fock representation corresponding to the bare trap levels . let us introduce the annihilation and creation operators of atoms in the ground ( excited ) state and in the trap level @xmath25 ( @xmath26 ) , which we denote by @xmath27 , @xmath28 ( @xmath29 , @xmath30 ) . these operators fulfill standard fermionic anticommutation relations : @xmath31 , @xmath32 . we use the standard theory of quantum - stochastic processes @xcite to develop the quantum me that governs the dynamics of cooling @xcite . the derivation of the me for fermions proceeds basically in a similar way to the corresponding derivation for bosons . the latter was presented in details in @xcite . the me that describes the dynamics of the density matrix @xmath33 of the atoms in the fl regime is @xmath34 where @xmath35 , \end{aligned}\ ] ] with @xmath36 here , @xmath37 is the single atom effective spontaneous emission rate , @xmath21 is the rabi frequency associated with the atom transition and the laser field , @xmath38 ( @xmath39 ) are the energies of the ground ( excited ) harmonic trap level @xmath25 ( @xmath26 ) , @xmath40 is the laser detuning from the atomic transition , @xmath41 , where @xmath42 is the fluorescence dipole pattern , and @xmath43 are the frank condon factors . in the following we assume that @xmath44 and @xmath45 , which allows for adiabatic elimination of excited state @xmath19 . to assure the fl regime we require also @xmath46 at least in the initial phase of the cooling process . the dynamics of the two - component system in the fl limit , where only the first component is laser cooled , may be described by the following me : @xmath47 the new term @xmath48 that appears in equation ( [ me1 ] ) describes the process of elastic collisions between the two species : @xmath49 , \end{aligned}\ ] ] where @xmath50 here , @xmath51 and @xmath52 are respectively the annihilation and creation operators of atoms of the second component in the trap state @xmath25 . since at low temperatures collisions between atoms are purely of @xmath14-wave type , the operator @xmath53 accounts only for collisions between atoms in the second component and those atoms in the ground electronic state @xmath18 of the first species . the collisions with the atoms in the @xmath19 can be neglected if the laser is sufficiently weak to assure that only a small fraction of atoms is excited in each pulse . the amplitude @xmath54 in equation ( [ hcoll ] ) reads @xmath55 where @xmath56 ( @xmath57 ) denotes the wavefunction of the @xmath58-th level of the @xmath18 , ( @xmath59 ) trap , and @xmath60 is the scattering length . the term @xmath61 from the me ( [ me1 ] ) , includes the effective hamiltonian @xmath62 , which in the presence of two components , accounts also for the energies of atoms of the second component @xmath63 for the assumed regime of parameters , the excited state may be adiabatically eliminated . to this end we use standard projection operator techniques @xcite . the corresponding calculus proceed in the same way as in the case of bosons , which was described in detail in @xcite . here , we present only the final results . for sufficiently weak interactions between the two species , the dynamics due to the collisions and the dynamics due to the laser cooling are independent @xcite . the process of laser cooling may be described by the me ( [ me ] ) without the collisional part , while the collisional part of dynamics after the adiabatic elimination takes the form of qbme , similar to the one formulated for bosons and presented in @xcite . the rate equations for the populations @xmath64 and @xmath65 in level @xmath58 and @xmath25 of the trap @xmath18 and @xmath59 respectively , are given by @xmath66 where @xmath67 is the rate of laser - induced transition from state @xmath25 to state @xmath58 of the @xmath18 trap @xmath68 ^ 2+\gamma^2 ( r_{l}+\xi_{lp})^2 } , \label{pmn}\end{aligned}\ ] ] with @xmath69 . the rate @xmath70 of a collision between two fermions of different species , occupying the states @xmath71 and @xmath72 of the trap respectively , resulting in a transition to the states @xmath58 and @xmath25 , is given by @xmath73 in the transition rate ( [ pmn ] ) one can clearly identify two contributions due to the quantum statistical character of fermions : ( a ) if the system becomes more degenerate , @xmath74 vanishes , and the atoms remain for very long times in the excited state @xmath26 , i.e. the spontaneous emission is inhibited . this affects negatively the cooling process in two ways : first , it prolongs the cooling , and second , the excited - ground collisions can occur and lead to heating and losses . in addition , the adiabatic elimination used in the derivation of equation . ( [ pmn ] ) ceases to be valid . ( b ) the fermionic inhibition factor @xmath75 appears in the numerator of the probabilities , introducing also a slowing - down of the cooling process . the negative influence of fermi statistics above can be overcome by dynamically changing of @xmath10 in raman cooling @xcite . one can increase it gradually during the cooling in order to avoid the inhibition effects , but still remaining in the fl regime . note that , the fl condition involves the spontaneous emission rate modified in the presence of other atoms : @xmath76 . still , even if one uses such approach some small fraction of the atoms will remain in the excited state after the cooling pulse , and has to be removed from the trap in order to avoid non - elastic collisions . the latter aim can be achieved by optically pumping the excited atoms to a third non - trapped level . this introduces a new loss mechanism which we take into account in the simulations . the calculations of the laser cooling were performed for the case of a three - dimensional harmonic trap . for simplicity we assumed that the trap is isotropic with frequency @xmath11 . the main difficulty for performing monte carlo simulations of the laser cooling for fermions is the large number of states in the calculation . since one state can be populated maximally by one fermion , in order to perform a simulation for a medium - size atomic sample ( in our case @xmath77 atoms ) , we had to consider rather large trap ( @xmath78 energy levels in our case ) . in addition if the simulation starts from a temperature above @xmath1 , the number of states must be much larger than the number of atoms in the trap . for an initial number of @xmath77 atoms , we calculated the cooling dynamics for the trap consisting of approximately @xmath79 states . for a single component gas of fermions , where the collisions between the atoms are absent , we were able to perform simulations treating each state of the trap separately . in this case the transition is only induced by the laser , and its rate given by ( [ pmn ] ) depends only on two states . the number of different probabilities that one have to recalculate after the change of the atomic distribution is roughly @xmath80 . in the case of the two component fermi gas , one has to account for the process of collisions , in which the transition rate ( [ pcoll ] ) depends on four numbers : two initial states and two final states . the number of different probabilities that one has to recalculate after a single transition takes place is roughly @xmath81 . the number of operations required to calculate the probabilities is orders of magnitude too large for numerical computations . therefore for simulating the cooling in two - species gas , we have applied the ergodic approximation . we assume that the populations of the states with the same energy are equal : @xmath82 , where @xmath83 , @xmath84 is the number of atoms of the @xmath58-th component occupying the energy shell @xmath85 and @xmath86 is the degeneracy of the energy shell @xmath85 . this approximation relies on the fact that for typical parameters , the collisional processes are much faster than the laser cooling . additionally the thermalization inside a single energy shell is much faster than the thermalization between energy shells of different energies . the assumption that collisions occur much more frequently than laser induced transitions , is necessary , since the laser cooling is not fully `` compatible '' with the ergodic approximation . the cooling is based on the process of absorption of a photon from one of three orthogonal laser beams , therefore the rate of transition depends on the three dimensional structure of the initial and the final state . due to this dependence , we have not been able to find a simpler formula , which would be a counterpart of equation ( [ pmn ] ) in ergodic approximation . in the case of the collisional process , however , one can further develop the rate ( [ pcoll ] ) in order to fully utilize the advantages of the ergodic approximation . the rate of collision , between fermions occupying the energy shells @xmath87 and @xmath88 , resulting in a transition to the energy shells @xmath89 and @xmath90 in the ergodic approximation is given by @xmath91 where the amplitude of the transition @xmath92 is defined as follows @xmath93 in principle , the amplitude of transition @xmath92 may be calculated directly from equation . ( [ uerg ] ) , by performing a summation of the terms @xmath94 given by formula ( [ udef ] ) . this requires , however , a lot of numerical efforts , since the number of operations is of the order @xmath95 , where @xmath96 is the number of states in the trap . this can be avoided , if the trap is isotropic , in which case we were able to derive a simplified expression @xmath97 where @xmath98 . the coefficients @xmath99 are defined by @xmath100 where @xmath101 denotes the @xmath58-th hermite polynomial . note that the summation over indices @xmath58 ( @xmath25 , @xmath71 , @xmath72 ) in equation ( [ uerg1 ] ) , runs from @xmath102 if @xmath103 ( @xmath104 , @xmath105 , @xmath106 ) is even , or from @xmath107 if @xmath103 ( @xmath104 , @xmath105 , @xmath106 ) is odd , up to @xmath103 ( @xmath104 , @xmath105 , @xmath106 ) with an increment equal to @xmath108 . in order to obtain equation ( [ uerg1 ] ) , we have used standard summation formulas for hermite polynomials @xcite . equation ( [ uerg1 ] ) can be further simplified , if the collisions involves distant energy shells , i.e. the minimum @xmath109 of the set of numbers @xmath110 is much smaller than its maximum @xmath111 . according to ref . @xcite , the following approximate formula holds @xmath112 with the help of the exact expression ( [ uerg1 ] ) , we have verified numerically that the approximation ( [ uappr ] ) gives correct results , already for @xmath113 . equation ( [ uappr ] ) leads to small errors for low lying energy shells , and for the transitions between neighboring shells . in the simulations , however , we have used the values calculated numerically with the help of the exact formula ( [ uerg1 ] ) . the numerical simulations of the cooling were performed using monte carlo techniques . in the calculation we assumed that the atoms are confined in a dipole trap , characterized by a lamb - dicke parameter @xmath114 , with @xmath115 being the size of the ground state of the trap , and @xmath116 the laser wavelength . in the context of laser cooling , an optical dipole trap was recently employed in experiments on all - optical production of a degenerate gas of two species of li @xcite , and it is also considered in current experiments in mg @xcite . due to numerical limitations we assumed lamb - dicke parameter @xmath117 @xcite . the assumed value of @xmath118 is not unrealistic , and can be realized for instance by trapping potassium atoms in a dipole trap with @xmath119 khz , which is the trapping frequency employed in experiments in @xmath120 @xcite , and a laser wavelength @xmath121 nm . in the calculation we use the above mentioned values to calculate real cooling time . we start our simulations with an initial number of atoms @xmath122 , which corresponds to a fermi energy @xmath123 . the three - dimensional isotropic harmonic trap contains @xmath124 energy levels , which corresponds to @xmath125 states . in the gas of fermions the main sources of losses are background collisions and photoassociation . background collisions result from non ideal vacuum conditions in the experiment , and introduce a relevant heating mechanism for temperatures well below fermi temperature @xcite . in the simulations we assumed a background collision rate @xmath126 hz @xcite . photoassociation losses ( when the laser is tuned between molecular resonances ) are typically of the order of @xmath127@xmath128/s for laser intensities of @xmath107mw/@xmath129 @xcite . in our case the laser intensities needed are typically @xmath130 times smaller and we estimate that for @xmath122 atoms , the atomic density is smaller than @xmath131@xmath132 . therefore the photoassociation losses can be safely neglected . the other type of losses which we account for in our calculations is the removal process of long - living excited atoms at the end of each cooling pulse . figure [ fig:1 ] shows the time - dependence of the temperature of a laser - cooled one - component gas of fermions . the inset shows the time dependence of the number of fermions , that decreases due to losses . as the initial state of the system we consider a thermal distribution with temperature @xmath133 . the cooling process was divided into three stages , each consisting of a sequence of two raman pulses . the employed pulses that are used are characterized by the following parameters : detuning @xmath134 , @xmath135 , @xmath136 respectively , rabi frequency @xmath137 , @xmath138 , @xmath139 respectively , and length @xmath140 , @xmath141 , @xmath142 respectively . the use of two pulses with slightly different detunings in each cooling stage allows to compensate for the oscillatory character of the frank - condon factors . the values of @xmath21 were chosen in such way that not more than @xmath143 of the atoms is excited during each pulse , in order to fulfill the condition of the adiabatic elimination . the temperature of the system was determined by fitting of the thermal distribution to the current distribution of fermions . at the end of cooling , the system reaches the temperature @xmath144 . time - dependence of the temperature during the laser cooling of a single component fermi gas by controlling the effective spontaneous raman rate . inset : time dependence of the number of atoms , which decreases due to background collisions and the removal of long - living excited atoms . the different raman cooling stages discussed in the text extend from @xmath102 @xmath14 to @xmath108 @xmath14 , from @xmath108 @xmath14 to @xmath145 @xmath14 , and from @xmath145 @xmath14 to @xmath146 @xmath14.,width=302 ] time - dependence of the effective single - atom spontaneous emission rate for the case considered in fig . [ fig:1 ] . the values for the first pulse in the sequence ( solid line with diamonds ) and those for the second pulse ( solid line with squares ) are indicated . the different raman cooling stages discussed in the text extend from @xmath102 @xmath14 to @xmath108 @xmath14 , from @xmath108 @xmath14 to @xmath145 @xmath14 , and from @xmath145 @xmath14 to @xmath146 @xmath14.,width=302 ] the time dependence of the spontaneous emission rate is presented in figure [ fig:2 ] . the two curves depict the values for the two pulses used at each stage of cooling . from this figure one can see how @xmath10 was controlled in order to speed up the cooling , and at the same time avoid inhibition of the spontaneous emission , and remain in the _ festina lente _ regime . in the simulation it was realized by increasing of @xmath10 gradually during the cooling process , in such a way that the average value of the spontaneous emission rate calculated for many - atom system : @xmath147 , remains approximately constant , and equal to @xmath148 , except for the last pulses . in the last cooling stage rather long pulses were applied , and therefore @xmath147 could be considered smaller than in the other pulses . the expectation value of @xmath147 is presented on figure [ fig:3 ] . it was calculated at the end of each pulse , averaging over all possible laser - induced transition that may take place . with each transition @xmath149 one can associate the spontaneous emission rate @xmath150 , where @xmath26 denote an intermediate state , when the atom is excited . time - dependence of the expectation value of the spontaneous emission rate for the case considered in fig . [ fig:1 ] . the spontaneous emission rate @xmath147 takes into account the inhibition due to the presence of the other fermions in the trap . the average was calculated for the atomic distribution after each cooling pulse . the different raman cooling stages discussed in the text extend from @xmath102 @xmath14 to @xmath108 @xmath14 , from @xmath108 @xmath14 to @xmath145 @xmath14 , and from @xmath145 @xmath14 to @xmath146 @xmath14.,width=302 ] mean distribution of fermions ( bars ) obtained by averaging in time during the prolonged last stage of the cooling process presented in fig . [ fig:1 ] . solid line - fit of the thermal distribution . , width=302 ] fluctuations of the distribution of fermions ( bars ) obtained by averaging in time during the prolonged last stage of the cooling process presented in fig . [ fig:1 ] . solid line - fluctuations of the thermal gas , with the temperature and the chemical potential determined from the fit of the mean distribution . , width=302 ] at the end of the cooling process , the temperature of the gas of fermions stabilizes and the system reaches an equilibrium distribution . in the absence of collisions , this distribution is purely determined by the laser induced transitions , and in general may be very different than the thermal distribution . in order to investigate this point , we studied the mean population and the fluctuations of the energy shells . the statistics was calculated by time - averaging during the last stage of cooling . to this end , the final stage of cooling was extended to the time @xmath151 . this time correspond , in average to four absorption and spontaneous emission cycles for each atom . the average distribution of atoms is plotted in figure [ fig:4 ] . the bars depict the distribution obtained from the simulation , whereas the solid line shows the fit of the thermal distribution . according to figure [ fig:4 ] , the mean population numbers are well described in the framework of the grand canonical ensemble . the fluctuations of the occupation numbers are presented in figure [ fig:5 ] . the solid line as before represents the fluctuations of the thermal gas , with the temperature and the chemical potential determined from the fit of the mean distribution . as can be seen , the fluctuations in the laser cooled gas have completely different properties than the fluctuations in the thermal distribution . in a small energy range below @xmath152 , the fluctuations are enhanced , whereas deeply in the fermi sea they completely vanish . from this peculiar behavior we may conclude that the laser cooling strongly affects the atoms near the fermi surface . on the other hand , it does not excite those atoms deeply in the fermi sea , due to the large negative detunings of the employed pulses . the existence of such non - thermal fluctuations is a peculiar feature of the proposed cooling method . the observed enhancement of the fluctuation near the fermi energy could have an important influence on the formation of cooper pairs in laser cooled fermi gases . this question , however , lies beyond the scope of this paper and it will be the subject of further investigations . it is worth stressing that our cooling scheme can be applied to a fermi mixture with scattering length modified in the vicinity of a feshbach resonance . with the help of this technique , the critical temperature for the bcs transition is predicted to reach @xmath153-@xmath154 @xmath1 @xcite . laser cooling should then allow to achieve such temperatures , presumably in a time scale shorter that @xmath107 @xmath14 ( see figure [ fig:1 ] ) . in the simulation of laser cooling in the two component fermi system we considered the same parameters as in the case of the single component gas . we assumed that the two species are placed in the same isotropic optical trap with frequency @xmath119 khz . the lamb - dicke parameter is @xmath155 and the wavelength of the cooling laser is @xmath121 nm . the trap considered in the simulation has the same depth for both species , and contains @xmath124 energy levels . the losses due to the background collisions affect in this case both components . in the simulations we assumed the same background - collision rate @xmath126 hz for both species . the two components have equal initial number of atoms @xmath122 . the scattering length for the interactions between the two species is equal to @xmath156 , where @xmath157 is the bohr radius . this value correspond to the interactions between two spin states @xmath158 and @xmath159 of @xmath160k @xcite . time dependence of the temperatures of the two components during the evolution in the presence of collisions only . the inset presents the final distribution of the first component ( upper plot ) and of the second component ( lower plot).,width=302 ] our first aim was to investigate the dynamics and the properties of the equilibrium distribution of the two component fermi gas where the collisions are described by the qbme , and the resulting probabilities ( [ pcoll ] ) . in this case we do not take into account the laser cooling mechanism . the energy levels in the trap are initially populated according to a thermal distribution with temperature @xmath161 . then , we let the system evolve in the presence of collisions during a time @xmath162 ms , corresponding to about @xmath163 collisions per atom in average . the time dependence of the temperatures of the two components is presented in figure [ fig:6 ] , whereas the inset shows their final distributions . as one can notice , the temperature in the system slightly fluctuates , but its mean value remains constant during the evolution . therefore the final equilibrium distribution can not be very different from the thermal distribution . the detailed comparison may be carried out on the basis of the mean populations and the fluctuations of the energy shells , depicted in figures [ fig:7 ] and [ fig:8 ] respectively . the bars present the averages calculated in the simulation , during the whole time of evolution . the solid lines depict the corresponding values for the thermal distribution . we see , that both the mean values and fluctuations in the two component system are very close to the functions calculated for a thermal distribution . this means that the qbme correctly describes the process of thermalization . in addition we conclude that for @xmath122 atoms , finite size effects are not observed , and the distributions correspond to the ones derived in the thermodynamic limit . average distribution of atoms of the first component calculated during the evolution in the presence of collisions . solid line - fit of the thermal distribution.,width=302 ] fluctuations of the atomic distribution for the first component , calculated during the evolution in the presence of collisions . solid line - fluctuations of the thermal gas with the temperature and the chemical potential determined from the fit of the mean distribution.,width=302 ] let us now discuss the results calculated with the inclusion of the cooling mechanism . the time dependence of the temperature for the cooling of the two - component gas of fermions is presented in figure [ fig:9 ] . the inset shows the number of fermions as a function of time . the initial state of the system is given by a thermal distribution with temperature @xmath164 . as for the case of single component fermi gas , the cooling process was divided into three stages , each consisting of a sequence of two raman pulses . the pulses that are used are characterized by the following parameters : detuning @xmath134 , @xmath135 , @xmath136 respectively , rabi frequency @xmath165 , @xmath166 , @xmath167 respectively , and length @xmath140 , @xmath168 , @xmath169 respectively . for the considered parameters , not more than @xmath143 of the atoms is excited during each pulse . as one can observe , a final temperature @xmath170 may be reached within @xmath171 s. time - dependence of the temperatures of the two components ( darker and lighter curves ) during the cooling of a two - component fermi gas by controlling the effective spontaneous emission rate . inset : time dependence of the number of atoms . the losses are due to background collisions and the removal of long - living excited atoms . the different raman cooling stages discussed in the text extend from @xmath102 @xmath14 to @xmath172 @xmath14 , from @xmath172 @xmath14 to @xmath173 @xmath14 , and from @xmath173 @xmath14 to @xmath174 @xmath14.,width=302 ] we have investigated also the statistical properties of the equilibrium distribution in the two - component system , in the presence of both laser - cooling and collisions . the results are presented in figure [ fig:10 ] ( mean occupation numbers ) and in figure [ fig:11 ] ( fluctuations ) . the bars represent the numerical results , calculated by time averaging during the last stage of the cooling . the solid curves depict the corresponding functions for the thermal distribution , with the temperature and the chemical potential determined from the fit of the average distribution . the mean occupation numbers are exactly the same as predicted by the thermal distribution . the fluctuations in the laser cooled system , however , are substantially larger than the fluctuations in thermal ensemble . the laser - cooling process induces additional migration of the atoms in the region close to the fermi energy , leading to the increased fluctuations in this region . finally we have studied the storage of the degenerate fermi gas produced in the cooling process . the main limitation in this case is provided by the background losses , which may generate holes deep within the fermi sea . the latter increases the corresponding energy per particle , and via collisional thermalization leads to an important source of heating @xcite . we calculate the dynamics of the system , which was already cooled to a temperature @xmath175 . the time dependence of the temperature is presented in figure [ fig:12 ] . the plot compares two cases : ( i ) the laser is turned off , and the gas is heated due to the background collisions , ( ii ) the laser is turned on , and the cooling pulses of the last stage are continuously applied . as we may observe from the figure , in the latter case , the laser cooling compensate for the heating induced by the creation of holes in the degenerate distribution . hence , it helps to maintain the degenerate gas for a relatively long time in the trap . we have also verified that the presence of laser - cooling does not lead to substantially larger losses . mean final distribution of fermions ( bars ) obtained by averaging in time during the last stage of the cooling process presented in fig . [ fig:9 ] . solid line - fit of the thermal distribution . , width=302 ] fluctuations of the final distribution of fermions ( bars ) obtained by averaging in time during the last stage of the cooling process presented in fig . [ fig:9 ] . solid line - fluctuations of the thermal gas , with the temperature and the chemical potential determined from the fit of the mean distribution . , width=302 ] in this paper we have analyzed the laser cooling of trapped one and two - component fermi gases . we have developed the quantum master equation that describes the laser cooling of fermions , and the process of collisions in the case of the two - component gas . we have considered the fl regime where the heating due to photon reabsorptions can be neglected . we have shown that the inhibition of the spontaneous emission due to pauli blocking can be overcome employing collective raman cooling with dynamically modified spontaneous raman rate . additionally , a slightly anharmonic trap could be employed . by means of monte carlo simulations of the corresponding rate equations , we have shown that it is possible to cool one- and two - component fermi gases down to temperatures of the order of a few percent of @xmath1 . the results obtained in this paper present by no means fundamental limits of laser cooling ; temperatures below 1% of @xmath1 can be achieved by optimizing the cooling procedure @xcite . the laser cooling is an external process that influences the dynamics of trapped atoms , therefore one can expect that the equilibrium distribution of the laser - cooled system may differ from the thermal distribution . we have analyzed , both for the single- and the two - component case , the statistics of the laser - cooled gas , comparing the results to the predictions of the grand canonical ensemble . we have shown that although the mean - number distribution can be well approximated by a thermal one , the fluctuations are significantly different . in particular , the fluctuations are considerably enhanced at the fermi surface . this effect could play a significant role in the bcs pairing in laser cooled gases , and it will be the subject of further analysis . additionally , we have discussed the possible loss sources that may affect the laser cooled gas . in particular , for the two - component gas , the creation of holes deep in the fermi sea due to background collisions could lead to a significant heating of the sample , which should importantly reduce the life - time of the degenerate gas . in this sense , we have shown that our laser - cooling scheme can be employed to maintain a two - component fermi gas at a fixed temperature without significant additional atom losses . heating due to background collisions in the degenerate two component fermi system . the background collisions rate is assumed to be @xmath176 hz . the temperature of the system slowly increases when the cooling lasers are turned off . in the presence of the cooling lasers the temperature of the system remain unchanged.,width=302 ] we acknowledge support from the alexander von humboldt stiftung , the deutsche forschungsgemeinschaft , the rtn `` cold quantum gases '' , the esf program bec2000 + , the russian foundation for fundamental research , the polish kbn grant no . 5-p03b-103 - 20 , and ist program `` equip '' . stoof and m. houbiers , in _ bose - einstein condensation in atomic gases _ , varenna school , 1999 , course cxl ; l. you and m. marinescu , phys . a * 60 * , 2324 ( 1999 ) ; m.a . et al . _ , phys . a * 66 * , 013606 ( 2002 ) . in particular , it would be difficult to realize the _ festina lente _ cooling for very large lamb - dicke parameters ( @xmath177 ) , since the smallness of @xmath11 , could set unfeasible requirements on @xmath10 .	the collective raman cooling of trapped one- and two - component fermi gases is considered . we obtain the quantum master equation that describes the laser cooling in the _ festina lente _ regime , for which the heating due to photon reabsorption can be neglected . for the two - component case the collisional processes are described within the formalism of quantum boltzmann master equation . the inhibition of the spontaneous emission can be overcome by properly adjusting the spontaneous raman rate during the cooling . our numerical results based in monte carlo simulations of the corresponding rate equations , show that three - dimensional temperatures of the order of @xmath0 @xmath1 ( single - component ) and @xmath2 @xmath1 ( two - component ) can be achieved . we investigate the statistical properties of the equilibrium distribution of the laser - cooled gas , showing that the number fluctuations are enhanced compared with the thermal distribution close to the fermi surface . finally , we analyze the heating related to the background losses , concluding that our laser - cooling scheme should maintain the temperature of the gas without significant additional losses .
lisa - laser interferometric space antenna - is a proposed esa - nasa mission which will use coherent laser beams exchanged between three identical spacecraft forming a giant ( almost ) equilateral triangle of side @xmath0 kilometres to observe and detect low frequency cosmic gw @xcite . laser frequency noise dominates the other secondary noises , such as optical path noise , acceleration noise by 7 or 8 orders of magnitude , and must be removed if lisa is to achieve the required sensitivity of @xmath1 , where @xmath2 is the metric perturbation caused by a gravitational wave . in lisa , six data streams arise from the exchange of laser beams between the three spacecraft approximately 5 million km apart . these six streams produce redundancy in the data which can be used to suppress the laser frequency noise by the technique called time - delay interferometry ( tdi ) in which the six data streams are combined with appropriate time - delays @xcite . a mathematical foundation for the tdi problem for the static lisa was given in @xcite , where it was shown that the data combinations cancelling laser frequency noise formed the _ module of syzygies _ over the polynomial ring of time - delay operators . for the static lisa , the polynomial ring was in three variables and also commutative . the generators of the module were obtained via grbner basis methods . this scheme can be extended in a straight forward way to include the ( sagnac ) effect arising from the rotation of lisa , where the up - down optical links are unequal and as a consequence , now one has six optical links , but the armlengths are still constant in time . these are the modified ist generation tdi . here the polynomial ring is still commutative although over six indeterminates . the generators of the module were obtained in @xcite . in the general case of time varying armlengths , the polynomial ring is in six variables ( the six optical links ) but now it is noncommutative . the algebraic problem in the general case has been discussed in @xcite , but the solution to the general problem seems extremely difficult . we still have a linear system which leads to a module ( a left module ) , but it seems difficult to obtain its generators in general . in the non - commutative case , even the grbner basis algorithm may not terminate . we must envisage the possibility that not all optical links of lisa will be operating at all times for various reasons like technical failure for instance or even the operating costs . therefore , it is important to discuss the question when not all the links operate . here we look at a specific situation where the data is available from only two arms or four optical links . this should not much affect the information that can be extracted from the data , because this is essentially a michelson configuration which is known to be quite useful . the practical advantage for this case is that the algebraic problem simplifies considerably and therefore becomes tractable . we can reduce the problem to that of only one linear constraint on two polynomials , although the equation is still noncommutative . we show that we can generate an infinity of solutions by using a combinatorial algebraic approach which lists all such solutions in a systematic way . the solutions - that is the laser frequency noise is suppressed for these time - delayed combinations - are approximate in the sense that the @xmath3 and @xmath4 terms are ignored in the calculation , where @xmath5 is the generic length of the lisa s arm or the optical link and the dot denotes derivative with respect to time . the solutions are based on vanishing commutators ( in this approximation ) . we enumerate such commutators and for each such commutator there is a corresponding solution . we present an algorithm to construct these solutions . a geometric combinatorial approach was adopted in @xcite where several solutions were exhibited . our approach is algebraic where several operations are algebraic operations on strings of operators . the algebraic approach has the advantage of easy manipulation of data streams , although some geometrical insight could be at a premium . the analysis presented here can be used also for other future space detectors like astrod where again a similar situation arises @xcite . in this section we recapitulate the results obtained in the earlier literature already mentioned and also set up notation . in the literature on tdi of lisa there are several notations - here we choose the one which seems simplest for our purpose . we follow the notation and conventions of @xcite and @xcite . the six links are denoted by @xmath6 . the time - delay operator for the link @xmath7 from s / c 1 to s / c 2 or @xmath8 is denoted by @xmath9 in @xcite and so on in a cyclic fashion . the delay operators in the other sense are denoted by @xmath10 ; the link @xmath11 from @xmath12 by @xmath13 and similarly the links @xmath14 are defined through cyclic permutation . figure [ schematic ] depicts the optical links as described . let @xmath15 represent the laser frequency noise in s / c @xmath16 . let @xmath17 be the delay operator corresponding to the variable armlength @xmath18 , i.e. @xmath19 . then we have , u^1 = c_1 - z c_3 , + v^1 = l c_2 - c_1 . the other links in terms of @xmath20 are obtained by cyclic permutations . also in the @xmath21 we have not included contributions from the secondary noises , gravitational wave signal etc . since here our aim is to deal with laser frequency noise only . any observable @xmath22 is written as : x = p_i v^i + q_i u^i , where @xmath23 are polynomials in the variables @xmath24 . thus @xmath22 is specified by giving the six tuple polynomial vector @xmath25 . writing out the @xmath26 in terms of the laser noises @xmath20 , and in order that the laser frequency noise cancel for arbitrary functions @xmath20 , the polynomials @xmath25 must satisfy the equations : p_1 - q_1 + q_2 x - p_3 n & = & 0 , + p_2 - q_2 + q_3 y - p_1 l & = & 0 , + p_3 - q_3 + q_1 z - p_2 m & = & 0 , [ lneq ] where one must remember that the order of the operators is important when the armlengths are time - varying . the solutions ( if they exist ) to eq . ( [ lneq ] ) are the second generation tdi . one finds that at least in the approximation we consider they do exist . for a reasonably optimised model of lisa , say given in @xcite , @xmath27 metres / sec@xmath28 and thus even if one considers say 20 successive optical paths , that is , about @xmath29 seconds of light travel time , @xmath30 metre . this is well below few metres and thus can be neglected in the residual laser noise computation . moreover , @xmath31 terms ( and higher order ) can be dropped since they are of the order of @xmath32 ( they come with a factor @xmath33 ) which is much smaller than 1 part in @xmath34 , which is the level at which the laser frequency noise must be cancelled . thus , we keep terms only to the first degree in @xmath35 and also neglect higher time derivative terms in @xmath36 . the solutions to these equations are the polynomial vectors @xmath37 and they form a left module @xmath38 over the non - commutative ring @xmath39 , where @xmath40 is the field of rational numbers and @xmath41 are the indeterminates . a left module means that one can multiply a solution @xmath25 from the _ left _ by any polynomial in @xmath42 , then it is also a solution to the equations ( [ lneq ] ) and therefore in the module @xmath38 . it is difficult to solve these equations in general for the non - commutative case . previously we have shown in @xcite that the ring @xmath42 can be reduced to the quotient ring @xmath43 which is defined by quotienting @xmath42 by the ideal @xmath44 generated by vanishing commutators . consider commutators of the form @xmath45 $ ] where @xmath46 and @xmath47 or @xmath48 represents any one of the delay operators @xmath41 . upto the order of approximation we work in : = _ k = 1^n l_x_k _ m = 1^n _ y_m - _ m = 1^n l_y_m _ k = 1^n _ x_k . from this equation it immediately follows that if the set of operators @xmath49 is a permutation of the operators @xmath50 , then the commutator : = 0 , [ xct_cmtr ] upto the order of the approximation . the ideal @xmath44 is generated by such commutators . although we may be able to apply non - commutative grbner basis methods , the general solution seems quite difficult . however , we could consider a simpler problem of one arm of lisa being nonfunctional . then we have a simpler situation which turns out to be tractable . we arbitrarily choose the non - functional arm to be the one connecting s / c 2 and s / c 3 . this means from our labelling that the polynomials are now restricted to only the four variables @xmath51 and we can put the polynomials @xmath52 . this simplifies the equations ( [ lneq ] ) considerably . the second and third of the equations in ( [ lneq ] ) simplify to : q_2 & = & - p_1 l , + p_3 & = & - q_1 z . [ eqsimp ] substituting these in the first equation , we need to solve just one equation : p_1 ( 1 - lx ) - q_1 ( 1 - zn ) = 0 , to obtain the full polynomial vector as a solution . it is clear that solutions are of the michelson type denoted by @xmath22 . one such solution has already been given in the literature @xcite . this solution in our notation is : p_1 & = & 1 - zn - znlx + lx ( zn)^2 + q_1 & = & 1 - lx - lxzn + zn ( lx)^2 . [ x1 ] writing , = p_1 ( 1 - lx ) - q_1 ( 1 - zn ) , we get for ( [ x1 ] ) , @xmath53 $ ] which is a commutator that vanishes . thus it is an element of @xmath44 and ( [ x1 ] ) is a solution ( over the quotient ring ) . this is already well known . _ what we would like to emphasise is that there are more solutions of this type - in fact there are infinite number of such solutions . _ in what follows we give a systematic way to enumerate and list such solutions . notice that the operators @xmath54 and @xmath55 occur together . physically , these operators correspond to round trips . we denote these round trip operators by @xmath56 and @xmath57 , then we can write the previous equation and solution respectively , in a compact way as : p_1 ( 1 - a ) - q_1 ( 1 - b ) = 0 , [ 4lnk ] and , p_1 & = & 1 - b - ba + ab^2 , + q_1 & = & 1 - a - ab + ba^2 . [ x1 ] we can substitute this in @xmath58 to obtain @xmath59 $ ] which vanishes and hence ( [ x1 ] ) is a solution . but we can proceed further and write : p_1 & = & p_1 - b a^2 b , + q_1 & = & q_1 - a b^2 a , to obtain : = b a^2 ba - a b^2 ab . @xmath60 however does not vanish - it is also not a commutator - and hence @xmath61 is not a solution . continuing further with , p_1 & = & p_1 + a b^2 a b , + q_1 & = & q_1 + b a^2 b a , we get , @xmath62 which also does not vanish , again not a commutator . this process can be continued in such a way so that lower degree terms cancel and so as to get another vanishing commutator but of higher degree . the next commutator is obtained by adding two more terms of successively higher degree in a similar way . the result is : p_1 & = & 1 - b - ba + ab^2 - b a^2 b + a b^2 ab + ab^2 aba - ba^2 bab^2 , + q_1 & = & 1 - a - ab + ba^2 - a b^2 a + b a^2 ba + ba^2 bab - ab^2 aba^2 . [ x2 ] note that @xmath63 is obtained from @xmath64 by interchanging @xmath65 and @xmath66 and vice versa . this is a symmetry . the expressions for @xmath64 and @xmath63 in eq . ( [ x2 ] ) give the vanishing commutator : = [ a b^2 a , b a^2 b ] , and hence ( [ x2 ] ) is a solution . there are two more solutions with eight terms . one of them is : p_1 & = & 1 + a - b^2 - b^2 a - b^2 a^2 - b^2 a^3 + a^2 b^4 + a^2 b^4 a + & = & ( 1 - b^2 - b^2 a^2 + a^2 b^4)(1 + a ) , + q_1 & = & 1 + b - a^2 - a^2 b - a^2 b^2 - a^2 b^3 + b^2 a^4 + b^2 a^4 b + & = & ( 1 - a^2 - a^2 b^2 + b^2 a^4)(1 + b ) , with @xmath67 $ ] which is also a vanishing commutator . this solution looks something like the square of the solution ( [ x1 ] ) obtained by tinto et al ( see @xcite ) multiplied by linear factors in order ensure successive term cancellation . here one starts with @xmath68 and continues until one reaches the above commutator . the third solution corresponds to the commutator @xmath69 $ ] . this solution is given by : p_1 & = & 1 - b + ab - bab - baba + abab^2 - baba^2 b + abab^2 ab , + q_1 & = & 1 - a + ba - aba - abab + baba^2 - abab^2 a + baba^2 ba . the following observations are in order : the solutions correspond to the vanishing commutators as follows ; there is only one commutator at degree 4 , namely , @xmath70 $ ] ; of degree 8 , there are three commutators : @xmath71,~ [ abab , baba],~ [ a b^2 a , b a^2 b]$ ] . and these are the only three commutators at degree 8 . also there are equal number of @xmath65 s and @xmath66 s in each string of the commutators . to each of the vanishing commutators , there is a corresponding polynomial vector solution @xmath72 . the rest of the solution vector is obtained from eq.([eqsimp ] ) which then yields @xmath73 and @xmath74 . these are solutions only upto degree 7 . clearly one can construct solutions of higher degree . we do this in a combinatorial way in the next section . we enumerate the solutions in a given order - called the lexicographic order - and also present an algorithm to obtain the polynomials @xmath64 and @xmath63 . in order to proceed further , let us first define an interchange operator @xmath75 between the symbols @xmath65 and @xmath66 , that is , given a string @xmath76 composed of @xmath65 and @xmath66 , the string @xmath77 is obtained by replacing @xmath65 by @xmath66 in @xmath76 . a string is a polynomial with only one term usually also called a monomial in the literature . the length of a string is the degree of the monomial . we will use these terms interchangably , but that should not lead to any confusion . note that all the vanishing commutators which have appeared corresponding to a solution have the form @xmath78 $ ] . also since the string @xmath79 must be a permutation of the string @xmath76 , in order that the commutator @xmath78 $ ] vanish , there must be an equal number of @xmath65 s and @xmath66 s in the string @xmath76 of the vanishing commutator . so if we have @xmath80 number of @xmath65 s in the string @xmath76 , it must also have the same @xmath80 number of @xmath66 s , so that the total string @xmath76 has length @xmath81 . the commutator is then made up of strings of length @xmath82 . note that the operator @xmath75 can be extended to a polynomial containing strings in an obvious way by defining : i ( s_1 + s_2 + ... ) = i s_1 + i s_2 + ... . then we easily check that in the foregoing @xmath83 , where @xmath72 constitute a solution to eq . ( [ 4lnk ] ) . also @xmath84 the identity mapping . @xmath75 can be considered to be an operator over the polynomial ring over the indeterminates @xmath65 and @xmath66 . in the foregoing we have then considered cases for @xmath85 . for @xmath86 there is one solution , while for @xmath87 there are three distinct solutions . in order to proceed further , we must first list the commutators at each @xmath80 . in order to carry out this programme , we must introduce an order among the strings of length @xmath81 . we choose this order to be _ lexicographic _ or if we consider strings of all lengths , then _ length lexicographic _ @xcite . but since at a time we deal only with strings of a given length @xmath81 , the lexicographic order suffices . we arbitrarily choose @xmath88 . then at length 2 , @xmath89 . at length 3 , @xmath90 and so on . more generally , for strings of the same length , @xmath91 if at the first position from the left the strings differ , the symbol in string @xmath76 is less than the symbol in string @xmath92 . in this way we may list all strings in the symbols @xmath93 of a given length this is all we require . thus at length @xmath94 we have just one commutator @xmath95 $ ] . at @xmath96 , we have @xmath97 , [ abab , baba ] , [ abba , baab]$ ] , that is , three commutators ; we have also three solutions corresponding to each commutator . we have also taken care to write the first string in the commutator to be less than the second string in the lexicigraphic order chosen . note that we could list the commutator just by the first string , because the second string in the commutator is obtained by applying the operator @xmath75 to the first string . we will give an algorithm to obtain a solution , that is , the polynomials @xmath64 and @xmath63 in a unique way given the commutator . thus listing all the commutators of a string of length @xmath81 amounts to listing the solutions @xmath72 having that commutator . the polynomials @xmath64 and @xmath63 have @xmath81 terms and degree @xmath98 . we are now ready to list the commutators @xmath78 $ ] with first string @xmath76 of length @xmath81 . consider @xmath99 . the string @xmath76 has length 6 with 3 @xmath65 s and 3 @xmath66 s . listing the strings in lexicographical order we have : @xmath100 . we stop here because , the next string is @xmath101 and @xmath102 has occurred in the listing before and hence would lead to minus the commutator and therefore not an independent commutator . this follows from the symmetry encoded in the interchange operator @xmath75 . thus at @xmath99 there are 10 independent commutators . proceeding in this way we find that at @xmath103 there are 35 such commutators . the general formula for the number of commutators is @xmath104 . this formula follows from the following argument . consider a string of length @xmath81 , with @xmath80 number of @xmath65 s and @xmath66 s . because of symmetry between @xmath65 and @xmath66 , the first @xmath65 is fixed in the string ; as seen , the string starting with @xmath66 does not lead to a new commutator . thus we must choose @xmath105 combinations of @xmath65 s among @xmath98 available positions in the remaining string . this is the number of independent commutators which is then just @xmath104 . the algorithm is obtained by observing the way we got the solutions of degree 7 for @xmath87 . we started with the lowest degree and proceeded towards obtaining the commutator by cancelling terms at successively increasing degree . but now we have been assigned a commutator before hand and we need to find the solutions @xmath72 . we therefore need to go in the reverse order . we freely make use of the operator @xmath75 in this algorithm . consider the commutator @xmath106 = s_{2n } i ( s_{2n } ) - i ( s_{2n } ) s_{2n}$ ] where @xmath107 is a string containing @xmath80 number of @xmath65 s and an equal number of @xmath66 s . 1 . consider the first string in the commutator @xmath108 which we denote by @xmath109 . the subscript denotes the length of the string or the degree of the monomial . now , either @xmath110 ends in @xmath65 or @xmath66 ; that is , @xmath111 or @xmath112 . * if @xmath111 , then write @xmath113 or ; * if @xmath112 , then write @xmath114 . 2 . replace @xmath110 by @xmath115 and repeat . that is either we have @xmath116 or @xmath117 . if @xmath116 , then write @xmath118 otherwise write @xmath119 . 3 . continue this process till one arrives at a string of length 0 , namely , @xmath120 . note that @xmath121 . 4 . then @xmath122 and @xmath83 . note that @xmath123 are strings of length @xmath124 . @xmath64 and @xmath63 are polynomials of degree @xmath125 . to illustrate the algorithm explicitly , let us take the case for @xmath99 and choose one commutator among the 10 possible ones . let us randomly choose the string @xmath126 , so that the commutator is @xmath127 $ ] . so we have @xmath128 . thus @xmath129 . now @xmath130 ends in @xmath66 . thus we write @xmath131 , where @xmath132 . thus choosing the second option in step ( ii ) , we have @xmath133 . this is the highest degree term in the polynomial @xmath64 and for @xmath99 , it is of degree 11 . the next term is obtained from @xmath134 . again @xmath134 ends in @xmath66 . following the steps of the algorithm , @xmath135 and hence @xmath136 . since , @xmath137 ends in @xmath65 , @xmath138 . we then have @xmath139 and so on . we must carry out the steps till we reach degree 0 . this terminates the algorithm . in this way , we can easily construct the 11th degree polynomial @xmath64 by adding up all the strings @xmath140 . collecting all the terms and writing out the polynomial @xmath64 explicitly : p_1 & = & 1 - b + ab - bab - baba + abab^2 - baba^2b + abab^2ab + & - & baba^2bab + abab^2abab + abab^2ababa - baba^2babab^2 . the polynomial @xmath63 is obtained from @xmath64 by just interchanging @xmath65 and @xmath66 , that is , @xmath83 . the other polynomials @xmath74 and @xmath73 are obtained from eq . ( [ eqsimp ] ) . these polynomials then constitute the polynomial vector which is a solution and is in the module of solutions over the quotient ring @xmath43 . in terms of the time - delay operators @xmath51 , the polynomials @xmath64 and @xmath63 are of degree 22 , while @xmath74 and @xmath73 are polynomials of degree 23 . for a general value of @xmath80 , the solution contains polynomials of maximum degree @xmath141 in the time - delay operators . clearly , from the mathematical point of view we have an infinite family of solutions . note that here we do not lay any claim to any exhaustive listing of solutions . however , we believe this family of solutions is sufficiently rich , because we can form linear combinations of these solutions and then these form a left submodule of the full left - module of solutions . we can also subtract solutions like ( [ x1 ] ) from ( [ x2 ] ) which is also a solution with each component polynomial having less number of terms ( or links in the usual terminology ) albeit of the same degree . further , by using cyclic permutations on @xmath142 and @xmath143 we can generate more solutions . for example , we may set @xmath144 and @xmath57 and obtain solutions for the four links , @xmath145 . the linear combinations of these solutions with the previous ones in the variables @xmath146 are also solutions in the full module pertaining to all six links . thus we see that a plethora of solutions can be so generated . from the physical point of view , since we have neglected terms in @xmath3 and @xmath4 and higher orders in our calculations , a limit on the degree of the polynomial solutions arises . that is upto certain degree of the polynomials , we can safely assume the commutators to vanish . but as the degree of the polynomials increases it is not possible to neglect these higher order terms and then such a limit becomes important . we now investigate this limit and make a very rough estimate of it . our lisa model in @xcite gives @xmath147 m / sec@xmath28 . the other term @xmath148 is way too small to matter . so the limit is set by @xmath149 . from @xmath149 we compute the error in @xmath36 , namely , @xmath150 . if we allow the error to be no more than say 10 metres , then we find @xmath151 sec . each time - delay is about @xmath152 secs which gives the number of successive time - delays to be about 270 . this is the maximum degree of the polynomials . this means one can go upto @xmath153 . if we set the limit more stringently at @xmath154 metre , then the highest degree of the polynomial reduces to about 80 which means one can go upto @xmath155 . this means there are a large number of tdi observables available to do the physics . we however point out that when we go to higher degree polynomials , the number of interpolations of data also increases leading to accumulation of numerical errors . this is a relevant issue whenever high degree polynomials are used . in this work we have given a family of approximate solutions for the general model of lisa where the up - down links are unequal and the armlengths are time varying , but with only two arms functional . such a situation can arise when one of the arms is non - functional or when it is not possible to obtain data from all optical links because high costs etc . we then have only four links to deal with and although the general non - commutative problem of obtaining second generation tdi is extremely difficult , the problem here reduces to a much simpler and tractable one . we have therefore been able to exhibit a family of solutions although not exhaustive , is quite rich and interesting both from the mathematical and physical point of view . also the mathematical analysis presented here could be extended to other space missions such as astrod where a similar situation arises although the physical parameters will be quite different . the authors s. v. dhurandhar and j - y vinet would like to thank the indo - french centre for the promotion of advanced research ( ifcpar ) project no . 3504 - 1 under which this work has been carried out . armstrong j w 2008 _ lrr-2006 - 1 : http://relativity.livingreviews.org / articles / lrr-2008 - 2_. s. v. dhurandhar , k. rajesh nayak , j - y . vinet , _ phys . * d 65 * , 102002(2002 ) . k. r. nayak and j - y vinet , _ phys . _ d * 70 * , 102003 ( 2004 ) . s. v. dhurandhar , jpcs , * 154 * , 012047 ( 2009 ) . m. vallisneri , _ phys . _ d * 72 * , 04003 ( 2005 ) . et al , advances in space research , * 32 * , 1437 ( 2003 ) . s. v. dhurandhar , j - y . vinet and r. nayak , class . quantum grav . * 25 * , 245002 ( 2008 ) . b. j. keller , _ algorithms and orders for finding noncommutative grbner basis _ , thesis , viginia 1997 .	in order to attain the requisite sensitivity for lisa - a joint space mission of the esa and nasa- the laser frequency noise must be suppressed below the secondary noises such as the optical path noise , acceleration noise etc . by combining six appropriately time - delayed data streams containing fractional doppler shifts - a technique called time delay interferometry ( tdi ) - the laser frequency noise may be adequately suppressed . we consider the general model of lisa where the armlengths vary with time , so that second generation tdi are relevant . however , we must envisage the possibility , that not all the optical links of lisa will be operating at all times , and therefore , we here consider the case of lisa operating with two arms only . as shown earlier in the literature , obtaining even approximate solutions of tdi to the general problem is very difficult . since here only four optical links are relevant , the algebraic problem simplifies considerably . we are then able to exhibit a large number of solutions ( from mathematical point of view an infinite number ) and further present an algorithm to generate these solutions .
a _ contact structure _ on a three - dimensional manifold @xmath4 is a smooth field @xmath7 of tangent two - dimensional planes , locally defined as a kernel of a differential @xmath8-form @xmath9 with non - vanishing @xmath10 . a contact structure is _ cooriented _ if the two - dimensional planes defining the contact structure are continuously cooriented ( transversally oriented ) . a contact structure is _ parallelized _ if the two dimensional vector bundle @xmath11 over @xmath4 is trivialized . a contact structure @xmath12 on a manifold @xmath4 is said to be _ overtwisted _ if there exists a @xmath13-disk @xmath14 embedded into @xmath4 such that the boundary @xmath15 is tangent to @xmath12 while the disk @xmath14 is transverse to @xmath12 along @xmath15 . not overtwisted contact structures are called _ tight_. a curve in @xmath4 is an immersion of @xmath5 into @xmath4 . a curve in a contact manifold is called _ legendrian ( resp . transverse ) _ if it is everywhere tangent ( resp . transverse ) to the planes of the contact structure . if the contact structure is cooriented , then every legendrian curve has a natural framing . similarly if the contact structure is parallelized , then every transverse curve has a natural framing . this means that every component of the space of legendrian curves in a manifold @xmath4 with a cooriented contact structure is a subspace of a component of the space of framed curves . similarly every component of the space of transverse curves in a manifold @xmath4 with a parallelized contact structure is a subspace of a component of the space of framed curves . a legendrian ( resp . transverse ) knot is a legendrian ( resp . transverse ) embedding of @xmath5 into a contact manifold . in @xcite d. fuchs and s. tabachnikov proved that the quotients of the groups of @xmath16-valued order @xmath17 invariants by the groups of order @xmath18 invariants of legendrian , of transverse , and of framed knots in the standard contact @xmath19 are canonically isomorphic . an analogous result was obtained by j. w. hill @xcite in the case of legendrian and of framed knots in the spherical cotangent bundle of @xmath20 ( with the standard contact structure ) . to prove these facts they used the fact of the existence of the universal vassiliev invariant of framed knots in these spaces ( also known as kontsevich integral ) . recently the existence of such an invariant was shown by j. e. andersen , j. mattes , and n. reshetikhin @xcite for framed knots in the spaces that are the product of @xmath21 and an oriented compact surface with nonempty boundary . however for most manifolds the existence of such a universal vassiliev invariant is still unknown . in this paper we use a different approach . it allows us to generalize these isomorphisms to the case of finite order invariants ( with values in an arbitrary abelian group ) of legendrian , of transverse , and of framed knots in a big class of contact manifolds . for a closed orientable manifold @xmath4 admitting a metric of a negative sectional curvature we show that for an arbitrary cooriented contact structure on @xmath4 the groups of order @xmath17 invariants with values in an arbitrary abelian group of legendrian and of framed knots ( from the corresponding components of the spaces of legendrian and of framed curves ) are canonically isomorphic . for such manifolds with parallelized contact structure we show the same fact for the groups of order @xmath17 invariants of transverse and of framed knots . the same fact is proven in the case of zero - homologous legendrian , transverse , and framed knots in an arbitrary contact manifold with the parallelized contact structure for a contact @xmath22-manifold @xmath4 ( with the parallelized contact structure ) such that it is a total space of a locally trivial @xmath5- or @xmath21-fibration over a surface we show that the groups of order @xmath17 invariants ( with values in an arbitrary abelian group ) of legendrian , of transverse , and of framed knots are canonically isomorphic , provided that @xmath23 and @xmath4 is not the quotient of @xmath1 by the equivalence relation @xmath24 ( the last manifold @xmath5-fibers over @xmath25 ) . for an arbitrary cooriented contact structure on @xmath1 with a nonzero euler class of the contact bundle we construct examples of legendrian homotopic legendrian knots @xmath2 and @xmath3 in @xmath1 such that @xmath2 and @xmath3 realize isotopic framed knots but can be distinguished by order one invariants of legendrian knots in @xmath1 . we also construct such examples for a big class of contact manifolds @xmath4 such that @xmath4 is a total space of a locally trivial @xmath5-fibration over a nonorientable surface . this means that in general finite order invariants of legendrian and of framed knots are different , and finite order invariants of legendrian knots can distinguish legendrian knots that realize isotopic framed knots . we show that in some of these examples the restrictions of the contact structure to the complements of @xmath2 and of @xmath3 are overtwisted . this seems to suggest that there is a nontrivial theory of legendrian knots that realize isotopic framed knots and have overtwisted complements . a _ contact structure _ on a three - dimensional manifold @xmath4 is a smooth field @xmath7 of tangent two - dimensional planes , locally defined as a kernel of a differential @xmath8-form @xmath9 with non - vanishing @xmath10 . a manifold with the contact structure possesses canonical orientation determined by the volume form @xmath26 . the standard contact structure in @xmath19 is given by the @xmath8-form @xmath27 . a _ contact element _ on the manifold is a hyperplane in the tangent space to the manifold at a point . for a surface @xmath28 we denote by @xmath29 the space of all cooriented ( transversally oriented ) contact elements of @xmath28 . this space is a spherical cotangent bundle of @xmath28 . its natural contact structure is a distribution of tangent hyperplanes given by a condition that a velocity vector of an incidence point of a contact element belongs to the element . a contact structure is _ cooriented _ if the two - dimensional planes defining the contact structure are continuously cooriented ( transversally oriented ) . a contact structure is _ oriented _ if the two - dimensional planes defining the contact structure are continuously oriented . since every contact manifold has a natural orientation we see that every cooriented contact structure is naturally oriented and every oriented contact structure is naturally cooriented . a contact structure is _ parallelizable _ ( _ parallelized _ ) if the two - dimensional vector bundle @xmath11 over @xmath4 is trivializable ( trivialized ) . since every contact manifold has a canonical orientation , one can see that every parallelized contact structure is naturally cooriented . a contact structure @xmath12 on a manifold @xmath4 is said to be _ overtwisted _ if there exists a @xmath13-disk @xmath14 embedded into @xmath4 such that the boundary @xmath15 is tangent to @xmath12 while the disk @xmath14 is transverse to @xmath12 along @xmath15 . not overtwisted contact structures are called _ tight_. a _ curve _ in @xmath4 is an immersion of @xmath5 into @xmath4 . ( all curves have the natural orientation induced by the orientation of @xmath5 . ) a _ framed curve _ in @xmath4 is a curve in @xmath4 equipped with the continuous unit normal vector field . a _ legendrian curve _ in a contact manifold @xmath30 is a curve in @xmath4 that is everywhere tangent to @xmath12 . if the contact structure on @xmath4 is cooriented , then every legendrian curve has a natural framing given by the unit normals to the planes of the contact structure that point in the direction specified by the coorientation . for @xmath30 a contact manifold with a cooriented contact structure , we put @xmath31 to be the total space of the fiberwise spherization of the contact bundle , and we put @xmath32 to be the corresponding locally trivial @xmath5-fibration . the @xmath33-principle proved for the legendrian curves by m. gromov @xcite says that the space of legendrian curves in @xmath30 is weak homotopy equivalent to @xmath34 the space of free loops in @xmath31 . the equivalence is given by mapping a point of a legendrian curve to the point of @xmath31 corresponding to the direction of the velocity vector of the curve at this point . to a legendrian curve @xmath35 in the contact manifold with the parallelized contact structure one can associate an integer that is the number of revolutions of the direction of the velocity vector of @xmath35 ( with respect to the chosen frames in @xmath12 ) under traversing @xmath35 according to the orientation . this integer is called the _ maslov number _ of @xmath35 . the @xmath33-principle implies that the set of maslov numbers enumerates the set of the connected components of the space of legendrian curves in @xmath19 . a _ transverse _ curve in a contact manifold @xmath30 is a curve in @xmath4 that is everywhere transverse to @xmath12 . if the contact structure on @xmath4 is parallelized , then a transverse curve has a natural framing given by the unit normals corresponding to the projections on the tangent two - planes ( orthogonal to the velocity vectors of the curve ) of the first of the two coordinate vectors of the contact planes . a transverse curve in the contact manifold with a cooriented contact structures is said to be _ positive _ if at every point the velocity vector of the curve points into the coorienting half - plane , and it is said to be _ negative _ otherwise . a _ knot ( framed knot ) _ in @xmath4 is an embedding ( framed embedding ) of @xmath5 into @xmath4 . in a similar way we define legendrian and transverse knots in @xmath4 . a _ singular ( framed ) _ knot with @xmath36-double points is a curve ( framed curve ) in @xmath4 whose only singularities are @xmath36 ( transverse ) double points . isotopy _ of a singular ( framed ) knot with @xmath36 double points is a path in the space of singular ( framed ) knots with @xmath36 double points under which the preimages of the double points on @xmath5 change continuously . an @xmath6-valued framed ( resp . legendrian , resp . transverse ) knot invariant is an @xmath6-valued function on the set of the isotopy classes of framed ( resp . legendrian , resp . transverse ) knots . a transverse double point @xmath37 of a singular knot can be resolved in two essentially different ways . we say that a resolution of a double point is positive ( resp . negative ) if the tangent vector to the first strand , the tangent vector to the second strand , and the vector from the second strand to the first form the positive @xmath22-frame ( this does not depend on the order of the strands ) . ( if the singular knot is legendrian ( resp . transverse ) , then these resolution can be made in the category of legendrian ( resp . transverse ) knots . ) a singular framed ( resp . legendrian , resp . transverse ) knot @xmath38 with @xmath39 transverse double points admits @xmath40 possible resolutions of the double points . a sign of the resolution is put to be @xmath41 if the number of negatively resolved double points is even , and it is put to be @xmath42 otherwise . let @xmath43 be an @xmath6-valued invariant of framed ( resp . legendrian , resp . transverse ) knots . the invariant @xmath43 is said to be of _ finite order _ ( or _ vassiliev invariant _ ) if there exists a positive integer @xmath39 such that for any singular knot @xmath44 with @xmath39 transverse double points the sum ( with appropriate signs ) of the values of @xmath43 on the nonsingular knots obtained by the @xmath45 resolutions of the double points is zero . an invariant is said to be of order not greater than @xmath36 ( of order @xmath17 ) if @xmath36 can be chosen as integer in the above definition . the group of @xmath46-valued finite order invariants has an increasing filtration by the subgroups of the invariants of order @xmath17 . [ description]_description of legendrian and of transverse knots in @xmath19 . _ the contact darboux theorem says that every contact manifold locally looks like @xmath19 with the standard contact structure @xmath27 . transverse and legendrian knots are conveniently presented by the projections into the plane @xmath47 . identify a point @xmath48 with the point @xmath49 furnished with a fixed direction of an unoriented straight line through @xmath47 with the slope @xmath50 . then the curve in @xmath19 is a one parameter family of points with directions in @xmath20 . a curve in @xmath19 is transverse if and only if the corresponding curve in @xmath20 is never tangent to the chosen directions along itself . while a generic regular curve has a regular projection onto the @xmath47-plane , the projection of a generic legendrian curve onto the @xmath47-plane has isolated critical points ( since all the planes of the contact structure are parallel to the @xmath50-axis ) . hence the projection of a generic legendrian curve may have cusps . a curve in @xmath19 is legendrian if and only if the corresponding planar curve with cusps is everywhere tangent to the field of directions , in particular this field is determined by the curve with cusps . let @xmath30 be a contact manifold with a cooriented contact structure . let @xmath51 be a component of the space of legendrian curves in @xmath30 . since a legendrian curve in @xmath30 has a natural framing , we see that @xmath51 is a subspace of a connected component @xmath52 of the space of framed curves in @xmath4 . in @xcite d. fuchs and s. tabachnikov proved that @xmath16-valued finite order invariants of legendrian knots in the standard contact @xmath19 can not distinguish legendrian knots that belong to the same component @xmath51 of the space of legendrian curves in @xmath19 and realize isotopic framed knots . [ conditions ] as it was later observed by d. fuchs and s. tabachnikov @xcite , the proof of the theorem can be generalized to a large class of contact manifolds with cooriented contact structures . moreover it can be generalized to the case of finite order invariants with values in an arbitrary abelian group @xmath6 . in order to state the generalized theorem of d. fuchs and s. tabachnikov we introduce the following definitions . let @xmath30 be a contact manifold with a cooriented contact structure , let @xmath51 be a component of the space of legendrian curves in @xmath53 , and let @xmath52 be the corresponding component of the space of framed curves in @xmath4 . we say that @xmath52 _ satisfies condition _ @xmath54 if for any @xmath55 every framed knot @xmath56 is not isotopic to the framed knot @xmath57 obtained by adding @xmath58 positive extra twists to the framing of @xmath38 . we say that @xmath52 _ satisfies condition _ @xmath59 if it contains infinitely many components of the space of legendrian curves ( see proposition [ interpretationconditionii ] for the homological interpretation of condition ) . below we formulate the generalized theorem of d. fuchs and s. tabachnikov for legendrian knots . [ thmtf ] let @xmath30 be a contact manifold with a cooriented contact structure . let @xmath51 be a component of the space of legendrian curves in @xmath30 , and let @xmath52 be the corresponding component of the space of framed curves in @xmath4 . let @xmath43 be an @xmath6-valued order @xmath17 invariant of legendrian knots in @xmath53 , and let @xmath60 be legendrian knots that represent isotopic framed knots . then @xmath61 , provided that @xmath52 satisfies conditions and . the proof of this theorem is a more or less straightforward generalization of the proof of the corresponding theorem for @xmath62 and @xmath63 proved by d. fuchs and s. tabachnikov @xcite . let @xmath30 be a contact manifold with a parallelized contact structure . let @xmath64 be a component of the space of transverse curves in @xmath30 . since a transverse curve in @xmath30 has a natural framing , we see that @xmath64 is a subspace of a connected component @xmath52 of the space of framed curves in @xmath4 . in @xcite d. fuchs and s. tabachnikov proved that @xmath16-valued finite order invariants of transverse knots in the standard contact @xmath19 can not distinguish transverse knots that belong to the same component @xmath64 of the space of transverse curves in @xmath19 and realize isotopic framed knots . as it was later observed by d. fuchs and s. tabachnikov @xcite , the proof of the theorem can be generalized to a large class of contact manifolds with parallelized contact structures . moreover it can be generalized to the case of finite order invariants with values in an arbitrary abelian group @xmath6 . in order to state the generalized theorem of d. fuchs and s. tabachnikov we introduce the following definition . let @xmath30 be a contact manifold with a parallelized contact structure , let @xmath64 be a component of the space of transverse curves in @xmath53 , and let @xmath52 be the corresponding component of the space of framed curves in @xmath4 . we say that @xmath52 _ satisfies condition _ @xmath54 if for any @xmath55 every framed knot @xmath56 is not isotopic to the framed knot @xmath57 obtained by adding @xmath58 positive extra twists to the framing of @xmath38 . below we formulate the generalized theorem of d. fuchs and s. tabachnikov for transverse knots . [ thmtftransverse ] let @xmath30 be a contact manifold with a parallelized contact structure . let @xmath64 be a component of the space of transverse curves in @xmath30 , and let @xmath52 be the corresponding component of the space of framed curves in @xmath4 . let @xmath43 be an @xmath6-valued order @xmath17 invariant of transverse knots in @xmath53 . let @xmath65 be transverse knots that represent isotopic framed knots . then @xmath61 , provided that @xmath52 satisfies condition . the proof of this theorem is a more or less straightforward generalization of the proof of the corresponding theorem for @xmath62 and @xmath63 proved by d. fuchs and s. tabachnikov @xcite . let @xmath30 be a contact manifold with a cooriented contact structure . let @xmath51 be a connected component of the space of legendrian curves in @xmath4 , and let @xmath52 be the connected component of the space of framed curves that contains @xmath51 . let @xmath66 ( resp . @xmath67 ) be the groups of @xmath6-valued order @xmath17 invariants of legendrian ( resp . framed ) knots from @xmath51 ( resp . from @xmath52 ) . clearly every invariant @xmath68 restricted to the category of legendrian knots in @xmath51 is an element @xmath69 . this gives a homomorphism @xmath70 . [ isomorphism ] let @xmath53 be a contact manifold with a cooriented contact structure . let @xmath51 be a connected component of the space of legendrian curves in @xmath4 , and let @xmath52 be the connected component of the space of framed curves that contains @xmath51 . then the following two statements * a : * and * b : * are equivalent . a : : for any @xmath71 and any knots @xmath72 that represent isotopic framed knots @xmath61 . b : : for @xmath71 there exists @xmath73 such that @xmath74 . such @xmath75 is unique and it defines a canonical isomorphism @xmath76 . for the proof of theorem [ isomorphism ] see subsection [ proofisomorphism ] . the following theorem is an immediate consequence of theorems [ thmtf ] and [ isomorphism ] . [ isomorphismobtained ] let @xmath53 be a contact manifold with a cooriented contact structure , and let @xmath51 be a connected component of the space of legendrian curves in @xmath4 . let @xmath52 be the connected component of the space of framed curves that contains @xmath51 . let @xmath66 ( resp . @xmath67 ) be the groups of @xmath6-valued order @xmath17 invariants of legendrian ( resp . framed ) knots from @xmath51 ( resp . from @xmath52 ) . then the groups @xmath66 and @xmath67 are canonically isomorphic , provided that @xmath52 satisfies conditions and . let @xmath52 be a connected component of the space of framed curves in @xmath4 that satisfies condition . theorem [ isomorphismobtained ] implies for any @xmath77 the group of @xmath6-valued order @xmath17 invariants of legendrian knots from a connected component of legendrian curves curves contained in @xmath52 does not depend on the choice of a cooriented contact structure , provided that for this choice @xmath52 satisfies condition . and hence this group can not be used to distinguish such cooriented contact structures . the @xmath33-principle proved for the legendrian curves by m. gromov @xcite implies that @xmath52 satisfies condition provided that the contact structure is parallelizable . clearly if @xmath52 consists of zero homologous knots , then it satisfies condition and we get that for an arbitrary contact manifold @xmath30 ( with the parallelizable contact structure ) and for an arbitrary component @xmath51 consisting of zero homologous legendrian curves the groups @xmath78 and @xmath67 are canonically isomorphic . since @xmath79 and the standard contact structure on @xmath19 is parallelizable , we get this isomorphism for the components of the space of legendrian and of framed curves in the standard contact @xmath19 . in @xcite d. fuchs and s. tabachnikov showed that for the standard contact @xmath19 and for @xmath62 the quotient groups @xmath80 and @xmath81 are canonically isomorphic . their proof was based on the fact that for the @xmath16-valued vassiliev invariants of framed knots in @xmath19 there exists the universal vassiliev invariant constructed by t. q. t. le and j. murakami @xcite . ( for unframed knots in @xmath19 the existence of the universal vassiliev invariant is the classical result of m. kontsevich @xcite . ) our results generalize the results of fuchs and tabachnikov . [ fibration ] let @xmath4 be an orientable manifold that is a total space of a locally trivial @xmath5-fibration over a ( not necessarily closed or orientable ) surface @xmath28 . assume moreover that @xmath23 and @xmath4 is not the quotient of @xmath1 by the equivalence relation that @xmath82 ( such @xmath4 fibers over @xmath25 ) . then all the components of the space of framed curves in @xmath4 satisfy condition . for the proof of theorem [ fibration ] see subsection [ prooffibration ] . [ corollariesfibration]_corollaries of theorem [ fibration ] . _ let @xmath4 be a manifold satisfying conditions of theorem [ fibration ] , let @xmath12 be a cooriented contact structure on @xmath4 , and let @xmath52 be a connected component of the space of framed curves in @xmath30 that satisfies condition . then for every connected component @xmath83 of the space of legendrian curves there is a natural isomorphism between the groups @xmath66 and @xmath67 of @xmath6-valued order @xmath17 invariants . thus if the contact structure on a manifold ( satisfying conditions of theorem [ fibration ] ) is parallelizable , then we get the isomorphisms of the groups of @xmath6-valued order @xmath17 invariants of legendrian and of framed knots ( from the corresponding components of the two spaces ) . a very interesting class of contact manifolds satisfying the conditions of theorem [ fibration ] is formed by the spherical cotangent bundles @xmath29 of surfaces @xmath28 with the natural contact structure on @xmath29 ( see [ definitions ] ) . the theory of the invariants of legendrian knots in @xmath29 is often referred to as the theory of v. i. arnold s @xcite @xmath84-type invariants of fronts on a surface @xmath28 . the natural contact structure @xmath12 on the spherical cotangent bundle @xmath29 of a surface @xmath28 is cooriented . ( the coorientation is induced from the coorientation of the contact elements of @xmath28 . ) one can verify that for orientable @xmath28 the standard contact structure on @xmath29 is parallelizable , and hence all the components of the space of framed curves satisfy condition . if @xmath28 is not orientable , then the standard cooriented contact structure on @xmath29 is not parallelizable , but one can still verify ( cf . proposition 8.2.4 @xcite ) that every connected component of the space of framed curves satisfies condition . hence for an arbitrary surface @xmath28 we obtain the canonical isomorphism of the groups of @xmath6-valued order @xmath17 invariants of legendrian and of framed knots ( from the corresponding components of the spaces of legendrian and of framed knots in @xmath29 with the standard contact structure ) . or equivalently we get that the groups of @xmath6-valued order @xmath17 @xmath84-type invariants of fronts on @xmath28 and the groups of @xmath6-valued order @xmath17 invariants of framed knots in @xmath29 ( from the corresponding components of the two spaces ) are canonically isomorphic . in @xcite j. w. hill showed that for the standard contact @xmath85 and for @xmath62 the quotient groups @xmath80 and @xmath81 are canonically isomorphic . his proof was based on the fact that for the @xmath16-valued vassiliev invariants of framed knots in @xmath85 there exists the universal vassiliev invariant constructed by v. goryunov @xcite . ( for unframed knots in @xmath19 the existence of the universal vassiliev invariant is the classical result of m. kontsevich @xcite . ) our results generalize the results of j. w. hill ( even in the case of @xmath86 . before we proceed further we clarify what it means that a connected component @xmath52 of the space of framed curves in a contact manifold @xmath53 with a cooriented contact structure satisfies condition . [ interpretationconditionii ] let @xmath30 be a contact manifold with a cooriented contact structure , and let @xmath52 be a component of the space of framed curves in @xmath4 . then @xmath52 does not satisfy condition if and only if there exists @xmath87 such that the value of the euler class of the contact bundle on @xmath9 is nonzero and @xmath9 is realizable by a mapping @xmath88 of the two torus with the property that @xmath89 of the meridian of @xmath90 is a loop free homotopic to loops realized by curves from @xmath52 . for the proof of proposition [ interpretationconditionii ] see subsection [ proofinterpretationconditionii ] . [ atoroidal ] let @xmath30 be a contact manifold ( with a cooriented contact structure ) such that @xmath91 , every @xmath92 is of infinite order , and for an arbitrary mapping @xmath88 of the two - torus the homomorphism @xmath93 is not injective . then all the components of the space of framed curves in @xmath4 satisfy conditions and , and hence the groups of @xmath46-valued order @xmath17 invariants of legendrian and of framed knots ( from the corresponding components ) are canonically isomorphic . for the proof of theorem [ atoroidal ] see subsection [ proofatoridal ] . a well known theorem by a. preissman ( see @xcite pp . 258 - 265 ) says that every nontrivial commutative subgroup of the fundamental group of a closed three - dimensional manifold @xmath4 of negative sectional curvature is infinite cyclic . it is also known that the universal covering of such @xmath4 is diffeomorphic to @xmath19 , and hence @xmath91 . thus every closed manifold @xmath4 admitting a metric of a negative sectional curvature satisfies all the conditions of theorem [ atoroidal ] and for an arbitrary cooriented contact structure on such @xmath4 we obtain the isomorphism of the groups of @xmath6-valued order @xmath17 invariants of legendrian and of framed knots from the corresponding components . theorem [ atoroidal ] implies ( see proposition [ preissman ] ) that for @xmath53 a contact manifold with a cooriented contact structure such that @xmath4 admits a structure of an @xmath21-fibration over @xmath94 ( the klein bottle ) we also get the isomorphism of the groups of the @xmath6-valued order @xmath17 invariants of legendrian and of framed knots from the corresponding components of the two spaces . ( from the work of t. fiedler @xcite one can easily see that if @xmath4 is an orientable manifold @xmath21-fibered over a ( not necessarily orientable ) surface @xmath28 , then all the components of the space of framed curves satisfy condition . hence for an arbitrary parallelizable contact structure on such @xmath4 the groups of @xmath6-valued order @xmath17 invariants of legendrian and of framed knots ( from the corresponding components ) are canonically isomorphic . ) in this subsection we construct a big class of contact manifolds @xmath53 and connected components @xmath51 of legendrian curves in @xmath53 for which there exist legendrian knots @xmath95 such that @xmath2 and @xmath3 realize isotopic framed knots but can be distinguished by finite order invariants of legendrian knots . the theorem of r. lutz @xcite says that for an arbitrary orientable three - manifold @xmath4 every homotopy class of the distribution of tangent two - planes to @xmath4 contains a contact structure . ( the theorem of ya . eliashberg @xcite says even more that every homotopy class of the distribution of tangent two - planes to @xmath4 contains a positive overtwisted contact structure . ) however in our constructions we will use only the euler classes of the contact bundles . for this reason we start with the following proposition . [ existcontact ] 1 : : let @xmath53 be a contact manifold with a cooriented contact structure , and let @xmath96 be the euler class of the contact bundle . then @xmath97 , for some @xmath98 . 2 : : let @xmath4 be an oriented manifold , and @xmath96 be such that @xmath99 for some @xmath100 . then there exists a cooriented contact structure @xmath12 on @xmath4 such that the the euler class of the contact bundle is @xmath101 . for the proof of proposition [ existcontact ] see subsubsection [ proofexistcontact ] . let @xmath12 be a cooriented contact structure on @xmath102 such that the euler class of the contact bundle is nonzero . let @xmath38 be a knot in @xmath103 that crosses exactly once one of the spheres @xmath104 . the theorem of chow @xcite and rashevskii @xcite says that there exists a legendrian knot @xmath105 that is @xmath106-small isotopic to @xmath38 as an unframed knot . let @xmath2 be a legendrian knot that is the same as @xmath2 except of a small piece located in a chart contactomorphic to the standard contact @xmath19 where it is changed as it is shown in figure [ change.fig ] ( see [ description ] ) . [ example1 ] a : : legendrian knots @xmath105 and @xmath2 belong to the same component of the space of legendrian curves and realize isotopic framed knots . b : : there exists a @xmath107-valued order one invariant @xmath108 of legendrian knots , such that @xmath109 . for the proof of theorem [ example1 ] see subsection [ proofexample1 ] . [ overtwisted]the contact structure and the knots @xmath105 and @xmath2 in theorem [ example1 ] can be chosen so that the complements of @xmath105 and of @xmath2 are overtwisted . let @xmath110 be an embedded into @xmath4 disk centered at a point @xmath111 . the theorem of eliashberg @xcite says that every homotopy class of the distribution of two - planes contains an overtwisted contact structure that has @xmath110 as the standard overtwisted disk . in the example of theorem [ example1 ] we can start with an unframed knot @xmath38 that is far away from @xmath110 . then since both @xmath105 and @xmath2 were constructed using a @xmath106-small approximation of @xmath38 we can assume that they are also far away from @xmath110 and thus the restrictions of the contact structure to the complements of @xmath105 and of @xmath2 are overtwisted . let @xmath112 , @xmath113 , be the knot that is the same as @xmath105 everywhere except of a small piece located in a chart contactomorphic to the standard contact @xmath19 where it is changed in the way described by the addition of @xmath114 zigzags shown in figure [ change.fig ] . the proof of theorem [ example1 ] implies that all @xmath112 s are legendrian homotopic knots that realize isotopic framed knots , but for all @xmath115 legendrian knots @xmath116 and @xmath117 are not isotopic and there exists a @xmath107-valued order one invariant @xmath108 of legendrian knots such that @xmath118 . ( and hence this @xmath108 distinguishes all the @xmath112 s . ) [ otherexamples ] _ below we describe another big family of examples when finite order invariants distinguish legendrian knots that realize isotopic framed knots . _ let @xmath28 be a nonorientable surface that can be decomposed as a sum of the klein bottle @xmath38 and a surface @xmath119 . let @xmath4 be an orientable manifold that admits a structure of a locally trivial @xmath5-fibration @xmath120 . ( for example one can take @xmath4 to be @xmath121 the spherical tangent bundle of @xmath28 . ) consider an @xmath5-fibration @xmath122 induced from @xmath123 by the mapping of @xmath5 to the solid loop in the surface shown in figure [ example2.fig ] . ( in this figure the enumeration of the end points of the arcs indicates which pairs of points should be identified to obtain the loop . ) since the solid loop is an orientation preserving loop in @xmath28 we get that @xmath124 . consider the natural map @xmath125 ( the one from the definition of the induced fibration ) . since a homology class in @xmath126 projecting to the dashed loop in figure [ example2.fig ] has a nonzero intersection with the class realized by @xmath127 , we get that the class realized by @xmath127 is nonzero in @xmath128 . let @xmath12 be a cooriented contact structure on @xmath4 such that the value of the euler class @xmath96 of the contact bundle on the homology class realized by @xmath127 is nonzero . and hence is equal to @xmath129 for some @xmath130 , see proposition [ existcontact ] . let @xmath38 be an arbitrary legendrian knot whose projection to @xmath28 ( considered as a loop ) is free homotopic to the solid loop in figure [ example2.fig ] . let @xmath131 be legendrian knots that are the same as @xmath38 everywhere except of a chart ( contactomorphic to the standard contact @xmath19 ) where @xmath2 and @xmath3 are different from @xmath38 as it is described in figure [ example4.fig ] , see [ description ] . ( here @xmath129 is the value of the euler class of the contact bundle on the homology class realized by @xmath127 . ) [ example2 ] the knots @xmath2 and @xmath3 described above belong to the same component @xmath51 of the space of legendrian curves and realize isotopic framed knots . there exists a @xmath107-valued order one invariant @xmath108 of legendrian knots from @xmath51 such that @xmath132 . for the proof of theorem [ example2 ] see subsection [ proofexample2 ] . in theorem [ example2 ] we do not need the cooriented contact structure @xmath12 to be tight or overtwisted . the only condition @xmath12 should satisfy is the condition on the euler class of the contact bundle . however the author is not currently aware of the existence of tight contact satisfying the conditions of theorem [ example2 ] . similar to [ overtwisted ] one can verify that the contact structure and the knots @xmath2 and @xmath3 in the statement of theorem [ example2 ] can be chosen so that the restrictions of the contact structure to the complements of @xmath2 and of @xmath3 are overtwisted . using the ideas of the proof of theorem [ example2 ] one can easily construct many much more complicated examples of contact manifolds @xmath30 with a cooriented contact structure and connected components @xmath51 of the space of legendrian curves in @xmath53 such that there exist legendrian knots @xmath133 that realize isotopic framed knots but are distinguishable by finite order invariants of legendrian knots . for example instead of the solid loop we could take any loop such that the number of double points that separate the loop into two orientation reversing loops is odd , and the value of the euler class of the contact bundle on the homology class realized by the mapping @xmath134 corresponding to the loop is nonzero . let @xmath4 be a contact manifold with a parallelized contact structure . let @xmath64 be a connected component of the space of transverse curves in @xmath4 , and let @xmath52 be the connected component of the space of framed curves that contains @xmath64 . ( such a component exists because a transverse curve in a manifold with a parallelized contact structure is naturally framed , and the path in the space of transverse curves corresponds to a path in the space of framed curves . ) let @xmath135 ( resp . @xmath67 ) be the groups of @xmath6-valued order @xmath17 invariants of transverse ( resp . framed ) knots from @xmath64 ( resp . from @xmath52 ) . clearly every invariant @xmath68 restricted to the category of transverse knots in @xmath64 is an element @xmath136 . this gives a homomorphism @xmath137 . [ isomorphismtransverse ] let @xmath53 be a contact manifold with a parallelized contact structure . let @xmath64 be a connected component of the space of transverse curves in @xmath30 , and let @xmath52 be the corresponding component of the space of framed curves . then the following two statements * a : * and * b : * are equivalent . a : : for any @xmath138 and any knots @xmath139 that represent isotopic framed knots @xmath61 . b : : for @xmath138 there exists @xmath73 such that @xmath74 . such @xmath75 is unique and it defines a canonical isomorphism @xmath140 . the proof of theorem [ isomorphismtransverse ] is analogous to the proof of the theorem [ isomorphism ] . the following theorem is an immediate consequence of theorems [ thmtftransverse ] and [ isomorphismtransverse ] . [ isomorphismobtainedtransverse ] let @xmath53 be a contact manifold with a parallelized contact structure , and let @xmath64 be a connected component of the space of transverse curves in @xmath4 . let @xmath52 be the connected component of the space of framed curves that contains @xmath64 . let @xmath135 ( resp . @xmath67 ) be the groups of @xmath6-valued order @xmath17 invariants of transverse ( resp . framed ) knots from @xmath64 ( resp . from @xmath52 ) . then the groups @xmath135 and @xmath67 are canonically isomorphic , provided that @xmath52 satisfies condition . let @xmath30 be a contact manifold with a parallelized contact structure , and let @xmath52 be a connected component of the space of framed curves in @xmath4 that satisfies condition . theorem [ isomorphismobtainedtransverse ] implies that for any @xmath77 the group of @xmath6-valued order @xmath17 invariants of transverse knots from a connected component of transverse curves contained in @xmath52 does not depend on the choice of a parallelized contact structure . and hence this group can not be used to distinguish parallelized contact structures on @xmath4 . clearly if @xmath52 consists of zero homologous knots , then it satisfies condition and we get that for an arbitrary contact manifold @xmath30 ( with a parallelized contact structure ) and for an arbitrary component @xmath141 consisting of zero homologous transverse curves the groups @xmath142 and @xmath67 are canonically isomorphic . let @xmath64 be a connected component of the space of transverse curves in @xmath19 , and let @xmath52 be the corresponding component of the space of framed curves in @xmath19 . since @xmath79 and the standard contact structure in @xmath19 is parallelized we get that the groups @xmath135 and @xmath67 are canonically isomorphic . this generalizes the result of d. fuchs and s.tabachnikov @xcite that says that for the standard contact @xmath19 and for @xmath62 the quotient groups @xmath143 and @xmath81 are canonically isomorphic . ( there are two components of the space of transverse curves in @xmath19 . they consist of respectively positive and negative transverse curves . the connected components of the space of framed curves in @xmath19 are enumerated by the @xmath144-reduced value of the self - linking number of knots from the component . ) for a big class of manifolds one can show that all the connected components of the space of framed curves satisfy condition . ( see theorems [ fibration ] and [ atoroidal ] and the remarks and corollaries after them . ) for an arbitrary parallelized contact structure on these manifolds one gets that the groups of @xmath6-valued order @xmath17 invariants of transverse and of framed knots ( from the corresponding components ) are canonically isomorphic . [ commute ] let @xmath145 be a locally trivial @xmath5-fibration of an oriented manifold @xmath146 over a ( not necessarily orientable ) manifold @xmath147 . let @xmath148 be the class of an oriented fiber of @xmath123 , and let @xmath9 be an element of @xmath149 . then : _ proof of proposition [ commute ] . _ if we move an oriented fiber along the loop @xmath153 , then in the end it comes to itself either with the same or with the opposite orientation . it is easy to see that it comes to itself with the opposite orientation if and only if @xmath151 is an orientation reversing loop in @xmath147 . _ proof of proposition [ preissman ] . _ it is well known that any closed @xmath28 , other than @xmath157 , admits a hyperbolic metric of a constant negative curvature . ( it is induced from the universal covering of @xmath28 by the hyperbolic plane @xmath158 . ) the theorem by a. preissman ( see @xcite pp . 258 - 265 ) says that if @xmath4 is a compact riemannian manifold with a negative curvature , then any nontrivial abelian subgroup @xmath159 is isomorphic to @xmath107 . thus if @xmath94 is closed , then any nontrivial commutative @xmath160 is infinite cyclic . [ toughandtechnical ] let @xmath161 ( klein bottle ) be a surface not necessarily closed or orientable . let @xmath4 be an orientable manifold , and let @xmath120 be a locally trivial @xmath5-fibration . let @xmath162 be the class of an oriented @xmath5-fiber of @xmath123 , and let @xmath163 be an element with @xmath164 . let @xmath165 be an element from @xmath166 the centralizer of @xmath9 . then : a : : if @xmath167 is an orientation preserving loop in @xmath28 , then there exist @xmath168 and @xmath169 such that @xmath170 . b : : if @xmath167 is an orientation reversing loop in @xmath28 , then there exist @xmath171 and @xmath169 such that @xmath172 . [ prooftoughandtechnical ] _ proof of proposition [ toughandtechnical ] . _ since @xmath173 , we have that @xmath9 and @xmath165 commute in @xmath174 . hence @xmath167 and @xmath175 commute in @xmath156 . proposition [ preissman ] and the fact that @xmath164 imply that there exist @xmath176 with @xmath177 , @xmath171 , and @xmath169 such that @xmath178 and @xmath179 . to prove statement * b : * we have to show that @xmath182 in , provided that @xmath167 is an orientation reversing loop in @xmath28 . since @xmath173 we have that @xmath183 commutes with @xmath9 . clearly @xmath184 is an orientation preserving loop in @xmath28 , and proposition [ commute ] implies that @xmath9 commutes with @xmath185 if and only if @xmath182 . ( note that @xmath186 has infinite order in @xmath174 for @xmath4 from the statement of the proposition . ) [ pi2nontrivial ] let @xmath4 with @xmath187 be a connected three dimensional orientable manifold that admits a structure of a locally - trivial @xmath5-fibration over a surface @xmath28 ( not necessarily compact or orientable ) . then @xmath4 is either @xmath1 or the quotient of @xmath1 by the equivalence relation @xmath188 . ( in the last case @xmath4 fibers over @xmath25 . ) let @xmath4 be a manifold with @xmath187 that fibers over @xmath189 . choose an orientation on @xmath4 . since both @xmath4 and @xmath189 are oriented we get a canonical orientation on all the @xmath5-fibers of @xmath123 . from the exact homotopy sequence of the fibration one gets that for such @xmath4 the class @xmath162 of the fiber of @xmath123 has infinite order in @xmath174 . on the other hand if @xmath190 is the euler class of the oriented plane bundle associated to @xmath123 , then @xmath191 . since @xmath174 is generated by @xmath186 , we get that @xmath192 , and hence @xmath193 . but the only plane bundle over @xmath189 with the zero euler class is @xmath194 and it corresponds to @xmath195 . assume now that @xmath4 fibers over @xmath25 . let @xmath196 be the locally trivial @xmath5-fibration induced from @xmath197 under the orientation double covering @xmath198 . clearly @xmath199 is orientable and @xmath199 double covers @xmath4 . since @xmath200 we have that @xmath201 , and hence @xmath202 . using this one verifies that @xmath4 is the quotient of @xmath1 by the equivalence relation @xmath203 . [ hansen ] let @xmath146 be a manifold , let @xmath204 be the space of free loops in @xmath146 , and let @xmath205 be a loop . a loop @xmath206 is a mapping @xmath207 , with @xmath208 and @xmath209 being the trace of the point @xmath210 under the homotopy of @xmath211 described by @xmath9 . let @xmath212 be the element corresponding to the trace of the point @xmath210 under the homotopy of @xmath211 described by @xmath9 . since @xmath213 is commutative , we get that @xmath214 is a surjective homomorphism of @xmath215 onto the centralizer @xmath216 of @xmath217 . if @xmath218 for the loops @xmath219 , then the mappings @xmath220 and @xmath221 of @xmath90 corresponding to these loops can be deformed to be identical on the @xmath8-skeleton of @xmath90 . clearly the obstruction for @xmath220 and @xmath221 to be homotopic as mappings of @xmath90 ( with the mapping of the @xmath8-skeleton of @xmath90 fixed under homotopy ) is an element of @xmath222 obtained by gluing together the @xmath13-cells of the two tori . in particular we get the proposition of v. l. hansen @xcite ) that @xmath223 is an isomorphism , provided that @xmath224 . [ h - principleforcurves ] _ @xmath33-principle for curves in @xmath4 . _ for a three dimensional manifold @xmath4 we put @xmath225 to be the manifold obtained by the fiberwise spherization of the tangent bundle of @xmath4 , and we put @xmath226 to be the corresponding locally trivial @xmath189-fibration . the @xmath33-principle says that the space of curves in @xmath4 is weak homotopy equivalent to @xmath227 the space of free loops in @xmath225 . the weak homotopy equivalence is given by mapping a curve @xmath38 to a loop @xmath228 that sends a point @xmath229 to the point of @xmath225 corresponding to the direction of the velocity vector of @xmath38 at @xmath230 . [ proofexistcontact]_proof of the proposition [ existcontact ] . _ since the contact structure is cooriented we get that the tangent bundle @xmath231 is isomorphic to a sum @xmath232 of an oriented contact bundle @xmath12 and a trivial oriented line bundle @xmath233 . but the tangent bundle of every orientable @xmath22-manifold is trivializable and we get that the second stiefel - whitney class of the contact bundle is zero . but the second stiefel - whitney of @xmath12 is the projection of the euler class of @xmath12 under the natural mapping @xmath234 and we get the first statement of the proposition . consider an oriented @xmath13-dimensional vector bundle @xmath235 over @xmath4 with the euler class @xmath236 . similar to the above we get that the second stiefel - whitney class @xmath237 of @xmath235 is zero , since it is the projection of @xmath238 . since @xmath235 is an oriented vector bundle we have @xmath239 . consider a sum @xmath240 of @xmath235 with the trivial oriented one - dimensional vector bundle @xmath233 . clearly the total stiefel - whitney class of the three dimensional oriented vector bundle @xmath240 is equal to @xmath8 , and the euler class of @xmath240 is equal to @xmath241 . using the interpretation of the stiefel - whitney and the euler classes of @xmath240 as obstructions for the trivialization of @xmath240 , we get that @xmath240 is trivializable . since the tangent bundle of an oriented three dimensional manifold is trivializable , we see that @xmath235 is isomorphic to an oriented sub - bundle of @xmath231 . since @xmath4 is oriented this sub - bundle of @xmath231 is also cooriented . now the theorem of lutz @xcite , that says that every homotopy class of a distribution of tangent two - planes to @xmath4 contains a contact structure , implies the second statement of the proposition . [ obstruction ] let @xmath2 and @xmath3 be two framed knots that coincide pointwise as embeddings of @xmath5 . then there is an integer obstruction @xmath243 for them to be isotopic as framed knots with the embeddings of @xmath5 fixed under the isotopy . this obstruction is calculated as follows . let @xmath244 be the knot obtained by shifting @xmath2 along the framing and reversing the orientation on the shifted copy . together @xmath2 and @xmath244 bound a thin strip . we put @xmath245 to be the intersection number of the strip with the very small shift of @xmath3 along its framing . for two singular framed knots @xmath250 and @xmath251 with @xmath36 transverse double points that coincide pointwise as immersions of @xmath5 , we put @xmath252 to be the value of @xmath253 on the nonsingular framed knots @xmath2 and @xmath3 that coincide pointwise as embeddings of @xmath5 and are obtained from @xmath250 and @xmath251 by resolving each pair of the corresponding double points of @xmath250 and of @xmath251 in the same way . ( the value @xmath254 does not depend on the resolution as soon as the corresponding double points of the two knots are resolved in exactly the same way . ) as before @xmath254 is the integer valued obstruction for @xmath250 and @xmath251 to be isotopic as singular framed knots with @xmath36 double points with the immersion of @xmath5 corresponding to the two knots fixed under isotopy . for a singular framed knot @xmath255 with @xmath36 double points we denote by @xmath256 the isotopy class of a singular framed knot with @xmath36 double points that coincides with @xmath255 as an immersion of @xmath5 and has @xmath257 . [ homotopy ] let @xmath2 and @xmath3 be framed knots ( resp . singular framed knots with @xmath36 double points ) that coincide pointwise as embeddings ( resp . immersions ) of @xmath5 . then @xmath2 and @xmath3 are homotopic as framed knots ( resp . singular framed knots with @xmath36 double points ) if and only if @xmath245 is even . [ proofhomotopy]_proof of proposition [ homotopy ] . _ clearly if @xmath245 is even , then @xmath2 and @xmath3 are framed homotopic . ( we can change the obstruction by two by creating a small kink and passing through a double point at it . ) every oriented three - dimensional manifold @xmath4 is parallelizable . hence all its stiefel - whitney classes vanish and @xmath4 admits a @xmath258-structure . a framed curve @xmath38 in @xmath4 represents a loop in the principal @xmath259-bundle of @xmath231 . ( the three - frame corresponding to a point of @xmath38 is the velocity vector , the framing vector , and the unique third vector of unit length such that the three - frame defines a positive orientation of @xmath4 . ) one observes that the values of the @xmath258-structure on the loops in the principal @xmath259-bundle of @xmath231 realized by @xmath2 and @xmath3 are different provided that @xmath245 is odd . but these values do not change under homotopy of framed curves . hence if @xmath245 is odd , then @xmath2 and @xmath3 are not framed homotopic . [ strange ] using the self - linking invariant of framed knots one can easily show that if @xmath2 and @xmath3 in [ obstruction ] are pointwise coinciding zero - homologous framed knots and @xmath260 , then @xmath2 is not isotopic to @xmath3 in the category of framed knots . however in general this is not true , see [ proofexample1 ] . this forces us to introduce the following definition . if for an unframed knot @xmath38 there exist isotopic framed knots @xmath2 and @xmath3 that coincide with @xmath38 pointwise and have @xmath260 , then we say that @xmath38 _ admits finitely many framings_. for @xmath38 that admits finitely many framings we put @xmath261 _ the number of framings of _ @xmath38 to be the minimal positive integer @xmath253 such that there exist isotopic framed knots @xmath2 and @xmath3 that coincide with @xmath38 pointwise and have @xmath262 . one can easily show that if @xmath38 admits finitely many framings , then there are exactly @xmath261 isotopy classes of framed knots realizing the isotopy class of the unframed knot @xmath38 . proposition [ homotopy ] implies that @xmath261 is even . a : : let @xmath38 be an unframed knot obtained by forgetting a framing on a knot from @xmath52 . then there exists a legendrian knot from @xmath263 realizing the isotopy class of @xmath38 . b : : if @xmath246 is an isotopy class of framed knots in @xmath52 that is realizable by a legendrian knot from @xmath51 , then the isotopy class of @xmath264 is also realizable by a legendrian knot from @xmath51 . c : : let @xmath44 be an unframed knot singular knot with @xmath36 double points obtained by forgetting a framing on a singular knot from @xmath52 . then there exists a legendrian knot from @xmath265 realizing the isotopy class of @xmath44 . d : : if @xmath266 is an isotopy class of framed knots in @xmath52 that is realizable by a singular legendrian knot from @xmath51 , then the isotopy class of @xmath267 is also realizable by a singular legendrian knot from @xmath51 . [ proofdecrease]_proof of proposition [ decrease ] . _ _ proof of statement [ a ] of proposition [ decrease ] . _ let @xmath31 be the fiberwise spherization of the two - dimensional contact vector bundle , and let @xmath32 be the corresponding locally trivial @xmath5-fibration . we denote by @xmath268 the class of an oriented @xmath5-fiber of @xmath269 . for a legendrian curve @xmath270 denote by @xmath271 the loop in @xmath31 obtained by mapping a point @xmath229 to the point of @xmath31 corresponding to the direction of the velocity vector of @xmath35 at @xmath272 . the @xmath33-principle proved for the legendrian curve by m. gromov @xcite says that that @xmath2 and @xmath3 legendrian curves in @xmath4 belong to the same component of the space of legendrian curves in @xmath4 if and only if @xmath273 and @xmath274 are free homotopic loops in @xmath31 . w. l. chow @xcite and p. k. rashevskii @xcite showed that every knot @xmath38 is isotopic to a legendrian knot @xmath35 ( and this isotopy can be made @xmath106-small ) . deforming @xmath38 we can assume ( see [ h - principleforcurves ] ) that : * a : * @xmath38 and @xmath35 coincide in the neighborhood of @xmath275 , * b : * @xmath38 and @xmath35 realize the same element @xmath276\in \pi_1(m , k_l(1))$ ] , and * c : * that liftings to @xmath31 of legendrian curves from @xmath51 are free homotopic to a loop @xmath9 in @xmath31 such that @xmath277 and @xmath278 \in \pi_1(m , k_l(1))$ ] . take a chart of @xmath4 ( that is contactomorphic to the standard contact @xmath19 ) containing a piece of the legendrian knot . from @xcite it is easy to see that insertions of two cusps shown in figure [ twocusp.fig ] into the projection of a legendrian knot in @xmath19 onto the @xmath281-plane induce multiplication by @xmath282 of the class in @xmath283 of a lifting of a knot @xmath35 to a loop in @xmath31 . ( here the sign depends on the choice of an orientation of the fiber used to define @xmath186 . ) performing this operation sufficiently many times we obtain the legendrian knot from @xmath51 realizing the unframed knot @xmath38 . _ proof of statement [ b ] of proposition [ decrease ] . _ take a local chart of @xmath4 ( where @xmath4 is contactomorphic to the standard contact @xmath19 ) containing a piece of the knot @xmath246 and perform the homotopy in @xmath51 shown in figure [ kink.fig ] . ( observe that a selftangency point of a projection of a legendrian curve in @xmath19 to the @xmath47-plane corresponds to a double point of the legendrian curve . ) a straightforward verification ( see @xcite ) shows that the legendrian knot we obtain in the end of the homotopy realizes @xmath264 . _ definition of @xmath284 . _ if an isotopy class of the knot @xmath285 is realizable by a legendrian knot @xmath286 , then put @xmath287 . the value @xmath288 is well defined because if @xmath289 is another knot realizing @xmath38 , then @xmath290 by statement * a : * of theorem [ isomorphism ] . let @xmath291 be the component of the space of unframed curves that corresponds to forgetting framings on the curves from @xmath52 . propositions [ decrease ] and [ homotopy ] imply that if an unframed knot @xmath292 admits finitely many framings , then all the isotopy classes of framed knots from @xmath52 realizing the unframed knot @xmath293 are realizable by legendrian knots from @xmath51 . thus we have defined the value of @xmath284 on all the framed knots from @xmath52 that realize unframed knots admitting finitely many framings . if @xmath292 admits infinitely many framings , then either * a ) * all the framed knots from @xmath52 realizing the isotopy class of @xmath293 are realizable by legendrian knots from @xmath51 or * b ) * there exists a knot @xmath294 realizing the isotopy class of @xmath293 such that @xmath246 is realizable by a legendrian knot from @xmath51 and @xmath295 is not realizable by a legendrian knot from @xmath51 ( in this case @xmath296 etc . also are not realizable by legendrian knots from @xmath51 , see [ decrease ] ) . in the case * a ) * the value of @xmath284 is already defined for all the framed knots from @xmath52 realizing @xmath293 . in the case * b ) * put @xmath297 ( proposition [ decrease ] implies that the sum on the right hand side is well - defined . ) similarly put @xmath298 now we have defined @xmath284 on all the framed knots ( from @xmath52 ) realizing @xmath293 . doing this for all @xmath293 for which case * b ) * holds we define the value of @xmath284 on all the knots from @xmath52 . if @xmath295 is realizable by a legendrian knot @xmath35 , then consider a singular legendrian knot @xmath300 with @xmath39 double points that are vertices of @xmath39 small kinks such that we get @xmath35 if we resolve all of them positively staying in the class of the legendrian knots . let @xmath301 be the set of the @xmath45 ways in which we can resolve @xmath300 . for @xmath302 put @xmath303 to be the sign of the resolution , and put @xmath304 to be the nonsingular legendrian knot obtained via the resolution @xmath305 . since @xmath43 is an order @xmath17 invariant of legendrian knots we get that @xmath306 this finishes the proof of the proposition . let @xmath307 be a singular framed knot with @xmath39 double points . let @xmath301 be the set of @xmath45 ways of resolving the double points . for @xmath308 put @xmath303 to be the sign of the resolution , and put @xmath309 to be the isotopy class of the knot obtained via the resolution @xmath305 . first we observe that the fact whether the identity holds or not depends only on the isotopy class of the singular knot @xmath44 with @xmath39 double points . if the isotopy class of @xmath44 is realizable by a singular legendrian knot with @xmath39 double points , then identity holds for @xmath44 , since @xmath43 is an order @xmath17 invariant of legendrian knots and the value of @xmath284 on a framed knot @xmath38 realizable by a legendrian knot @xmath35 was put to be @xmath312 . if @xmath313 admits finitely many framings , then all the singular framed knots from @xmath52 realizing the isotopy class of @xmath313 are realizable by singular legendrian knots from @xmath51 and we get that identity holds for @xmath44 . if @xmath313 admits infinitely many framings and all the singular framed knots from @xmath52 realizing @xmath313 are realizable by singular legendrian knots from @xmath51 , then automatically holds for @xmath44 . if @xmath313 admits infinitely many framings but not all the isotopy classes of singular framed knots from @xmath52 realizing @xmath313 are realizable by singular legendrian knots from @xmath51 . then put @xmath314 to be the framed knot realizing @xmath313 that is realizable by a singular legendrian knot from @xmath51 and such that @xmath315 , @xmath316 , are not realizable by singular legendrian knots from @xmath51 . proposition [ decrease ] says that @xmath317 , @xmath316 , are realizable by singular legendrian knots from @xmath51 and hence identity holds for @xmath317 , @xmath318 . clearly @xmath322 is an injective homomorphism . since an invariant @xmath68 should satisfy identity we get that @xmath323 is also injective . finally we observe that @xmath324 and this implies that @xmath76 is an isomorphism . to see that @xmath322 establishing the isomorphism is unique consider a singular framed knot with @xmath39 double points at the vertexes of @xmath39 small singular kinks . since we want @xmath284 to be an order @xmath17 invariant , the sum of the values of it ( with appropriate signs ) on the knots obtained by the @xmath45 possible resolutions of the double points should be zero . this forces us to use the formula to define the value of @xmath284 on the framed knots that are not realizable by legendrian knots . assume that @xmath52 does not satisfy condition , then there exists a framed knot @xmath325 such that @xmath326 is isotopic to @xmath327 for some @xmath328 . ( see [ obstruction ] for the definition of @xmath327 . ) let @xmath291 be the connected component of the space of unframed curves obtained by forgetting the framings on curves from @xmath52 . to a framed knot @xmath326 one can associate a link @xmath329 of two unframed knots with the first knot @xmath2 being @xmath326 with forgotten framing and the second knot @xmath3 being obtained by a small shift of @xmath2 along the framing of @xmath326 . let @xmath9 be a generic closed path in @xmath330 starting with @xmath331 . put @xmath332 to be set of instances when the link becomes singular under the deformation @xmath9 . at these instances @xmath9 crosses the discriminant in @xmath330 . ( the discriminant is a subspace of @xmath330 formed by singular links . ) put @xmath333 , @xmath334 , to be the signs of these crossings . put @xmath335 to be those instances when the arising double point of the singular link is an intersection of the two different components of the link . for a generic closed path @xmath9 starting with @xmath331 put @xmath336 . consider the instances when the first knot of the link becomes singular under @xmath9 . the double point of @xmath2 corresponding to such an instance separates @xmath2 into two loops . let @xmath337 be the set of instances when both loops corresponding to a double point of @xmath2 project to orientation reversing loops in @xmath28 . ( if @xmath28 is orientable or if @xmath291 consists of loops whose projections to @xmath28 are orientation reversing loops , then @xmath338 . ) similarly to the above put @xmath339 . for @xmath340 put @xmath341 to be the trace in @xmath4 of the point of the knot @xmath2 corresponding to the point @xmath8 ( on the unit complex circle that parameterizes @xmath2 ) under the deformation described by @xmath165 . it is easy to see that @xmath342 is a homomorphism . let @xmath343 be the analogous homomorphism constructed using the trace in @xmath4 of the point on @xmath3 corresponding to the point @xmath8 under the deformation of the link . clearly if @xmath326 is isotopic to @xmath327 , then there exists a generic closed path @xmath9 in @xmath330 starting at @xmath331 such that @xmath346 and the restriction of @xmath9 to each of the two components of the link is an isotopy . moreover this @xmath9 should satisfy the condition that @xmath347 . [ mainidea ] the codimension two stratum of the discriminant consists of links with two distinct transverse double points . it is easy to see that @xmath350 and @xmath351 , for any small loop @xmath165 going around the codimension two stratum . this implies that @xmath352 and @xmath353 depend only on the element of @xmath354 realized by a generic loop @xmath9 . hence @xmath355 and @xmath356 are homomorphisms from @xmath354 to @xmath107 . clearly @xmath357 and @xmath358 , for every @xmath359 and @xmath360 . moreover if the class @xmath9 satisfies conditions * a : * and * b : * of [ idea ] , then @xmath361 also satisfies these conditions . hence to prove the theorem it suffices to show that for every @xmath362 such that @xmath363 there exists @xmath364 such that at least of the following holds : 1 : : @xmath365 , 2 : : @xmath366 and hence @xmath9 is not realizable by a deformation of @xmath331 the restriction of which to every component of the link is an isotopy . ( the upper index @xmath367 in @xmath368 stands for obstruction . ) [ obstructiondecrease ] the @xmath33-principle ( see [ h - principleforcurves ] ) says that @xmath291 is weak homotopy equivalent to @xmath227 , the space of free loops in @xmath225 the spherical tangent bundle of @xmath4 . in particular @xmath369 . in subsubsection [ hansen ] we introduced a surjective homomorphism @xmath37 from @xmath370 onto @xmath371 . let @xmath372 be loops such that @xmath373 . as it was explained in [ hansen ] the obstruction for @xmath9 and @xmath165 to be homotopic is an element of @xmath374 . since every closed orientable manifold is parallelizable we get that @xmath375 . proposition [ pi2nontrivial ] says that @xmath91 for @xmath4 from the statement of the theorem [ fibration ] . hence @xmath376 . let @xmath377 be the obstruction for @xmath9 and @xmath165 to be homotopic . consider the loop @xmath378 that looks the same as @xmath9 except for a small period of time when we perform the deformation shown in figure [ obstruction.fig ] . clearly @xmath379 , and a straightforward verification show that the obstruction for @xmath378 and @xmath165 to be homotopic is @xmath380 ( for an appropriate choice of the generator of @xmath381 ) . performing this operation sufficiently many times we can change @xmath9 to be homotopic to @xmath165 . the kink participating in the deformation in figure [ obstruction.fig ] can be made very small , so that the deformation happens far away from the second component of the link . thus @xmath382 . since one of the two loops corresponding to a singular knot in figure [ obstruction.fig ] is contractible , it can not project to an orientation reversing loop in @xmath28 . hence @xmath383 . clearly @xmath384 and @xmath385 ( here the equalities denote the canonical isomorphisms and @xmath386 ( resp . @xmath387 ) is the centralizer of @xmath2 ( resp . @xmath3 ) in @xmath344 . ) hence the above observations imply that for @xmath388 such that @xmath389 the quantities @xmath390 and @xmath391 depend only on the element of @xmath392 realized by @xmath393 . [ mainidea2 ] combining this with [ mainidea ] we get that to prove the theorem it suffices to show that for every @xmath394 there exist @xmath169 and @xmath395 such that @xmath396 and either @xmath397 or @xmath398 . [ notcontractible ] _ consider the case of @xmath403 for @xmath404 . _ proposition [ toughandtechnical ] says that if @xmath405 is an orientation reversing loop in @xmath28 , then for every @xmath406 there exist @xmath169 and @xmath171 such that @xmath407 . the loop @xmath408 that has @xmath409 can be realized as the @xmath114-th power of @xmath410 , for the _ loop _ @xmath411 that is the sliding of both knots @xmath2 and @xmath3 along themselves according to their orientations . ( this deformation is induced by the rotation of the circle that parameterizes @xmath2 and @xmath3 . ) clearly no singular links occur under the deformation described by @xmath410 . hence @xmath412 and we have proved the statement of the theorem in the case of @xmath405 being an orientation reversing loop in @xmath161 . proposition [ toughandtechnical ] says that if @xmath405 is an orientation preserving loop in @xmath28 , then for every @xmath406 there exist @xmath169 and @xmath168 such that @xmath413 . the loop @xmath408 that has @xmath409 can be realized as a product of the @xmath414-th power of the loop @xmath415 and the @xmath114-th power of the loop @xmath410 , for @xmath415 described below ( and for @xmath410 described above ) . _ loop @xmath415 . _ since @xmath405 is an orientation preserving loop and @xmath4 is orientable , we get that the @xmath5-fibration over @xmath5 ( parameterizing the knots ) induced by @xmath416 is trivializable . hence we can coherently orient the fibers of this fibration . the orientation of the @xmath5-fiber over @xmath229 induces the orientation of the @xmath5-fiber of @xmath123 that contains @xmath417 . similarly for each @xmath229 we have an induced orientation of the @xmath5-fiber of @xmath123 that contains @xmath418 . ( in these two constructions we use the same orientation of the @xmath5-fibers of @xmath419 so that for every @xmath229 the orientations of the fibers containing @xmath417 and @xmath418 are coherent . ) the loop @xmath415 is the deformation of @xmath331 under which every point of @xmath331 slides once around the fiber of @xmath123 ( staying inside the fiber that contains the point ) in the direction specified by the orientation of the fiber . the input into @xmath421 could come only from singular links whose double points are located over the crossing points of @xmath422 . similarly the input into @xmath423 could come only from the double points of @xmath2 ( arising under the deformation @xmath415 ) and they could occur only over the selfintersection points of @xmath405 . the important crossing points of @xmath422 are located in pairs in small neighborhoods of the selfintersection points of @xmath405 , see figure [ doublepoint.fig ] . ( other points of @xmath422 correspond to extra twists of the framing . the two orientations of the fiber over them are the same , and hence the movement of branches of @xmath2 and of @xmath3 over such points happen in the same direction and does not give rise to singular links . ) if the double point @xmath43 of @xmath405 corresponding to a pair @xmath424 of crossing points of @xmath422 separates @xmath405 into two orientation preserving loops , then the sliding of the branches of @xmath2 and of @xmath3 located over @xmath425 and @xmath426 happen in the same direction and no double points of @xmath427 occur under this process . hence such @xmath43 corresponds to the zero input into @xmath421 . on the other hand the sliding of the two branches of @xmath2 located over such @xmath43 also happens in the same direction . thus the input into @xmath423 corresponding to such @xmath43 is also zero . if the double point @xmath43 of @xmath405 corresponding to a pair @xmath424 of crossing points of @xmath428 separates @xmath405 into two orientation reversing loops , then the sliding of the branches of @xmath2 and of @xmath3 located over @xmath425 ( and over @xmath426 ) happens in the opposite directions and such @xmath43 corresponds to the nontrivial input into @xmath421 . on the other hand the sliding of the two branches of @xmath2 located over @xmath43 also happens in the opposite directions , and such @xmath43 corresponds to a nontrivial input @xmath423 . a straightforward verification show that for such @xmath43 the input into @xmath421 corresponding to @xmath43 is twice the input into @xmath423 corresponding to @xmath43 . thus we have @xmath429 . clearly @xmath430 . substituting these expressions into @xmath431 and into @xmath432 we obtain identity . this finishes the proof of theorem [ fibration ] for @xmath405 being not contractible in @xmath433 . [ contractible ] _ consider the case of @xmath434 and @xmath435 . _ in this case @xmath436 . for all these surfaces @xmath186 has infinite order in @xmath174 . proposition [ commute ] implies that if @xmath437 , then @xmath386 is @xmath438 , the subgroup of @xmath174 consisting of the elements whose projection to @xmath28 are orientation preserving loops in @xmath28 . such @xmath447 can be realized as follows . consider a homotopy @xmath448 that changes @xmath2 and @xmath3 so that @xmath405 and @xmath449 are located in two small disks that do not intersect . ( we can assume that @xmath450 and @xmath451 do not move under this homotopy . ) next consider the deformation of the link described by the loop @xmath9 in @xmath330 such that @xmath452 , @xmath453 , and at each moment of the deformation @xmath405 and @xmath449 are in small disks that do not intersect . ( the position of the disks in @xmath28 changes under @xmath9 . ) finally deform @xmath2 and @xmath3 back to the original shape via @xmath454 . clearly @xmath348 and the inputs of @xmath448 and @xmath454 into @xmath390 cancel out . thus @xmath397 and we finished the proof of the theorem in the case of @xmath161 . _ proof of theorem [ fibration ] for @xmath455 . _ in this case @xmath174 is generated by the class @xmath186 of the fiber of @xmath456 . proposition [ pi2nontrivial ] says that @xmath457 is zero ( for @xmath4 from the statement of the theorem ) . hence @xmath186 has finite order in @xmath174 , and a certain degree of every @xmath458 is @xmath459 . the loop @xmath459 is realizable as @xmath460 for a trivial deformation @xmath447 ( under which nothing happens and hence no singular link occur ) . this finishes the proof of the theorem for @xmath455 . _ proof of theorem [ fibration ] for @xmath461 . _ proposition [ pi2nontrivial ] says that @xmath457 is zero ( for @xmath4 from the statement of the theorem ) . hence @xmath186 has finite order @xmath462 in @xmath174 . using proposition [ commute ] and the fact that @xmath463 we get that @xmath464 , for every @xmath465 . the loop @xmath459 is realizable as @xmath460 for a trivial deformation @xmath447 ( under which nothing happens and hence no singular link occur ) and we have proved the theorem for @xmath466 . _ proof of theorem [ fibration ] for @xmath467 . _ let @xmath468 be the fibration . since @xmath4 is oriented and the class @xmath186 of the fiber of @xmath123 is in the center of @xmath174 ( see [ commute ] ) we get that from one gets that @xmath473 every element of @xmath174 can be uniquely presented as @xmath474 , for some @xmath475 . let @xmath476 . using and the fact that @xmath186 is in the center of @xmath174 one verifies that @xmath477 commutes with @xmath478 if and only if @xmath479 . from this one concludes that if @xmath480 and @xmath481 , then a certain degree of every element of @xmath482 is expressible as @xmath460 for @xmath447 that is a product of the powers of the loops @xmath415 and @xmath410 described in [ notcontractible ] . similarly to [ notcontractible ] we get the proof of the theorem for this case . if @xmath483 and @xmath480 , then the proof of the theorem is analogous to [ contractible ] . if @xmath484 , then @xmath485 and @xmath4 has three structures of an @xmath5-fibration over @xmath90 . ( they are obtained by projecting @xmath4 on the three different pairs of circles . ) then every element of @xmath174 is realizable as @xmath460 for @xmath447 being a product of loops that are the powers of the versions of the loop @xmath415 with respect to the three @xmath5-fibration structures . we get that @xmath397 and this finishes the proof of theorem [ fibration ] in the case of @xmath467 . [ k]_proof of theorem [ fibration ] in the case of @xmath486 . _ using proposition [ commute ] , the fact that the class @xmath162 of the @xmath5-fiber of @xmath487 is of infinite order in @xmath174 , and the fact that a relation on the generators of @xmath174 projects ( under @xmath488 ) to a relation on the generators of @xmath489 we obtain that @xmath490 ( here @xmath491 and @xmath33 project respectively to @xmath492 and @xmath493 in figure [ klein1.fig ] depicting the klein bottle . ) every element of @xmath174 can be uniquely presented as @xmath495 , for some @xmath475 . taking @xmath496 and @xmath497 we find all the values of @xmath498 such that @xmath499 . we obtain that @xmath500 if and only if one of the following holds : 1 : : @xmath253 and @xmath501 are even , and @xmath502 ; 2 : : @xmath253 is even , @xmath501 is odd , @xmath503 , and @xmath504 ; 3 : : @xmath253 is odd , @xmath501 is even , @xmath505 , and @xmath506 ; 4 : : @xmath253 and @xmath501 are odd , @xmath507 , and @xmath508 . a : : if @xmath253 is odd , then a certain nonzero power of every @xmath509 can be realized as @xmath510 , for some @xmath511 ( see [ notcontractible ] for the definition of @xmath410 ) . and similarly to [ notcontractible ] we get the proof of the theorem for this case . b : : if @xmath253 is even and @xmath462 in is nonzero , then a certain nonzero power of every @xmath509 can be realized as @xmath512 , for some @xmath513 ( see [ notcontractible ] for the definitions of @xmath514 ) . and similarly to [ notcontractible ] we get the proof for this case . c : : if @xmath253 is even , @xmath484 , and @xmath515 , then @xmath482 is generated by @xmath516 . d : : if @xmath253 is even , @xmath484 , and @xmath517 with at least one of @xmath518 and @xmath36 being nonzero , then @xmath482 is generated by @xmath519 . consider an @xmath5-fibration @xmath520 induced by the orientation double cover @xmath521 . clearly @xmath199 is orientable and using we get that @xmath522 . using the fact that @xmath462 in is the euler class of the @xmath5-fibration over @xmath90 , we see that @xmath523 . this implies that either @xmath4 is @xmath524 ( the spherical tangent bundle of the klein bottle ) or the fiberwise projectivization of @xmath4 is @xmath524 . consider a loop @xmath89 in the space of autodiffeomorphisms of @xmath524 that is induced by the sliding of @xmath38 along the unit vector field parallel to the curve @xmath493 on @xmath38 ( see figure [ klein1.fig ] ) . ( note that @xmath38 has to slide twice along itself under this loop before every point of it comes to the original position . ) this @xmath89 induces an isotopy @xmath526 of @xmath331 such that @xmath527 , for some @xmath171 . since @xmath526 is an isotopy we have that @xmath528 . as usually @xmath529 and @xmath429 . next we observe that @xmath524 has another structure of the @xmath5-fibration over @xmath38 for which @xmath491 is the class of the @xmath5-fiber and the elements @xmath530 project to the generators of @xmath489 . let @xmath531 be the loop that is the version of @xmath415 with respect to this new structure of the @xmath5-fibration on @xmath524 . clearly @xmath532 and @xmath533 . similarly to [ notcontractible ] this finishes the proof of theorem [ fibration ] in the case when @xmath525 . the @xmath33-principle for curves [ h - principleforcurves ] says that the set @xmath291 of the connected components of the space of curves in @xmath4 is naturally identified with the set of the connected components of the space of free loops in @xmath225 the spherical tangent bundle of @xmath4 , or which is the same with the set of conjugacy classes of the elements of @xmath534 . from the long homotopy sequence of the fibration @xmath226 we see that it is also naturally identified with the set of conjugacy classes of the elements of @xmath174 . choose a @xmath258-structure on @xmath4 . it is easy to see ( cf . [ homotopy ] and [ proofhomotopy ] ) that the set @xmath535 of the connected components of the space of framed curves in @xmath4 is identified with the product @xmath536 . here the @xmath144-factor is the value of the @xmath258-structure on the loop in the principal @xmath259-bundle of @xmath231 that corresponds to a framed curve from the connected component , see [ proofhomotopy ] . ( this value does not depend on the choice of the framed curve in the component . ) the @xmath33-principle for the legendrian curves says that the set of the connected components of the space of legendrian curves is naturally identified with the set of homotopy classes of free loops in @xmath31 ( the spherical contact bundle of @xmath4 ) , or which is the same with the set of conjugacy classes of the elements of @xmath537 . since every contact manifold is oriented and the contact structure is cooriented , we see that the planes of the contact structure are naturally oriented . this orientation induces the orientation of the @xmath5-fibers of @xmath32 . put @xmath538 to be the class of the naturally oriented @xmath5-fiber of @xmath32 . the theorem of chow @xcite and rashevskii @xcite says that every connected component of the space of curves contains a legendrian curve . a straightforward verification shows that the insertion of the zig - zag into the legendrian curve @xmath38 ( see figure [ twocusp.fig ] ) changes the value of the @xmath258-structure on the corresponding framed curve . it is easy to verify ( see @xcite ) that the two connected components of the space of legendrian curves that contain @xmath38 and @xmath38 with the zigzag correspond respectively to the conjugacy classes of @xmath539 and of @xmath540 ( or of @xmath541 ) in @xmath537 . ( here the fact whether we get @xmath540 or @xmath541 depends on which of the two possible zig - zags we insert . ) let @xmath542 be a connected component of the space of legendrian curves in @xmath53 that corresponds to the conjugacy class of @xmath543 . then every connected component @xmath544 of the space of legendrian curves corresponds to the conjugacy class of @xmath545 , for some @xmath546 . put @xmath101 to the euler class of the contact bundle . consider the locally - trivial @xmath5-fibration @xmath557 induced by @xmath89 . one can verify that @xmath558 is the euler class of @xmath123 . on the other hand the euler class of @xmath123 is @xmath559 and is naturally identified with the value of @xmath101 on the homology class realized by @xmath127 . this implies that if @xmath52 does not satisfy condition , then there exists a homology class @xmath9 from the statement of the proposition . on the other hand the existence of the class @xmath9 from the statement of the proposition implies that there exists a legendrian curve @xmath560 such that @xmath539 is conjugate to @xmath561 , for @xmath36 being the value of @xmath101 ( the euler class of the contact bundle ) on the homology class realized by @xmath127 . ( proposition [ existcontact ] says that @xmath97 , for some @xmath562 , and hence @xmath36 is even . ) this means that @xmath52 does not satisfy condition and we have proved proposition [ interpretationconditionii ] . [ proofconditionii ] similar to the proof [ proofinterpretationconditionii ] of proposition [ interpretationconditionii ] we get that to prove that all the components of the space of curves satisfy condition it suffices to show that @xmath539 and @xmath548 are not conjugate in @xmath537 , for all @xmath169 and @xmath563 . let @xmath549 and @xmath546 be such that @xmath564 to prove the statement we have to show that @xmath439 . identity implies that @xmath551 and @xmath565 commute in @xmath174 . consider a mapping of the two - torus @xmath566 such that @xmath567 and @xmath568 . ( such a mapping exists since @xmath569 and @xmath570 commute in @xmath174 . ) by the assumption of the theorem @xmath571 has a nontrivial kernel . thus there exist @xmath168 with at least one of @xmath114 or @xmath414 being nonzero such that @xmath572 , and hence @xmath573 thus @xmath575 . applying to the last identity we get that @xmath576 . since @xmath91 we see that @xmath186 has infinite order in @xmath537 , and hence @xmath577 . if @xmath36 is zero , then we are done . hence we have to look at the case of @xmath578 . ( we assumed that @xmath574 . ) if @xmath578 , then @xmath579 . since we assumed that @xmath574 the assumption of the theorem implies that @xmath580 . thus @xmath165 is a power of @xmath186 and hence is in the center of @xmath537 . thus @xmath439 in and we have proved that all the components of the space of framed curves satisfy condition . [ proofconditioni ] the proof of the fact that every component @xmath52 of the space of framed curves satisfies condition is obtained by a straightforward generalization of the proof of theorem [ fibration ] ( see [ prooffibration ] ) and in fact is much simpler . from the condition that for every @xmath88 the homomorphism @xmath93 is not injective one gets that for every @xmath581 there exist @xmath582 such that at least one of @xmath36 and @xmath114 is nonzero and @xmath583 . if the connected component of the space of unframed curves that contains @xmath2 consists of noncontractible loops , then @xmath437 ( otherwise @xmath584 is an element of finite order which contradicts to the assumptions of the theorem ) . and then a certain nonzero degree of every @xmath394 can be realized as @xmath585 , for the loop @xmath410 introduced in [ notcontractible ] . similar to [ notcontractible ] this finishes the proof in the case of curves from @xmath52 realizing noncontractible loops . _ proof of statement * a : * of theorem [ example1 ] . _ clearly ( see figure [ homotopy.fig ] ) the two legendrian knots @xmath105 and @xmath2 belong to the same component of the space of legendrian curves . it is easy to see that if @xmath105 realizes the isotopy class of the framed knot @xmath586 , then @xmath2 realizes the isotopy class of @xmath587 ( see [ obstruction ] for the definition of @xmath587 ) . below we show that @xmath586 and @xmath587 are isotopic framed knots . let @xmath588 be the sphere that crosses @xmath586 at exactly one point , and let @xmath589\times s^2 $ ] be the tubular neighborhood of @xmath590 . fix @xmath591 ( below called the north pole ) and the direction in @xmath592 ( below called the zero meridian ) . we can assume that the knot @xmath586 inside @xmath589\times s^2 $ ] looks as follows : it intersects each @xmath593\times s^2 $ ] at the north pole of the corresponding sphere , and the framing of the knot is parallel to the zero meridian . consider an automorphism @xmath594 that is identical outside of @xmath589\times s^2 $ ] such that it rotates each @xmath595\times s^2 $ ] by @xmath596 around the north pole in the clockwise direction . clearly under this automorphism @xmath586 gets two extra negative twists of the framing and @xmath597 . on the other hand it is easy to see that @xmath598 is diffeotopic to the identity , since it corresponds to the contractible loop in @xmath599 . hence we see that @xmath586 and @xmath587 are isotopic framed knots . this finishes the proof of the statement * a : * of theorem [ example1 ] . [ props1xs2 ] let @xmath12 be a cooriented contact structure on @xmath195 with a nonzero euler class @xmath101 of the contact bundle . let @xmath31 be the spherical contact bundle , let @xmath32 be the corresponding locally trivial @xmath5-fibration , and let @xmath538 be the class of an oriented @xmath5-fiber of @xmath269 . then @xmath186 is of finite order in @xmath537 and @xmath600 . _ proof of proposition [ props1xs2 ] . _ consider the oriented @xmath13-plane bundle @xmath601 that is the restriction of the contact bundle over @xmath4 to a sphere @xmath602 . the euler class of @xmath235 is the value of @xmath101 on the homology class realized by @xmath603 , and hence is nonzero . let @xmath604 be the manifold obtained by the fiberwise spherization of @xmath123 , and let @xmath605 be the corresponding locally trivial @xmath5-fibration . since the euler class of @xmath235 is nonzero we get that a certain multiple of the class of the fiber of @xmath606 is homologous to zero . but @xmath607 is generated by the class of the fiber and hence the class of the fiber of @xmath606 is of finite order in @xmath607 . this implies that @xmath538 is of finite order . _ proof of statement * b : * of theorem [ example1 ] . _ let @xmath51 be the connected component of the space of legendrian curves that contains @xmath105 and @xmath2 . figure [ homotopy.fig ] shows that @xmath105 can be changed to @xmath2 ( in the space of legendrian curves ) by a sequence of isotopies and one passage through a transverse double point . hence if there exists a @xmath107-valued invariant @xmath108 of legendrian knots from @xmath51 that increases by one under every positive passage through a transverse double point of a legendrian knot , then it distinguishes @xmath105 and @xmath2 . ( clearly if such @xmath108 does exist , then it is an order one invariant of legendrian knots . ) below we show the existence of such @xmath108 in the connected component @xmath51 of the space of legendrian curves that contains @xmath105 . put @xmath608 . let @xmath609 be a legendrian knot , and let @xmath447 be a generic path in @xmath51 connecting @xmath105 and @xmath610 . let @xmath611 be the set of moments when @xmath447 crosses the discriminant ( i.e. a subspace of singular knots ) in @xmath51 , and let @xmath612 , @xmath613 , be the signs of these crossings . for a generic path @xmath614 put @xmath615 . it is clear that if @xmath108 ( with @xmath608 ) does exist , then @xmath616 . to show that @xmath108 does exist we have to verify that @xmath390 does not depend on the choice of a generic path @xmath447 connecting @xmath105 and @xmath610 , or which is the same we have to show that @xmath352 is zero for every generic closed loop @xmath9 connecting @xmath105 to itself . there are two codimension two strata of the discriminant of @xmath51 . they are formed respectively by singular legendrian knots with two transverse double points , and by legendrian knots with one double point at which the two intersecting branches are tangent of order one . it is easy to see that @xmath350 , for every small closed loop @xmath165 going around a codimension two stratum of @xmath51 . the @xmath33-principle says that the space of legendrian curves in @xmath30 is weak homotopy equivalent to the space @xmath34 of free loops in the spherical contact bundle @xmath31 of @xmath195 . ( the mapping giving the equivalence lifts a legendrian curve @xmath38 to a loop @xmath539 in @xmath31 by sending @xmath229 to the point in @xmath31 that corresponds to the direction of the velocity vector of @xmath38 at @xmath230 . ) hence @xmath617 is naturally isomorphic to @xmath622 . since @xmath600 by proposition [ props1xs2 ] , we get ( see [ hansen ] ) that @xmath623 is isomorphic to @xmath624 the centralizer of @xmath625 . using propositions [ commute ] and [ props1xs2 ] we see that @xmath626 , for some nonzero @xmath627 , or @xmath628 . hence there exists @xmath546 and nonzero @xmath629 such that @xmath630 . ( one should take @xmath36 and @xmath253 to be divisible by @xmath123 if @xmath626 . ) but the loop @xmath9 in @xmath617 corresponding to @xmath631 is just the sliding @xmath36-times of @xmath105 along itself according to the orientation . ( this deformation is induced by the rotation of the parameterizing circle . ) this loop does not intersect the discriminant , and hence @xmath348 . this finishes the proof of the statement * b : * of theorem [ example1 ] . [ part1]_@xmath2 and @xmath3 are homotopic legendrian curves , and @xmath2 and @xmath3 realize isotopic framed knots . _ let @xmath632 be the class of the @xmath5-fiber of the fibration @xmath32 . the @xmath33-principle says that the connected component of the space of legendrian curves that contains @xmath38 corresponds to the conjugacy class of @xmath563 . from the work of fuchs and tabachnikov it is easy too see that the connected components containing @xmath2 and @xmath3 correspond to the conjugacy classes of @xmath633 and of @xmath634 . let @xmath635 be an element projecting to the class @xmath162 of the @xmath5-fiber of @xmath120 . the value of the euler class of the contact bundle on the homology class realized by @xmath127 is equal to @xmath636 . ( here @xmath89 is the mapping from the description of the euler class of the contact bundle . ) and because of the reasons explained in the proof of proposition [ interpretationconditionii ] , we get that @xmath637 . now proposition [ commute ] implies that @xmath273 and @xmath274 are conjugate in @xmath537 and hence @xmath2 and @xmath3 are in the same component of the space of legendrian curves . the fact that @xmath2 and @xmath3 realize isotopic framed knots is clear , because as unframed knots they are the same , and as it is shown in @xcite every pair of extra cusps corresponds to the negative extra turn of the framing . [ ideaexample2]_idea of the proof of of the fact that @xmath2 and @xmath3 can be distinguished by an order one invariant of legendrian knots . _ let @xmath493 be a point in @xmath4 . let @xmath44 be a singular unframed knot with one double point . the double point separates @xmath44 into two oriented loops . deform @xmath44 preserving the double point , so that the double point is located at @xmath493 . choosing one of the two loops of @xmath44 we obtain an ordered set of two elements @xmath638 , or which is the same an element @xmath639 . clearly there is a unique element of the set @xmath640 that corresponds to the original singular unframed knot @xmath44 , where @xmath640 is is the quotient set of @xmath641 via the consequent actions of the following groups : 1 : : @xmath174 whose element @xmath235 acts on @xmath642 by sending it to @xmath643 . ( this corresponds to the ambiguity in deforming @xmath44 , so that the double point is located at @xmath493 . ) 2 : : @xmath144 that acts via the cyclic permutation of the two summands . ( this corresponds to the ambiguity in the choice of one of the two loops of @xmath44 . ) assume that @xmath649 is an invariant of legendrian knots from @xmath51 such that under every ( generic transverse ) positive passage through a discriminant in @xmath51 it increases by @xmath650 , where @xmath44 is the unframed singular knot corresponding to the crossing of the discriminant . clearly such @xmath649 is an order one invariant of framed knots from @xmath51 . to prove the theorem we show the existence of such @xmath649 , and then we show that it distinguishes @xmath2 and @xmath3 . [ existence]_the existence of @xmath649 . _ let @xmath447 be a generic closed loop in @xmath51 that starts with @xmath2 . let @xmath611 be the set of instances when @xmath447 crosses the discriminant ( i.e. a subspace of singular knots ) in @xmath51 and let @xmath612 , @xmath613 , be the signs of these crossings . let @xmath651 be those instances for which the value of @xmath652 on the corresponding singular unframed knots is @xmath8 . for a generic path @xmath614 put @xmath653 . let @xmath291 be the connected component of the space of unframed curves obtained by forgetting the framings on curves from @xmath52 , and let @xmath244 be the unframed knot obtained by forgetting the framing on @xmath654 . similarly to the above for a generic closed loop @xmath447 in @xmath291 starting with @xmath244 we put @xmath655 . ( as above @xmath656 is the set of instances when the value of @xmath652 on the singular unframed knots obtained under @xmath447 is equal to @xmath8 , and @xmath657 , @xmath658 are the signs of the corresponding crossings of the discriminant . ) the @xmath33-principle says that the space of legendrian curves in @xmath30 is weak homotopy equivalent to the space of free loops in @xmath31 the spherical contact bundle of @xmath4 . ( the mapping that gives an equivalence lifts a legendrian curve @xmath38 in @xmath53 to a loop @xmath539 in @xmath31 by mapping @xmath665 to the point of @xmath31 that corresponds to the velocity vector of @xmath38 at @xmath230 . ) since @xmath600 for @xmath4 from the statement of the theorem , we obtain ( see [ hansen ] ) the natural isomorphism @xmath666 . since @xmath667 and @xmath107 is torsion free , we get that to show the existence of @xmath649 it suffices to show that for every @xmath668 there exist @xmath169 and @xmath669 such that @xmath670 and @xmath671 . clearly @xmath685 and hence by the @xmath33-principle there is a loop @xmath686 such that @xmath687 . let @xmath688 be the loop corresponding to the deformation under which @xmath2 slides once around itself according to the orientation of @xmath2 . ( this deformation is induced by the rotation of the circle parameterizing @xmath2 . ) clearly @xmath689 . thus to prove the existence of @xmath649 it suffices to show that @xmath690 , for @xmath669 such that @xmath691 . but this @xmath447 is @xmath692 . thus to prove the theorem it suffices to show that @xmath693 . since @xmath689 we get that @xmath694 and thus it suffices to show that @xmath695 . using the fact that @xmath699 , the @xmath33-principle for curves [ h - principleforcurves ] , and observations of [ mainidea ] , and of [ obstructiondecrease ] we obtain that @xmath696 can be realized by a sequence of deformations shown in figure [ obstruction.fig ] . one of the loops of a singular knot arising under the deformation shown in figure [ obstruction.fig ] is contractible and the value of @xmath9 on such a singular knot was put to be zero . thus we have that @xmath700 , and this finishes the proof of the existence of @xmath649 . [ distinguish ] _ let us show that @xmath649 distinguishes @xmath2 and @xmath3 . _ as before let @xmath244 be the unframed knot obtained by forgetting the framing on @xmath2 . let @xmath701 be a generic path connecting @xmath2 and @xmath3 . to prove the theorem we have to show that @xmath702 . the homotopy @xmath703 gives rise to a homotopy @xmath704 connecting @xmath244 to itself . this @xmath705 has the property that the value of the euler class of the contact bundle on the homology class realized by the corresponding mapping @xmath134 is nonzero . moreover it is clear that @xmath706 . it is easy to see that the element of @xmath710 realized by the deformation in figure [ obstruction.fig ] is in the center of @xmath710 . now from the proof of theorem [ fibration ] one concludes that for every @xmath707 there exist @xmath169 such that @xmath708 can be expressed as a product @xmath711 of the powers of the loops @xmath415 , @xmath410 described in [ notcontractible ] and of the loop @xmath712 that is the deformation described in figure [ obstruction.fig ] . it is easy to see that if @xmath713 then @xmath714 , and that if @xmath578 then the value of the euler class on the homology class realized by the corresponding mapping of @xmath90 is zero . this finishes the proof of theorem [ example2 ] . * acknowledgements . * i am very grateful to o. ya . viro for the valuable discussions and suggestions . i am deeply thankful to h. geiges , a. stoimenow and s. tabachnikov for the valuable suggestions , and to o. baues , m. bhupal , n. a campo , a. cattaneo , j. frhlich , a. shumakovich and v. turaev for many valuable discussions . 99999 v.i . invarianty i perestroiki ploskih frontov , _ trudy mat . inst . steklov . * 209 * ( 1995 ) ; english translation : _ invariants and perestroikas of wave fronts on the plane , _ singularities of smooth mappings with additional structures , _ proc . steklov inst . * 209 * ( 1995 ) , pp . j. e. andersen , j. mattes , n. reshetikhin : _ quantization of the algebra of chord diagrams , _ math . cambridge philos . , vol . * 124 * no . 3 ( 1998 ) , 451467 w. l. chow , _ ber systemevon linearen partiellen differentialgleichungen erster ordnung , _ math . * 117 * ( 1939 ) , pp . 98105 m. do carmo : _ riemannian geometry , _ birkhuser , boston ( 1992 ) ya . eliashberg : _ classification of overtwisted contact structures on @xmath22-manifolds , _ invent . * 98 * ( 1989 ) , pp . 623637 t. fiedler : _ a small state sum for knots , _ topology vol . * 32 * no . 2 ( 1993 ) , pp . 281 - 294 d. fuchs and s. tabachnikov : _ invariants of legendrian and transverse knots in the standard contact space , _ topology vol . * 36 * , no . 5 ( 1997 ) , 10251053 d. fuchs and s. tabachnikov joint results , private communication with s. tabachnikov ( 1999 ) v. goryunov : _ vassiliev type invariants in arnold s @xmath715-theory of plane curves without direct self - tangencies , _ topology vol . * 37 * no . 3 ( 1998 ) , pp . m. gromov : _ partial differential relations , _ springer - verlag , berlin heidelberg ( 1986 ) , pp . 338339 v. l. hansen : _ on the fundamental group of the mapping space , _ compositio math . * 28 * ( 1974 ) , pp . 3336 j. w. hill : _ vassiliev - type invariants in @xmath84-theory of planar fronts without dangerous self - tangencies , _ c. r. acad . paris sr . , vol . * 324 * no . 5 ( 1997 ) , 537542 m. kontsevich : _ vassiliev s knot invariants , _ gelfand seminar , pp . 137150 , adv . soviet math . , 16 , part 2 , amer . soc . , providence , ri ( 1993 ) . r. lutz : _ structures de contact sur les fibrs principaux en cercles de dimension @xmath22 , _ ann . fourier , vol . * 3 * ( 1977 ) , pp . 115 t. q. t. le and j. murakami : _ the universal vassiliev - kontsevich invariant for framed oriented links , _ compositio math . , vol . * 102 * no . 1 ( 1996 ) pp . k. rashevskii : _ about the possibility to connect any two points of a completely nonholonomic space by an admissible curve , _ uchen . libknecht ped . ser . mat . * 2 * ( 1938 ) , pp . 8394 v. tchernov : _ arnold - type invariants of wave fronts on surfaces _ , http://xxx.lanl.gov/ math.gt/9901133 ( 1999 )	we show that for a big class of contact manifolds the groups of order @xmath0 invariants ( with values in an arbitrary abelian group ) of legendrian , of transverse , and of framed knots are canonically isomorphic . on the other hand for an arbitrary cooriented contact structure on @xmath1 with the nonzero euler class of the contact bundle we construct examples of legendrian homotopic legendrian knots @xmath2 and @xmath3 such that they realize isotopic framed knots but can be distinguished by finite order invariants of legendrian knots in @xmath1 . we construct similar examples for a big class of contact manifolds @xmath4 such that @xmath4 is a total space of a locally trivial @xmath5-fibration over a nonorientable surface . we show that in some of these examples the complements of @xmath2 and of @xmath3 are overtwisted . in this paper @xmath6 is a not necessarily torsion free abelian group , and @xmath4 is a not necessarily compact connected orientable three dimensional riemannian manifold .
theoretical models for gamma - ray bursts ( grbs ) can be divided into two categories : those named fireball models that treat the shock interaction of relativistic outflows and production of gamma rays and afterglows in other wavelengths ( see , e.g. , mszros 2002 and zhang & mszros 2004 for reviews ) , and those that explore the central engine of the fireball , i.e. , the energy source of relativistic outflows . most popular models in the latter category are in common invoking a stellar - mass black hole accreting with a hypercritical rate , of the order of 1 . the main problem in these models is how to convert some fraction of the released gravitational energy of the accreted matter into a relativistic outflow , creating an explosion with energy @xmath0 @xmath1-@xmath2 ergs ( depending on whether emission is isotropic or it is beamed ) . two mechanisms have been proposed to tackle this problem : the neutrino emission and annihilation , and the energy extraction from the accretion disk and/or the black hole via magnetohydrodynamical processes ( see , e.g. , popham et al . 1999 and di matteo et al . 2002 for references ) . the former mechanism is easier to understand and can be calculated more accurately , and is the topic we wish to discuss in this paper . in the inner region of such a hyperaccretion disk the density and temperature are so high ( @xmath3 @xmath4 g @xmath5 , @xmath6 k ) that photons are totally trapped , and large amounts of energetic neutrinos are emitted , carrying away the viscous dissipation energy of accreted gas . annihilation of some fraction of emitted neutrinos produces a relativistic electron - positron outflow to power a grb . the properties of such a neutrino - cooled accretion disk model were first worked out in details by popham et al . ( 1999 ) . from the observational point of view , the key question to be answered in this model is whether the neutrino annihilation can indeed provide sufficient energy for grbs . popham et al . ( 1999 ) gave a positive answer to this question , but they assumed a priori that the accretion disk is transparent for neutrinos , thus their neutrino radiation luminosity ( before annihilation ) @xmath7 and accordingly neutrino annihilation luminosity @xmath8 might be overestimated . later , di matteo et al . ( 2002 ) recalculated @xmath7 and @xmath8 in the neutrino - cooled accretion disk model , taking the neutrino opacity into account . they obtained that even for a modest mass accretion rate the inner region of the disk becomes opaque and neutrinos are sufficiently trapped , the resulting @xmath8 is @xmath9 @xmath1 ergs s@xmath10 and is inadequate for grbs . in a recent work ( gu et al . 2006 ) we showed that when the general relativistic effect is considered and the contribution from the optically thick region is included , the neutrino - cooled accretion disk can work as the central engine for grbs , although the correct @xmath8 is somewhat lower than that of popham et al . ( 1999 ) because of the effect of neutrino opacity . in addition to the general relativity and the neutrino opacity , there are certainly other factors that may influence the neutrino radiation and annihilation of a neutrino - cooled accretion disk , such as the electron degeneracy and the electron fraction . in our previous work ( gu et al . 2006 ) the treatment of electron degeneracy was oversimplified , and the electron fraction @xmath11 was simply taken to be 0.5 , i.e. , assuming an equal mix of protons and neutrons since @xmath12 = @xmath13/(@xmath13 + @xmath14 ) , where @xmath13 and @xmath14 are the total number density of protons and that of neutrons , respectively . kohri & mineshige ( 2002 ) pointed out that when electrons are degenerate , there emerges an important consequence that the electron - positron pair creation and accordingly the neutrino emission are suppressed . kohri et al . ( 2005 ) took great care to calculate the ratio of free neutrons to free protons @xmath15@xmath16@xmath17 since this ratio has a large effect on the neutrino emission rates : the true ratio @xmath15@xmath16@xmath17@xmath18 1 ( or @xmath19 ) resulted from neutronization processes will also lead to a suppression of neutrino emission ( note that they did not distinguish between @xmath14 , @xmath13 and @xmath20,@xmath17 , see 2.3.1 ) . they made calculations for the neutrino emission rates and other quantities even in the delicate regime where the electron degeneracy is moderate , which is also a significant improvement over previous works . lee et al . ( 2005 ) also considered the effects of electron degeneracy and electron fraction in neutrino - cooled accretion disks : they used an expression for the pressure of ultra - relativistic electrons with arbitrary degeneracy and calculated @xmath12 with an approximate bridging formula that allows for a transition from the neutrino optically thin to optically thick regime . these works ( kohri & mineshige 2002 ; kohri et al . 2005 ; lee et al . 2005 ) made advances in microphysics , but they were within the newtonian framework . very recently , chen & beloborodov ( 2006 ) presented fully relativistic calculations of the structure of neutrino - cooled accretion disks around kerr black holes and proved that both the electron degeneracy and the electron fraction dramatically affect the disk and its neutrino emission . it is seen from the above brief review that the electron degeneracy and the lower electron fraction certainly suppress the neutrino emission and reduce the neutrino annihilation luminosity of a neutrino - cooled accretion disk . the question remains whether the reduced neutrino annihilation luminosity is still adequate for grbs . in this paper we try to refine our previous results of the structure and luminosity of neutrino - cooled accretion disks by considering the relevant microphysics more completely and more accurately . we limit the central accreting black hole to be a non - rotating one , its general relativistic effect is simulated by the well - known paczyski & wiita ( 1980 ) potential @xmath21 , where @xmath22 is the black hole mass , @xmath23 is the radius , and @xmath24 is the schwarzschild radius . as in the relevant previous works , the hydrodynamics of hyperaccretion disks is expected to be similar to that of normal accretion disks in x - ray binaries , and is well approximated by that of steady axisymmetric height - averaged accretion flows ( e.g. , chap . 3 of kato et al . accretion in the disk is driven by viscous stress , and the kinematic viscosity coefficient is expressed as @xmath25 , where @xmath26 is the half thickness of the disk ; @xmath27 is the isothermal sound speed , with @xmath28 and @xmath29 being the pressure and mass density , respectively ; and @xmath30 is a dimensionless constant parameter that absorbs all the detailed microphysics of viscous processes . the angular velocity is approximately keplerian , i.e. , @xmath31 . the disk is in the vertical hydrostatic equilibrium , and this gives @xmath32 . with these simplifications the problem is reduced to be one - dimensional , i.e. , all physical quantities depend on @xmath23 only . the constant mass accretion rate @xmath33 is expressed from the continuity equation as = -4rh v , where @xmath34 is the radial velocity that can be read from the angular momentum equation as v = h c_s(1 - j/ _ k r^ 2)^- 1 , where @xmath35 is an integration constant determined by the zero - torque boundary condition at the last stable orbit , and it represents the specific angular momentum ( per unit mass ) of the matter accreted into the black hole . the energy equation is generally written as the balance between the viscous heating and the cooling rates ( per unit area of a half - disk above or below the equator ) , = q^- . the viscous heating rate @xmath36 is similar to that of normal accretion disks , = - _ k r ( 1-j/_k r^2 ) . but the cooling rate @xmath37 is crucially different , it has three contributions : q^-=+ + . in this equation there is no cooling term of photon radiation ( it is practically zero in our calculations ) . instead , photons are totally trapped in the disk , so they contribute to the advective cooling and the pressure ( see below ) . the cooling rate by photodisintegration of @xmath30-particles @xmath38 is = 6.8 10 ^ 28 _ 10 v h , where @xmath39 , and @xmath40 is the mass fraction of free nucleons ( e.g. , kohri et al . 2005 ) . the advective cooling rate @xmath41 is = v h t - v t ( a t^3 + + ) , where @xmath42 is the specific entropy , @xmath43 is taken to be equal to 1 , i.e. , @xmath44 is approximated as @xmath45 ( kohri & mineshige 2002 ) , and @xmath46 is the mean mass of a nucleon . the entropy of degenerate particles is small and can be neglected . the three terms in the brackets of equation ( 7 ) are the entropy density of photons , of free nucleons and @xmath30-particles , and of neutrinos , respectively ; and @xmath47 is the energy density of neutrinos , for which we adopt a bridging formula valid in both the optically thin and thick regimes ( popham & narayan 1995 ; di matteo et al . 2002 ) , u_=_i , where @xmath48 is the total optical depth for neutrinos , @xmath49 is the absorption optical depth for neutrinos , and the subscript @xmath50 runs for the three species of neutrinos @xmath51 , @xmath52 , and @xmath53 . the cooling rate due to neutrino loss @xmath54 is expressed in accordance with the above equation(kohri et al . 2005 ) , q_=_i . the equation of state is also very different from that of normal accretion disks , as the contributions to the pressure from degenerate electrons ( we assume that nucleons are not degenerate throughout the present paper ) and from neutrinos should be included , it is written as p = p_gas+p_rad+p_e+p_. the gas pressure from free nucleons and @xmath30-particles @xmath55 is p_gas= . the photon radiation pressure @xmath56 is p_rad = at^4/3 . the electron pressure @xmath57 is from both electrons and positrons and should be calculated using the exact fermi - dirac distribution . no asymptotic expansions are valid because at different radii electrons may be with different degrees of degeneracy and may be relativistic or nonrelativistic . it reads p_e = p_e^-+ p_e^+ , with p_e^= _ 0^ d p , where @xmath58 is the chemical potential of electrons , and the electron degeneracy is measured by the degeneracy parameter defined as @xmath59 . we agree with lee et al . ( 2005 ) that since the presence of relativistic @xmath60 pairs is automatically taken into account in the expression for @xmath57 , there is no alteration to the numerical factor 1/3 in the expression for @xmath56 , nor to the factor 4/3 for the photon entropy in equation ( 7 ) . the neutrino pressure @xmath61 is p_=u _ / 3 . all the equations in the above two subsections can be combined into only two equations , i.e. , equations ( 1 ) and ( 3 ) . in these two equations there are seven unknown quantities , namely @xmath29 , @xmath62 , , @xmath63(@xmath64 ) , and @xmath58(or @xmath65 ) . therefore , one has to find more equations relating these unknowns with the knowledge of microphysics . the total optical depth for neutrinos is _ i + _ a,_i . the optical depth for neutrinos through scattering off free nucleons , @xmath30-particles , and electrons @xmath66 is given by _ s , _ i = h/__i = h [ _ p,_i _ p+_n,_i _ n+_,_i n_+_e,_i ( n_e^-+n_e^+ ) ] , where @xmath67 is the mean free path ; @xmath68 , @xmath69 , @xmath70 and @xmath71 are the cross sections of scattering on protons , neutrons , @xmath30-particles , and electrons ; @xmath17 , @xmath15 , @xmath72 , @xmath73 and @xmath74 are the number densities of free protons , free neutrons , @xmath30-particles , electrons , and positrons , respectively . the four cross sections are given by ( burrows & thompson 2002 ) _ p,_i=[(c_v,_i-1)^2 + 3 g_a^2 ( c_a,_i-1)^2 ] , _ n,_i= , _ , _ i=4 _ 0 e__i^2 ^4 _ w , _ e,_i= ( 1 + ) [ ( c_v,_i+c_a,_i)^2 + ( c_v,_i - c_a,_i)^2],where @xmath75,@xmath76 is the mean energy of neutrinos in units of @xmath77 , @xmath78 , @xmath79 , @xmath80 , @xmath81 , @xmath82 , @xmath83 . since @xmath84 and @xmath85 , the free proton fraction @xmath86 , the free neutron fraction , and the @xmath30-particle fraction @xmath87 are related to @xmath11 and @xmath88 as y_p= = y_e- , y_n==1-y_e- , y_= = , and @xmath73 , @xmath74 are given by the fermi - dirac integration , n_e^= _ 0^d p p^2 ( cf . kohri et al . 2005 , they took @xmath89 and @xmath90 , and thus @xmath91 and @xmath92 , which are valid only for the completely dissociated matter , i.e. , @xmath93 ) . the absorption depth for neutrinos @xmath94 is defined by _ a , _ i= , where @xmath95 is the total neutrino cooling rate ( per unit volume ) and is the sum of four terms , q__i = q _ urca+q_e^- e^++q_brem+q_plasmon.the neutrino cooling rate due to the urca processes @xmath96 relates only to @xmath51 and is represented by the sum of three terms ( chap . 11 of shapiro & teukolsky 1983 ; yuan 2005 ) , q_urca = q_p+e^-n+_e+q_n+e^+p+_e+q_n p+e^-+_e , with q_p+e^-n+_e = ( 1 + 3 g_a^2 ) _ p _ q^d e_e e_e ( e_e - q)^3 f_e^- , q_n+e^+p+_e=(1 + 3 g_a^2 ) _ n _ m_e c^2^d e_e e_e ( e_e+q)^3 f_e^+ , q_n p+e^-+_e=(1 + 3 g_a^2 ) _ n _ m_e c^2^q d e_e e_e ( q - e_e)^3 ( 1-f_e^- ) , where @xmath97 , @xmath98 , @xmath99 , and @xmath100 + 1\}^{-1}$]is the fermi - dirac function . the last term in the right hand side of equation ( 28 ) , i.e. , equation ( 31 ) , is small comparing with the other two terms , and was usually not included in the literature . the electron - positron pair annihilation rate into neutrinos @xmath101 is ( e.g. , itoh et al . 1989 ) q_e^-e^+ _ e _ e 3.4 10 ^ 33 t_11 ^ 9 ergs cm^-3 s^-1 , q_e^-e^+ _ _ = q_e^-e^+ _ _ 0.7 10 ^ 33 t_11 ^ 9 ergs cm^-3 s^-1,where @xmath102 . expressions ( 32 ) and ( 33 ) are valid in the nondegenerate limit @xmath103 , and when electrons are degenerate @xmath104 becomes negligible . the nucleon - nucleon bremsstrahlung rate @xmath105 through the processes @xmath106 is the same for the three species of neutrinos ( hannestad & raffelt 1998 ; burrows et al . 2000 ) , q_brem 1.5 10 ^ 27 _ 10 ^ 2 t_11 ^ 5.5 ergs cm^-3 s^-1.as to the plasmon decay rate @xmath107 , only that through the process @xmath108 needs to be considered , where plasmons @xmath109 are photons interacting with electrons , q_plasmon 1.5 10 ^ 32 t_11 ^ 9 _ p^6 ( 1 + _ p ) ( 2 + ) ergs cm^-3 s^-1,where @xmath110}^{1/2}$](ruffert et al . it is expected that @xmath105 and @xmath107 can become important only at very high electron degeneracy . it is seen that the electron fraction @xmath11 mentioned in introduction has appeared in the expression of neutrino optical depth @xmath63 . in order to calculate @xmath63 and accordingly the neutrino cooling rate @xmath111 ( see eq . [ 9 ] ) , more knowledge about @xmath11 is required . beloborodov ( 2003 ) proved that @xmath112-equilibrium among free neutrons , free protons , and electrons is established in disks with @xmath113 , a condition that is likely to be fulfilled for hyperaccretion disks ; and noted that a distinction needs to be made to determine the equilibrium composition depending on the optical depth of the disk material . kohri et al.(2005 ) discussed various timescales in neutrino - cooled accretion disks in details , and showed that for sufficiently large @xmath33 and not too large @xmath23 the timescale for the reactions from proton to neutron and from neutron to proton is shorter than the dynamical timescale ( the accretion time ) , i.e. , the @xmath112-equilibrium is likely to be realized . if the disk material is opaque to neutrinos , there are reversible reactions @xmath114 and @xmath115 . the chemical potential of neutrinos can be ignored because the number density of neutrinos and that of antineutrinos are likely to be equal , then the @xmath112-equilibrium condition is _ n=_p+_e , where @xmath116 and @xmath117 are chemical potentials of neutrons and protons , respectively . on the other hand , if the material is transparent to neutrinos , there are no reversible reactions as for the neutrino opaque material , but the @xmath112-equilibrium can be reached when the rate of reaction @xmath118 is equal to that of reaction @xmath119 . yuan(2005 ) calculated these two rates and obtained _ n=_p+2 _ e. the neutron - to - proton ratio in @xmath112-equilibrium is given by = , which results in = , for the neutrino opaque limit , and = , for the neutrino transparent limit . in order to allow for a transition from the optically thin to optically thick regime , we adopt a treatment similar to that in lee et al . ( 2005 ) , i.e. introducing a weight factor @xmath120 and writing in a combined form,= f ( _ ) + [ 1-f ( _ ) ] . by using the relation @xmath121 and equations ( 22 ) , ( 23 ) , and ( 24 ) we finally arrive at y_e=(1-x_nuc)+ . some comments on equation ( 42 ) are in order . first , this equation is more exact than equation ( 12 ) of lee et al . ( 2005 ) , theirs is the first term of the expansion of ours . second , we notice that kohri et al . ( 2005 ) discussed the ratio @xmath122 at length . they started with considering the transition rates from proton to neutron and from neutron to proton , then obtained the expression of @xmath123 , i.e. , equation ( 39 ) for the neutrino opaque limit , and adopted an approximate procedure to estimate this ratio for the neutrino optically thin regime . we stress that equation ( 40 ) is rigorous for the neutrino transparent limit , since it is also derived from the transition rates between protons and neutrons ( yuan 2005 ) ; then we construct equation ( 42 ) as a bridging formula between the opaque and transparent limits and expect it to work for the regime where the neutrino optical depth is moderate . an additional relation between @xmath11 and @xmath40 can be found from the equation of nuclear statistical equilibrium ( e.g. , meyer 1994 ) , y(z , a ) = g(z , a)[(3)^a-1 ^(1-a)/2 2^(3a-5)/2 ] a^3/2 ( ) ^3(a-1)/2 ^1-a y_p^z y_n^a - z , where @xmath124 is the mass fraction of a kind of particles with the charge number @xmath125 and mass number @xmath126 , @xmath127 is the nuclear partition function , @xmath128 is the riemann zeta function of argument 3 , @xmath129/ [ \pi^2 ( \hbar c)^3 \rho n_a]$ ] is the photon - to - baryon ratio , @xmath130 is the avagadro constant , and @xmath131 is the binding energy of the nucleus . as in all the previous works , we assume that all heavy nuclei are @xmath30-particles . this should be a reasonable assumption because all nuclei heavier than @xmath30-particles contain approximately equal numbers of neutrons and protons . then by using equations ( 22 ) , ( 23 ) , and ( 24 ) , equation ( 43 ) yields 4(y_e- ) ( 1-y_e- ) ( 1-x_nuc)^-1/2=8.71 10 ^ 4 _ 10 ^ -3/2 t_11 ^ 9/4 .only chen & beloborodov ( 2006 ) adopted this equation , while other previous works used a simple expression of @xmath40 that was obtained by taking @xmath132 and did not reflect the interdependence between @xmath11 and @xmath40 ( e.g. , eq . [ 40 ] of kohri et al . 2005 ) . the electron chemical potential is determined by the condition of charge neutrality among protons , electrons , and positrons , n_p= = n_e^--n_e^+,with @xmath73 and @xmath133 given in equation ( 25 ) . the system of equations is closed , as there are eight equations , i.e. , equations ( 1 ) , ( 3 ) , ( 42 ) , ( 44 ) , ( 45 ) , and three equations ( 16 ) for eight independent unknowns @xmath29,@xmath62,@xmath134,@xmath135,@xmath136,@xmath40 , @xmath11,and @xmath58(or @xmath65 ) , which can be numerically solved as functions of @xmath23 with given constant parameters @xmath22,@xmath33,@xmath30 , and @xmath35 . all the other quantities such as @xmath28 and its components , @xmath37 and its components , and composition fractions , , and @xmath87 are obtained accordingly . our equation set has the following advantages : ( 1 ) effects of relevant factors are taken into account in a combined way , including the general relativity , the inner boundary condition of the disk , various processes that contribute to the neutrino cooling and the neutrino opacity , the electron degeneracy , the electron fraction , and the coexistence of electrons , positrons , free protons , free neutrons , and @xmath30-particles . ( 2 ) whenever the electron degeneracy is concerned we use the exact fermi - dirac integration rather than the analytical approximations that are valid only for extreme cases . ( 3 ) we take great care to calculate the neutrino optical depth and the electron fraction . in doing so , we make a careful distinction between the total nucleon number densities @xmath137 , @xmath138 and the free nucleon number densities @xmath139,@xmath140 , so that the composition of the disk matter can be exactly described by fractions @xmath11 , @xmath141 , @xmath142 , and @xmath87 ; we propose a new bridging formula for @xmath11 ( eq . [ 42 ] ) from the @xmath112-equilibrium condition , which is applicable to both the neutrino optically thin and optically thick regimes ; and we adopt equation ( 44 ) that connects @xmath11 with . this section presents our results for the structure of a neutrino - cooled accretion disk , obtained by numerically solving the set of eight equations in 2 . figures 1 - 6 show physical quantities of the disk matter as functions of @xmath23 . in all these figures the necessary constant parameters are fixed to be their most typical values , i.e. , @xmath143 , @xmath33= 1 ( except for fig . 5 where @xmath33= 5),@xmath144 , and @xmath145 ( see , e.g. , gu et al . 2006 for the discussion of @xmath35 ) . figure 1 is for the density @xmath29 , temperature @xmath62 , and electron degeneracy parameter @xmath65 . it is seen that from @xmath146 inward to @xmath147 , @xmath29 increases by about four orders of magnitudes ( fig . 1a ) , @xmath62 increases by about eight times ( fig . 1b ) ; and in the innermost region of the disk @xmath29 reaches to @xmath148 , and @xmath62 reaches to @xmath149 . with such densities and temperatures , @xmath150 increases first with decreasing @xmath23 , reaches to its maximum value at @xmath151 , and then decreases with decreasing @xmath23 because of increasing temperature ( fig . the behavior of @xmath65 obtained by us is consistent with that of kohri et al . ( 2005 , fig . 4 there ) ; but is not so with that of chen & beloborodov ( 2006 , fig . 3 there ) , who got that @xmath65 always increases with decreasing @xmath23 in the very inner region . we think that @xmath65 ought to decrease when temperature is very high . it is important to note that the value of @xmath65 is of order a few , i.e. , the electron degeneracy is moderate . this justifies that the exact fermi - dirac distribution must be used in the calculations , and analytic approximations for either extremely degenerate electrons ( @xmath152 ) or fully nondegenerate electrons ( @xmath103 ) are invalid for hyperaccretion disks . figure 2 shows the composition of disk matter . in the outer region between @xmath153 and @xmath154 , almost all @xmath30-particles are not disintegrated , so that the @xmath30-particles fraction @xmath87 keeps to be @xmath155 , the electron fraction @xmath11 keeps to be @xmath156 , and the free proton fraction , the free neutron fraction , and the free nucleon fraction @xmath40 are all keeping @xmath157 . from @xmath154 inwards , the disintegration of @xmath30-particles causes @xmath87 to decreases and @xmath88 to increase dramatically . because of the neutronization processes favored by high temperature , @xmath142 greatly exceeds , and @xmath11 decreases accordingly . in the innermost region @xmath158 , @xmath30-particles are almost fully disintegrated , i.e. , @xmath159 , @xmath160 ; and the neutronization makes @xmath142 larger than 0.9 , and @xmath141 and @xmath11 smaller than 0.1 . contributions to the total pressure @xmath28 from free nucleons and @xmath30-particles @xmath55 , from degenerate electrons and positrons @xmath57 , from photon radiation @xmath56 , and from neutrinos @xmath61 are drawn in figure 3 . it is seen that @xmath161 in the outer region @xmath162 , and @xmath55 becomes dominant at smaller radii because both of the disintegration of @xmath30-particles and the decreasing degeneracy of electrons ; @xmath56 is of some importance only in the outer region , and @xmath61 is small even in the innermost region @xmath158 where the disk becomes optically thick to neutrinos . the neutrino optical depth @xmath63 is plotted in figure 4 . two comparisons are made in figure 4a . first , the optical depth for electron neutrinos @xmath163 is several times larger than that for @xmath164-neutrinos @xmath135 and @xmath165-neutrinos @xmath136 , and only @xmath163 can become larger than 2/3 in the innermost region @xmath158 , therefore @xmath163 ought to be taken as the representative of @xmath166 . second , the contribution to @xmath163 from absorption @xmath167 is more important than that from scattering @xmath168 , this is a surprising result as the general understanding in the literature was that scattering off nucleons is a far more important opacity source than absorption(e.g . , di matteo et al . 2002 ; lee et al . we notice that in calculating the neutrino cooling rate due to the urca processes @xmath96 ( which is dominant over other neutrino cooling rates , see fig . 4b ) those authors used an approximate formula that is valid only in the nondegeneracy limit . in fact , the electron degeneracy causes @xmath96 to increase greatly as in our calculations . according to equation ( 46 ) of kohri & mineshige ( 2002 ) , in the complete electron degeneracy limit @xmath96 becomes extremely large because @xmath169 ! it is indeed the case as seen from figure 4b that the contribution to @xmath167 from absorption relating to the urca processes @xmath170 is most important , the contribution from neutrino annihilation into @xmath171 pairs @xmath172 is of some relative importance only in the very outer region , and the contributions from nucleon - nucleon bremsstrahlung @xmath173 and from the inverse process of plasmon decay @xmath174 are totally negligible because the electron degeneracy is not very high . as to @xmath175 ( fig . 4c ) , the contribution through scattering off free neutrons @xmath176 is dominant over other contributions due to free protons @xmath177 , electrons @xmath178 , and @xmath30-particles @xmath179 ( @xmath179 is too small to be seen in the figure ) , obviously because of the richness of neutrons . having @xmath163 in mind , we now check the validity of our bridging formula for @xmath11 , i.e. , equation ( 42 ) . in figure 5 , @xmath11 calculated using equation ( 42 ) is drawn by the solid line . here @xmath180= 5is taken because only at such a high accretion rate the inner region of the disk can become noticeably optically thick . for comparison , the dashed line in the figure draws @xmath11 in the case that the disk material is assumed to be globally opaque to neutrinos , which is calculated with ( see eq . [ 39 ] ) y_e= ( 1-x_nuc)+x_nuc[1+]^-1 , and the dotted line draws @xmath11 if the material is globally transparent to neutrinos , obtained with the approximate formula of beloborodov ( 2003 ) , y_e= ( 1-x_nuc)+x_nuc [ + 0.487 ( ) ] . these two equations do not include the effect of varying neutrino optical depth . it is clear that our equation ( 42 ) does represent a bridge between the optically very thick and optically very thin limits , and it provides a reasonable estimate of @xmath11 in the intermediate regime where @xmath11 is overestimated by equation ( 46 ) and underestimated by equation ( 47 ) . in figure 6 we show the ratios of various cooling rates to the viscous heating rate @xmath36 . the photon radiation cooling is never important in hyperaccretion disks ( it is practically zero when we try to calculate it ) . advective cooling @xmath41 is dominant only in the outer region of the disk @xmath181 , because in this region the photodisintegration of @xmath30-particles has not started and the neutrino emission is weak . in the middle region @xmath182 , the @xmath30-particle photodisintegration cooling @xmath38 dominates , corresponding to a sharp decrease of @xmath87 and increase of ( see fig . 2 ) . in the inner region @xmath183 , the neutrino cooling @xmath111 becomes dominant as expected . we do not present here numerical results for the disk structure with varying values of @xmath30 and @xmath33 . differences caused by changing these two parameters are only quantitative and have been discussed in , e.g. , kohri et al . ( 2005 ) and chen & beloborodov ( 2006 ) . briefly speaking , @xmath30 smaller than 0.1 will make the density higher , the electron degeneracy higher , the electron fraction lower , and the neutrino - dominated region larger ; and at still larger accretion rates , advection can become dominant over neutrino cooling again in a very small region ( @xmath184 for @xmath33=5 , see fig . 3 of gu et al . 2006 ) , because in that region the optical depth is very large and neutrinos are trapped in the disk . having the neutrino cooling rate @xmath111 , the neutrino radiation luminosity ( before annihilation ) @xmath185 is obtained as l_=4 _ r_in^r_out q _ r d r. in our calculations the inner and outer edge of the disk are taken to be @xmath186 and @xmath187 , respectively . for the calculation of the neutrino annihilation luminosity we follow the approach in ruffert et al . ( 1997 ) , popham et al . ( 1999 ) , and rosswog et al . the disk is modeled as a grid of cells in the equatorial plane . a cell @xmath188 has its mean neutrino energy @xmath189 , neutrino radiation luminosity @xmath190 , and distance to a space point above ( or below ) the disk @xmath191 . the angle at which neutrinos from cell @xmath188 encounter antineutrinos from another cell @xmath192 at that point is denoted as @xmath193 . then the neutrino annihilation luminosity at that point is given by the summation over all pairs of cells , l_=_i a_1,i _ k _ k ( _ _ i^k + _ _ i^k ) ( 1-)^2 + + _ i a_2,i _ k _ k , where @xmath194[{(c_{v,\nu_{i}}-c_{a,\nu_{i}})}^2 + { ( c_{v,\nu_{i}}+c_{a,\nu_{i}})}^2]$ ] , @xmath195 @xmath196 , with @xmath197 and @xmath198 given in 2.3.1 . the total neutrino annihilation luminosity is obtained by the integration over the whole space outside the black hole and the disk , l_=4 _ r_g^_h^l _ r d r d z. figure 7 shows @xmath185 ( the thick dashed line ) and @xmath199 ( the thick solid line ) with varying @xmath33 ( @xmath200 , @xmath144 , and @xmath201 are kept ) . for comparison , these two luminosities calculated in our previous work ( gu et al . 2006 ) , where the electron degeneracy was not correctly considered and @xmath11 was taken to be equal to 0.5 , are also given in the figure by the thin dashed line and thin solid line , respectively . it is clear that the electron degeneracy and the lower @xmath11 resulted from the neutronization processes indeed suppress the neutrino emission considerably , the resulting @xmath202 and @xmath8 are reduced by a factor @xmath203 comparing with their overestimated values in gu et al . even so , the correct @xmath8 is still well above @xmath204 provided @xmath205 1 , and reaches to @xmath206 when @xmath207 10 . therefore , based on the energy consideration , neutrino - cooled accretion disks can work as the central engine of grbs . note that our calculations are for a nonrotating black hole , both popham et al . ( 1999 ) and chen & beloborodov ( 2006 ) have shown that a spinning ( kerr ) black hole will enhance the neutrino radiation efficiency , this only strengthens our conclusion here . in figure 8 the corresponding neutrino radiation efficiency @xmath208 and neutrino annihilation efficiency @xmath209 ) are shown by the dashed line and solid line , respectively . it is seen that @xmath210 does not change much with varying @xmath33 ; while @xmath211 increases rapidly with increasing @xmath33 , this is because at higher accretion rates more neutrinos are emitted and in turn have higher probabilities to encounter with each other . to see the spatial distribution of neutrino annihilation luminosity , we plot in figure 9 contours of ( @xmath212 ) in units of ( @xmath213 ) , i.e. , the neutrino annihilation luminosity of a circle with cylindrical coordinates @xmath23 and @xmath125 , for @xmath33= 1 . the figure has two important implications . first , it demonstrates the strong focusing of neutrino annihilation towards the central region of space . in fact , by performing the integration of equation ( 50 ) out to each radius we get that nearly 60% of the total annihilation luminosity is ejected from the region @xmath214 . second , it shows that the annihilation luminosity varies more rapidly along the @xmath125 coordinate than along the @xmath23 coordinate , indicating that this luminosity is anisotropic and most of the annihilation energy escapes outward along the angular momentum axis of the disk . in this paper we wish to discuss whether the annihilation of neutrinos emitted from hyperaccretion disks can provide sufficient energy for grbs , i.e. , to estimate the neutrino annihilation luminosity @xmath8 . to do this , we need to know the neutrino optical depth @xmath63 , because it determines the neutrino cooling rate @xmath111 ( eq . [ 9 ] ) and the neutrino radiation luminosity @xmath7 ( eq . [ 48 ] ) . to calculate contributions to @xmath63 from various absorption and scattering processes of neutrinos ( eq . [ 16 ] ) , we need to know the composition and physical state of disk matter , namely the fractions of electrons @xmath11 and free nucleons , and the electron chemical potential @xmath58 ( or electron degeneracy @xmath65 ) , with given density @xmath29 and temperature @xmath62 . we give exact definitions of the free proton fraction @xmath141 and free neutron fraction @xmath142 and their relations to @xmath11 and @xmath40 , and get three equations ( eqs . [ 42 ] , [ 44 ] , and [ 45 ] ) from the conditions of @xmath112-equilibrium , nuclear statistical equilibrium , and charge neutrality that describe the interdependence of @xmath141 , @xmath88 , @xmath58 , and @xmath63 . we prove that the electron degeneracy has important effects indeed , but mainly not in the sense that it suppresses the creation of neutrinos from @xmath215 pairs and enlarges the electron pressure ( kohri & mineshige 2002 ; kohri et al . as seen from figures 4b and 3 , the neutrino cooling due to @xmath215 pair annihilation @xmath101 is much smaller than that due to the urca processes @xmath96 , and the electron pressure is important only in the outer region of the disk where the @xmath30-particle disintegration has not started . instead , the main effects of electron degeneracy are : ( 1 ) it increases @xmath96 greatly , so that the corresponding absorption makes a major contribution to the neutrino optical depth ( fig . 4 ) ; ( 2 ) it along with the neutronization processes cause @xmath11 to become smaller than 0.1 in the inner region of the disk ( fig . 2 ) where the neutrino cooling is dominant ( fig . 6 ) . the resulting @xmath8 is considerably reduced comparing with that in previous works where the electron degeneracy was not considered and @xmath11 was taken to be 0.5 , however it is still likely to be adequate for grbs , and its spatial distribution is likely to be anisotropic . perhaps the main limitation of our calculations here is that they are one - dimensional . as in most of previous works in the field , we do not study the vertical structure of the disk and instead use the simple one - zone approximation of the vertically - averaged model . in particular , we do not consider the distribution and transport of neutrinos in the vertical direction of the disk , a problem that remains unsolved or even rarely touched ( sawyer 2003 ) . strictly speaking , a reliable quantitative evaluation of the neutrino annihilation luminosity should require two - dimensional calculations , in which the vertical structure and neutrino transport are self - consistently included . but in view of the fact that most popular models of normal accretion disks ( the shakura - sunyaev disk model , the slim disk model , and the advection - dominated accretion flow model ) are also one - dimensional and have been proved successful , our results here may provide a plausible , though rough , estimate of the luminosity of neutrino - cooled accretion disks . beloborodov , a. m. 2003 , , 588 , 931 burrows , a. , & thompson , t. a. 2002 , astro - ph/0211404 burrows , a. , young , t. , pinto , p. , eastman , r. , & thompson , t. a. 2000 , , 539 , 865 chen , w. , & beloborodov , a. m. 2006 , , in press ( astro - ph/0607145 ) di matteo , t. , perna , r. , & narayan , r. 2002 , , 579 , 706 gu , w .- m . , liu , t. , & lu , j .- f . 2006 , , 643 , l87 hannestad , s. , & raffelt , g. 1998 , , 507 , 339 itoh , n. , adachi , t. , nakagawa , m. , kohyama , y. , & munakata , h. 1989 , , 339 , 354 ( erratum 360 , 741 [ 1990 ] ) kato , s. , fukue , j. , & mineshige , s. 1998 , black hole accretion disks ( kyoto : kyoto univ . press ) kohri , k. , & mineshige , s. 2002 , , 577 , 311 kohri k. , narayan , r. , & piran , t. 2005 , , 629 , 341 lee , w. h. , ramirez - ruiz , e. , & page , d. 2005 , , 632 , 421 mszros , p. 2002 , , 40 , 137 meyer , b. s. 1994 , , 32 , 153 paczyski , b. & wiita , p. j. 1980 , , 88 , 23 popham , r. , & narayan , r. 1995 , , 442 , 337 popham , r. , woosley , s. e. , & fryer , c. 1999 , , 518 , 356 rosswog , s. , ramirez - ruiz , e. , & davies , m. b. 2003 , , 345 , 1077 ruffert , m. , janka , h .- , & schfer , g. 1996 , , 311 , 532 ruffert , m. , janka , h .- t . , takahashi , k. , & schfer , g. 1997 , , 319 , 122 sawyer , r. f. 2003 , phys . d , 68 , 063001 shapiro , s. l. , & teukolsky , s. l. 1983 , black holes , white dwarfs and neutron stars(new york : wiley ) yuan , y .- f . 2005 , , 72 , 013007 zhang , b. , & mszros , p. 2004 , int . j. mod . phys . a , 19 , 2385	neutrino - cooled hyperaccretion disks around stellar mass black holes are plausible candidates for the central engine of gamma - ray bursts . we calculate the one - dimensional structure and the annihilation luminosity of such disks . the neutrino optical depth is of crucial importance in determining the neutrino cooling rate and is in turn dependent on the electron fraction , the free nucleon fraction , and the electron degeneracy , with given density and temperature of the disk matter . we construct a bridging formula for the electron fraction that works for various neutrino optical depths , and give exact definitions for the free proton fraction and free neutron fraction . we show that the electron degeneracy has important effects in the sense that it enlarges the absorption optical depth for neutrinos , and it along with the neutronization processes favored by high temperature cause the electron fraction to drop to be below 0.1 in the inner region of the disk . the resulting neutrino annihilation luminosity is considerably reduced comparing with that obtained in previous works where the electron degeneracy was not considered and the electron fraction was simply taken to be 0.5 , but it is still likely to be adequate for gamma - ray bursts , and it is ejected mainly from the inner region of the disk and has an anisotropic distribution .
for @xmath16 in the open unit disk @xmath17 consider the compact set @xmath18 given by @xmath19 it is the attractor for the iterated function system ( ifs ) @xmath20 , where @xmath21 , that is , @xmath10 is the unique non - empty compact set such that @xmath22 , see @xcite . the connectedness locus @xmath23 was introduced in @xcite and studied by several authors , see @xcite in particular . there is the same dichotomy as for julia sets @xmath13 of quadratic maps @xmath24 : the attractor @xmath10 is either connected or totally disconnected . it is well - known that @xmath10 is connected if and only if @xmath25 it follows that @xmath26 has a characterization as the set of zeros of @xmath27 power series : @xmath28 figure 1 shows the part of @xmath26 in @xmath29 ( the set @xmath26 is symmetric with respect to both axes , and all @xmath30 , with @xmath31 , are known to be in @xmath26 ) . bousch @xcite proved that @xmath26 is connected and locally connected . here we consider the following subset of @xmath26 : @xmath32 it is known @xcite ( and not hard to see ) that @xmath33 if and only if @xmath10 is a _ dendrite _ , that is , a connected , locally connected , nowhere dense compact set in the plane with connected complement . ( note that a connected self - similar set is necessarily locally connected , see @xcite . ) the set @xmath10 is invariant under the involution @xmath34 . ( on the symbolic level this map just switches the coefficients 0 and 1 in the power series representation . ) it follows that @xmath35 is invariant under @xmath36 , hence for @xmath33 we have @xmath37 , the single common point of @xmath38 and @xmath39 . bandt @xcite observed that it is possible to define quadratic - like dynamics on @xmath10 for @xmath33 . indeed , @xmath10 is also the attractor of the ifs @xmath40 , and these two contractions are the inverse branches of a function @xmath41 note that @xmath42 is well - defined , hence @xmath11 is continuous , it maps @xmath10 onto itself 2-to-1 , except at the single critical point @xmath43 ; at all other points it is a local homeomorphism . this is an example of what kameyama @xcite called an _ invertible ifs_. we say that the system @xmath14 is _ post - critically finite _ ( p.c.f . ) if @xmath43 is strictly preperiodic under @xmath11 . bandt @xcite observed that by results of kameyama @xcite , if @xmath14 is p.c.f . , then it is topologically conjugate with @xmath15 for a certain @xmath44 . the sets @xmath1 and @xmath5 in figure 2 and 3 are `` more similar '' than mere topological equivalence would explain . it appears visually that they are quasiconformal images of each other . in fact , the attempt to justify this visual observation was the starting point of our work and led to the following . [ th - quad1 ] suppose @xmath33 is such that @xmath14 is post - critically finite . then the conjugacy with @xmath15 is quasisymmetric and extends to a quasiconformal homeomorphism of the riemann sphere . we do not know if the same holds for all @xmath33 , without the p.c.f . assumption , but we have been able to prove it for an uncountable set @xmath45 , see ( [ def - t0 ] ) and corollary [ cor - new ] . the first ingredient of the proof is establishing the _ bounded turning property _ , defined in section [ btpi ] , closely related to the concept of a _ john domain_. we prove the bounded turning property for a large class of self - similar fractals , which includes p.c.f . @xmath10 for @xmath33 . this part does not depend on the invertibility of the ifs . our proof of quasisymmetry is similar in spirit to the proof of mcmullen and sullivan @xcite of quasiconformality of conjugacies between hyperbolic rational maps , and its extension to non - hyperbolic settings as in @xcite . see @xcite for an exposition of some of these ideas in a general setting . our proof applies not only to the sets @xmath10 , but to other examples as well , such as the sierpiski gasket and the map @xmath46 , see section 4 . in the proof we use a recent result of bandt and rao @xcite that a planar self - similar ifs with a connected attractor and a finite `` overlap set '' is non - recurrent . * examples . * the simplest example is @xmath47 and @xmath48 , which is a well - known pair of quasisymmetrically conjugate maps , since @xmath11 is just the tent - map for @xmath47 . in @xcite there are many examples of @xmath16 for which it is rigorously proven that @xmath10 is a dendrite ( including an uncountable set which we denote @xmath49 in this paper ) , and many more can be found numerically , using pictures and a version of bandt s algorithm from @xcite . in the figures below we show two examples to which our theorem applies . for @xmath33 we indicate the ( unique ) @xmath27 power series vanishing at @xmath16 , by the sequence of its coefficients , e.g. , @xmath50 ; here @xmath51 and @xmath52 correspond to @xmath53 s and parentheses indicate a period . for the julia set @xmath13 we also indicate the external parameter angle . see section 5 for an explanation of these examples . @xmath54 \mbox{self - similar set $ a_{\lambda}$ } \ \ & \ \ \mbox{julia set $ j_c$ } \end{array}$ ] @xmath55 \mbox{self - similar set $ a_{\lambda}$ } \ \ & \ \ \mbox{julia set $ j_c$ } \end{array}$ ] the rest of the paper is organized as follows . in the next section we prove the bounded turning property for the sets @xmath10 , using a `` dynamical '' argument . in section 3 we prove , in a more general setting , that if there is a conjugacy between piecewise linear and rational dynamics , then it is quasisymmetric . sections 2 and 3 yield theorem [ th - quad1 ] . section 4 is devoted to examples of rational ( non - polynomial ) maps for which our theorem applies . sections 5 and 6 deal with various extensions , in particular , to the non - recurrent case . * notation . * throughout this paper , @xmath56 and @xmath57 denotes affine maps ( similitudes ) , @xmath4 denotes a polynomial or rational map , @xmath12 , @xmath3 is a piecewise linear expanding map , and @xmath58 is the closed disk of radius @xmath59 centered at @xmath60 a connected , locally connected set @xmath61 has _ bounded turning _ if there exists @xmath62 such that for every @xmath63 there is a connected set @xmath64 \subset x$ ] satisfying @xmath65 \le l \|z_0-z_1\|.\ ] ] [ th - bt1 ] let @xmath33 , @xmath10 be given by ( [ eq - al ] ) and @xmath11 by ( [ eq - ql ] ) , and suppose that @xmath14 is post - critically finite . then @xmath10 has the bounded turning property . _ let @xmath66 , @xmath67 , @xmath68 for @xmath69 , so that @xmath70 . recall that @xmath43 is the only `` critical point , '' the intersection of @xmath71 and @xmath72 . the assumption that @xmath6 is post - critically finite means that for some @xmath73 and @xmath74 , we have @xmath75 , and @xmath76 for @xmath77 . consider @xmath78 in @xmath1 , with the goal of proving ( [ eq - bt ] ) . we can assume that @xmath79 do not lie in the same subset @xmath80 , @xmath69 . indeed , otherwise we can pass to @xmath81 , @xmath69 , using the fact that @xmath3 is an expanding similitude on @xmath80 , so the `` local geometry '' is preserved . eventually , the points will be separated by @xmath43 . so we may suppose that @xmath82 , @xmath69 , and none of the points is @xmath43 . let @xmath83 , to be determined later . we have @xmath84 in view of @xmath85 . therefore , if @xmath86 , then ( [ eq - bt ] ) holds with @xmath87= a$ ] and @xmath88 and we are done . we will reduce the general case to this case by using the self - similarity of @xmath1 at @xmath89 . let @xmath90 be the smallest integer such that @xmath91 the idea is to `` enlarge '' the picture by applying @xmath92 for a suitable branch of the inverse . in order to avoid dealing with different branches , we use the symmetry of @xmath1 about @xmath43 and consider @xmath93 we choose @xmath94 so small that @xmath95 for all @xmath96 . then @xmath97 is an expanding similitude . there are two branches of @xmath98 that map @xmath89 back to @xmath99 denote @xmath100 the one that has @xmath101 . then @xmath102 is an expanding similitude on @xmath103 which fixes @xmath43 and maps @xmath104 to points @xmath105 set @xmath106 then @xmath107 in view of ( [ eq - tuchka ] ) . hence @xmath108 by ( [ eq - eta ] ) and therefore @xmath109 here we used that @xmath110 extends as an affine linear map to the plane which fixes @xmath111 , so it commutes with the involution @xmath36 . since @xmath112 contains @xmath113 and @xmath43 , the set @xmath64 = r^{-1}(a)\cup s(r^{-1}(a))$ ] has @xmath114 \le 2\|{\lambda}\|^{pk}\diam a \leq \frac{2\diam a}{\eta(\delta ) } \|z_0-z_1\|,\ ] ] as desired . here we show that the homeomorphism @xmath115 conjugating the quadratic polynomial @xmath116 on its julia set @xmath13 to the piecewise linear map @xmath11 on @xmath10 is quasisymmetric on @xmath13 , in the setting of theorem [ th - quad1 ] . in fact , the proof works in much greater generality , so we start with a precise description of our setting . we suppose that @xmath1 is the attractor of an ifs @xmath117 in the complex plane , with @xmath118 , where @xmath119 are similarities , all having the same contraction ratio @xmath120 ( so we can write @xmath121 for some @xmath122 and @xmath123 , if @xmath119 is orientation - preserving , and @xmath124 otherwise ) . thus , @xmath1 is the unique nonempty compact set satisfying @xmath125 we further assume that the ifs @xmath117 is _ invertible _ in the sense of kameyama @xcite , that is , that there exists a continuous map @xmath126 , such that @xmath127 is the inverse of @xmath128 . this means that there is `` compatibility '' of the ifs on the _ overlap set _ @xmath129 . thus , we can consider the dynamical system @xmath6 , and @xmath0 is the critical set of @xmath3 in the sense that @xmath3 is not a local homeomorphism precisely at @xmath0 . we also assume that the overlap set is finite , which implies , by a recent result of bandt and rao @xcite , that the ifs is _ critically non - recurrent _ ( or simply non - recurrent ) . for an invertible ifs this means , by definition , that for every @xmath130 , the limit set of @xmath131 does not contain @xmath132 . being post - critically finite , of course , is a stronger property . finally , we assume that @xmath1 is connected , so that @xmath0 is non - empty . a rational map @xmath4 is called _ semi - hyperbolic _ if @xmath4 has no parabolic periodic points and if all critical points in the julia set @xmath5 are non - recurrent . [ thm - qs ] let @xmath6 be as above with finite non - empty overlap set @xmath0 , and suppose that @xmath133 conjugates @xmath4 and @xmath3 , where @xmath4 is a rational map and @xmath5 is its julia set . if @xmath1 is of bounded turning and if @xmath4 has no parabolic cycles , then @xmath115 is quasisymmetric . notice that under the assumptions of theorem [ thm - qs ] @xmath4 is semi - hyperbolic : indeed , the critical points of @xmath4 on @xmath5 are precisely those points where @xmath134 is not a local homeomorphism , hence the conjugacy @xmath115 takes critical points into critical points . intuitively , a homeomorphism is quasisymmetric if images of disks are `` roundish '' ( see the discussion below for details ) . in order to prove quasisymmetry of the conjugacy , we follow the strategy of @xcite in a non - hyperbolic setting , as in @xcite : the @xmath115 image of a small disk @xmath58 centered at @xmath135 is analyzed by first passing to the forward iterate @xmath136 for suitable @xmath137 ( such that the diameter is large ) , second applying @xmath115 ( noticing that on large scale , images of roundish sets stay roundish just by continuity ) , and finally iterating backwards using @xmath3 . that it is possible to pass from small scale to large scale with bounded distortion in the critically non - recurrent setting is due to carleson , jones and yoccoz @xcite , whereas the same fact in the linear setting is very simple . now for the details . recall ( see @xcite ) that a homeomorphism @xmath57 between metric spaces @xmath138 and @xmath139 is _ quasisymmetric _ ( qs for short ) if there is a homeomorphism @xmath140 such that for every @xmath141 and @xmath142 , @xmath143 implies @xmath144 . we will use the polish notation @xmath145 instead of @xmath146 the homeomorphism @xmath57 is called _ weakly qs _ if there is @xmath147 such that @xmath148 implies @xmath149 by a result of tukia and visl ( @xcite , @xcite chapter 10 ) , weak quasisymmetry implies quasisymmetry if @xmath150 is doubling and connected ( as subsets of @xmath151 our spaces are automatically doubling ) . recall that a space @xmath150 is of _ bounded turning _ , if there exists a constant @xmath152 such that for every @xmath153 there is a curve @xmath154\subset x$ ] satisfying @xmath155 \le l \|z_1-z_2\|.\ ] ] for @xmath156 and @xmath157 , denote the inradius of @xmath158 with respect to @xmath159 by @xmath160 where @xmath58 is the closed ball of radius @xmath59 centered at @xmath159 . we say that a set @xmath156 is @xmath161-_roundish at x _ if @xmath162 thus balls are 2-roundish ( at their center ) . [ lem - qstest ] let @xmath163 be a homeomorphism and assume that @xmath150 is of bounded turning . then @xmath57 is qs if and only if the images of balls @xmath58 under @xmath57 are uniformly roundish at @xmath164 . _ first assume that @xmath57 is @xmath165 let @xmath166 and @xmath167 then @xmath168 and hence @xmath169 so that @xmath170 by taking the supremum over @xmath132 and the infimum over @xmath171 for the converse , assume @xmath172 for all @xmath173 . if @xmath174 , then @xmath175 if @xmath150 is bounded turning , this implies weak quasisymmetry and hence quasisymmetry , as follows : given @xmath142 with @xmath176 let @xmath177 $ ] be a curve as in the definition of bounded turning . similar to the proof that weak qs implies qs ( theorem 10.19 in @xcite ) , it is easy to construct a sequence of points @xmath178 on @xmath177 $ ] with @xmath179 , such that @xmath180 and so that @xmath137 is bounded in terms of the bounded turning constant @xmath161 and the doubling constant ( a universal constant in @xmath181 ) only . then @xmath182 and summation yields @xmath183 we will need the following version of results of carleson , jones and yoccoz @xcite . [ thm - cjy ] if @xmath4 is semi - hyperbolic , then @xmath5 is of bounded turning . furthermore , if @xmath5 is not the whole sphere , then there are @xmath184 such that for every @xmath185 and every @xmath135 there is @xmath186 such that @xmath187 and such that the degree of @xmath188 on @xmath189 is bounded above by the degree of @xmath4 . in addition , we can choose @xmath190 for a given @xmath191 . here @xmath192 denotes the connected component of @xmath158 containing @xmath60 the paper @xcite deals with polynomials @xmath4 . the bounded turning property for rational functions can be found in @xcite , corollary 2 ( see the remark regarding corollary 2 in section 4 of @xcite ) . for a proof of ( [ eq - cjy ] ) in the rational case , see the proof of theorem a in @xcite ( using the notation of that proof , notice that @xmath193 , use that the sequence @xmath194 satisfies @xmath195 for some @xmath196 by @xcite , proposition 2.5 , and notice that @xmath197 is bounded from below ; it follows that for every sufficiently small @xmath59 there is @xmath137 with @xmath198 comparable to @xmath59 ) . [ lem - diam ] for a critically non - recurrent map @xmath3 there exist constants @xmath199 and @xmath200 such that if @xmath201 is connected , then @xmath202 provided either @xmath203 or @xmath204 . _ suppose that @xmath205 is connected . if @xmath206 , then @xmath207 lies in one of the @xmath80 s , hence @xmath208 . if @xmath209 and @xmath210 then @xmath211 is a single point , say @xmath132 , and @xmath212 indeed , the left - hand side follows by taking @xmath213 such that @xmath214 , and the right - hand side follows by writing , for @xmath215 : @xmath216 by ( [ eq - diam1 ] ) , @xmath217 provided @xmath218 for all @xmath219 , where @xmath220 is the number of times the iterates of @xmath207 `` hit '' the critical set . by non - recurrence of @xmath3 , there exists @xmath221 such that for every @xmath130 and every @xmath222 , @xmath223 . if @xmath224 for some @xmath130 , then @xmath225 now suppose that @xmath226 by induction on @xmath227 ( using ( [ eq - diam2 ] ) with @xmath228 ) , we conclude that each @xmath130 can belong to @xmath229 for at most one value of @xmath230 . thus ( [ eq - diam2 ] ) yields ( [ eq - diam ] ) with @xmath231 . the same argument applies when @xmath204 . the following lemma is an analog of ( [ eq - cjy ] ) in the piecewise linear setting . [ lem - roundish ] let @xmath3 be non - recurrent and assume that @xmath1 is of bounded turning , with the bounded turning constant @xmath62 . then @xmath232 whenever @xmath233 is connected , @xmath234 and @xmath235 , provided that @xmath236 , where @xmath161 and @xmath237 are from lemma [ lem - diam ] . _ from lemma [ lem - diam ] we have @xmath238 if @xmath239 is such that @xmath240 , then there is a curve @xmath241 joining @xmath159 and @xmath242 such that @xmath243 , and lemma [ lem - diam ] yields @xmath244 it follows that @xmath245 and hence @xmath246 , in particular , @xmath247 . thus @xmath248 and the lemma follows . _ proof of theorem [ thm - qs ] . _ by lemma [ lem - qstest ] , it suffices to show that images of balls are roundish . for balls of large radius this is true by continuity , so we can fix @xmath135 , @xmath249 and let @xmath137 be obtained by theorem [ thm - cjy ] . because @xmath115 is a conjugacy , we have @xmath250 for all sets @xmath158 and @xmath251 thus ( [ eq - cjy ] ) gives @xmath252 by continuity , @xmath253 is uniformly roundish at @xmath254 , and its diameter is bounded away from zero and thus uniformly comparable to the diameter of @xmath255 . we can choose @xmath94 sufficiently small , so that the @xmath256 image of any @xmath94 ball is contained in a @xmath237 ball . then lemma [ lem - roundish ] shows that the diameters of @xmath257 and @xmath258 are comparable , and that the inradius of @xmath257 is comparable to its diameter . the theorem follows . _ proof of theorem [ th - quad1 ] . _ by theorem [ th - bt1 ] , @xmath10 is of bounded turning , so the complement of @xmath10 is a john domain by ( * ? ? ? * theorem 4.5 ) . theorem [ thm - qs ] yields that the conjugacy @xmath115 is quasisymmetric . ( note that parabolic cycles are excluded by the assumption that @xmath13 is a dendrite . ) since quasisymmetric maps between boundaries of john domains extend to quasiconformal homeomorphisms of the sphere , ( * ? ? ? * theorem 4.2 ) , it follows that @xmath115 extends to a globally quasiconformal map . we present two examples ; most likely , one can find many more . \1 . let @xmath1 be the standard sierpiski gasket , the attractor of the ifs @xmath259 the first ifs is the one commonly used , but the second ifs has the advantage of being invertible . let @xmath3 be the continuous map on @xmath1 whose inverse branches produce the second ifs . it is known @xcite that @xmath6 is conjugate to @xmath7 , where @xmath260 and @xmath5 is its julia set . ( this claim is easy to verify : @xmath4 has three critical points , one of which is mapped to a fixed point , and another two are mapped into a 2-cycle ; this combinatorics agrees with that of the map @xmath3 . ) the rational map @xmath4 is semi - hyperbolic , and @xmath1 is obviously of bounded turning . thus , the conjugacy is quasisymmetric by theorem [ thm - qs ] . consider the `` hexagasket '' @xmath261 shown in figure 4 , the attractor of the ifs @xmath262 where @xmath263 . it is , however , not invertible . let @xmath264 where @xmath265 is the piece of @xmath261 containing @xmath266 . we observe that @xmath261 can also be represented as a repelling invariant set of the piecewise linear function @xmath267 the critical set consists of six points : @xmath268 , @xmath269 , @xmath270 , @xmath271 , etc . , and one easily checks that @xmath3 is well - defined on them , hence continuous . it is easy to check that @xmath3 maps the critical points as follows : @xmath272 , @xmath273 , and @xmath274 . the critical values form a 3-cycle : @xmath275 ; the remaining `` vertices '' @xmath276 are also mapped into this cycle . now consider the rational map @xmath277 whose julia set @xmath5 is also shown in figure 4 . note that @xmath4 has six critical points @xmath278 where @xmath279 are the sixth roots of @xmath280 . the critical values @xmath281 form a 3-cycle . symmetry considerations ( or a straightforward computation ) yield that @xmath4 permutes the ( straight line ) rays to infinity through the critical values , and that @xmath282 is increasing on these rays . thus the critical values are in the boundary of @xmath283 , the unbounded component of @xmath284 . a point @xmath242 in @xmath283 near a critical value , let s say near @xmath285 , has six preimages under @xmath4 . four of them are in @xmath283 and two are in @xmath286 , the component of @xmath284 containing @xmath287 . one of the four is near @xmath288 ( @xmath289 ) , and two are near the critical points that map to @xmath285 . it follows that the critical points are in the boundary of @xmath283 . the two critical points that map to @xmath285 can contribute only one preimage each ( the fourth preimage of @xmath242 in @xmath283 comes from a rotation of @xmath285 by @xmath290 ) . so the other two preimages of @xmath242 near the critical point are the two preimages in @xmath286 , and it follows that the critical points are in the boundary of @xmath286 . this proves that the critical points are `` cut - points , '' and @xmath291 is a `` misiurewicz - sierpiski '' map studied by devaney et . @xcite , whose julia set @xmath292 is a generalized sierpiski gasket . the conjugacy of @xmath293 to @xmath7 is well - defined on the critical orbits and extends by symbolic dynamics to the entire set . by theorem [ thm - qs ] , it is quasisymmetric . here we revisit the question of conjugacy between @xmath14 and @xmath15 . recall that @xmath295 for @xmath296 , and @xmath13 is its julia set . bandt ( * ? ? ? * theorem 8.1 ) claimed that if @xmath10 is a dendrite , then there exists @xmath297 such that @xmath14 and @xmath15 are topologically conjugate . however , his argument works only in the post - critically finite case . in fact , @xcite refers to a theorem of kameyama who used thurston s theorem on the topological characterization of post - critically finite rational maps . we review kameyama s approach briefly , since it is not explained in @xcite . we quote some definitions from @xcite . _ the pair @xmath298 is called a _ self - similar system _ if @xmath299 is a compact set and there exists a continuous surjection @xmath300 ( the coding map ) such that @xmath301 for all @xmath302 and for all @xmath303 . the set @xmath304 is called the _ overlap set _ , where @xmath305 . the system is _ invertible _ if @xmath306 are injective and there exists a continuous map @xmath307 such that @xmath308 for all @xmath302 . the self - similar system is _ post - critically finite _ if @xmath0 is finite and all its points are strictly preperiodic for @xmath309 . we can consider @xmath310 as a dynamical system with the critical set @xmath0 . _ we have already seen invertible ifs in a special case , at the beginning of the last section . note that `` self - similarity '' in the definition above is very different from self - similarity used in this paper ; the former is just a way to define the ifs , whereas the latter requires that the maps are similitudes . note also that kameyama uses the term `` connecting set '' instead of `` overlap set '' used here . kameyama ( * ? ? ? * th.b ) ( see also ( * ? ? ? * th.6.1 ) ) proved that if @xmath298 is an invertible post - critically finite self - similar system , with @xmath299 connected and simply connected , then @xmath311 is topologically conjugate to @xmath312 for some @xmath137 , where @xmath291 is a polynomial and @xmath313 is its julia set . moreover , @xmath314 if @xmath309 preserves the cyclic order of `` arms '' at the vertices of the hubbard tree , defined as the connected hull of the critical orbit . the vertices are defined as the branch points of the tree intersected with @xmath315 , @xmath316 . in our case , for @xmath33 , @xmath317 is an invertible self - similar system , with @xmath318 , see ( [ eq - ql ] ) . the dendrite @xmath10 is connected and simply connected , so in the post - critically finite case kameyama s theorem applies . we have @xmath314 because the action of @xmath11 on the subsets @xmath319 is linear , so the cyclic order at branch points is preserved . thus , @xmath14 is conjugate to @xmath312 for some polynomial , which has to be quadratic and can be chosen in the form @xmath320 . next we present an alternative approach to the question of conjugacy , based on symbolic dynamics . we are grateful to henk bruin for many helpful comments and suggestions which are used in this section . let @xmath321 ; denote by @xmath322 the left shift . we fix @xmath33 and let @xmath323 we have @xmath324 . the sequence @xmath325 is the symbolic `` address '' of @xmath326 , corresponding to the ifs @xmath20 . in order to determine the @xmath11-itinerary , we need to `` rewrite '' this address in terms of the ifs @xmath327 . let @xmath36 be the `` flip '' map on @xmath328 which switches @xmath287 s and @xmath329 s . from the definition ( [ eq - ql ] ) of @xmath11 it follows that @xmath330 that is , @xmath11 acts on the address as a left shift or a shift and a flip , depending on the digit @xmath331 . we assign the symbols @xmath287 , @xmath329 , and @xmath332 to @xmath333 , @xmath334 , and @xmath335 respectively , and define itineraries and the kneading sequence @xmath336 , in the same way as for @xmath116 on a connected , locally connected julia set . by ( [ eq - qact ] ) , if the orbit of @xmath132 does not hit @xmath43 , then the _ itinerary _ @xmath337 of @xmath338 satisfies @xmath339 conversely , @xmath340 or @xmath329 according to whether @xmath341 contains an even number of @xmath329 s . the _ kneading sequence _ is basically the itinerary of the critical value . however , it is customary to start it with a @xmath329 , so it is defined by @xmath342 or @xmath343 according to whether @xmath42 is in @xmath72 or @xmath71 . explicitly : @xmath344 as was already mentioned , by a result of bandt and rao @xcite , in our case ( @xmath33 , i.e.@xmath10 is a dendrite ) the system @xmath14 is critically non - recurrent , hence the orbit of @xmath43 does not return to @xmath43 and the kneading sequence is in @xmath328 . as shown in @xcite , @xmath33 if and only if there is a unique powers series @xmath345 , with @xmath346 , satisfying @xmath347 , and moreover , all @xmath348 are in @xmath349 . this power series is obtained from the address of the critical point : @xmath350 in view of the above , the sequence @xmath348 is expressed in terms of the kneading sequence as follows : @xmath351 once we have the power series @xmath352 , the parameter @xmath16 is determined ( essentially ) uniquely , since we know by @xcite that a power series with coefficients @xmath53 can have at most one zero in @xmath353 . the two power series obtained from ( [ eq - knead2 ] ) will be related by @xmath354 , so will correspond to the choice of @xmath16 or @xmath355 . since the maps @xmath11 and @xmath356 are conjugate ( the conjugacy is @xmath357 from @xmath10 to @xmath358 ) , we can always assume , without loss of generality that @xmath359 and @xmath360 . suppose that @xmath13 , the julia set of @xmath320 , is a dendrite . then there is a standard way to write @xmath361 ( a disjoint union ) , using external rays , so that @xmath362 . then the kneading sequence of @xmath116 is defined as the itinerary of the critical value @xmath297 . [ prop - conj ] let @xmath33 and @xmath296 is such that @xmath13 is a dendrite , and the kneading sequences of @xmath11 and @xmath116 are identical . then @xmath14 and @xmath15 are conjugate . _ both systems have the same `` symbolic model '' : the sequence space @xmath328 factored by the equivalence relation : @xmath363 iff there exists @xmath364 such that @xmath365 for @xmath366 and @xmath367 , @xmath368 are the two addresses of the critical point @xmath43 . we map a point in @xmath10 with a given @xmath3-itinerary into the point in @xmath13 with the same itinerary . we will show that this map , let s denote it @xmath369 , is well - defined . indeed , it is easy to see that @xmath116 is critically non - recurrent since the kneading sequence has an initial block which never occurs again . we assumed that @xmath13 is a dendrite , in particular , that it is locally connected . this implies that the conformal map from @xmath370 to @xmath371 ( where @xmath372 denotes the riemann sphere ) , conjugating @xmath373 to @xmath116 , which is used to define symbolic dynamics on @xmath13 , extends continuously to the unit circle , and hence the `` cylinder sets '' in @xmath13 , defined as sets of points with a common itinerary of length @xmath137 , shrink to a singleton , as @xmath374 . this proves that @xmath369 is well - defined and continuous . it is straightforward to check that it is a conjugacy . in view of the proposition , the following strategy seems natural : given @xmath33 , determine the kneading sequence @xmath336 of @xmath11 , and try to find the quadratic map with a dendrite @xmath13 , with the same kneading sequence . there is a connection between the kneading sequences of quadratic maps and those for the doubling map on the circle . suppose that @xmath13 is locally connected and the critical value @xmath297 is a landing point for an external ray with angle @xmath375 . then all dynamical external rays land at points of @xmath13 , see ( * ? ? ? * theorem 18.3 ) . the rays with angles @xmath376 and @xmath377 land at the critical point @xmath287 and divide @xmath378 into the `` pieces '' @xmath379 and @xmath380 , with @xmath381 . now , the itinerary of @xmath297 under @xmath116 with respect to this partition coincides with the itinerary of @xmath382 under the angle - doubling map on the circle @xmath383 with respect to the partition @xmath384 of @xmath383 , with the the arc @xmath385 corresponding to @xmath329 , and the complementary one to @xmath287 . we call this the kneading sequence of the angle - doubling corresponding to @xmath382 . a sequence in @xmath328 is called _ admissible _ if it is a kneading sequence for the angle - doubling map for some @xmath386 . it is well - known that not every sequence is admissible . there are several criteria for admissibility , see the book in progress by bruin and schleicher @xcite ( see also @xcite ) . we do not discuss them here in detail . sometimes , the following simple - minded approach works . [ lem - admis ] let @xmath387 , with @xmath388 , and suppose that @xmath389 where @xmath390 is the lexicographical order on @xmath328 . then @xmath391 is admissible . _ _ this is proved in @xcite , but we present a short direct proof . consider @xmath392 and its kneading sequence for the angle - doubling , as defined above . we claim that this is precisely the sequence @xmath391 . indeed , the iterates of @xmath382 under the doubling map have binary expansions @xmath393 for @xmath394 . if @xmath395 , then the corresponding number @xmath396 is greater than @xmath397 by ( [ eq - admis ] ) and @xmath398 ( in binary ) , and so will be assigned the digit 1 in the itinerary . if @xmath399 , then the corresponding number @xmath400 is greater than @xmath401 and @xmath402 ( in binary ) , and so will be assigned the digit 0 in the itinerary . this concludes the proof of the claim , and of the lemma . we summarize the discussion in the following statement . [ prop - flaky ] suppose that @xmath33 is such that \(i ) the kneading sequence @xmath336 of @xmath11 is admissible , and corresponds to an angle @xmath382 ; \(ii ) there is a quadratic julia set @xmath13 which is locally connected , with an external angle @xmath382 . then @xmath14 and @xmath15 are conjugate . this is almost immediate from proposition [ prop - conj ] ; we only need to note that the kneading sequence is non - recurrent , hence @xmath116 is critically non - recurrent , and then @xmath13 can only be a dendrite . [ rem - land ] _ it is folklore , although we do not know a reference , that in the above proposition ( i ) implies ( ii ) . _ [ rem - mis ] _ if @xmath11 is post - critically finite , then @xmath336 is strictly preperiodic . it is admissible , since we have conjugacy by kameyama s theorem , and the corresponding angle @xmath382 is rational , with an even denominator . in this case it is well - known @xcite that the external parameter ray of angle @xmath382 lands at a _ misiurewicz point _ @xmath297 , for which @xmath13 is a dendrite . _ our systems @xmath14 satisfy all the axioms of an abstract julia set of bruin and schleicher ( * ? ? * def.14.4 ) . the connected hull of the orbit of @xmath43 in @xmath10 is the ( abstract ) hubbard tree or hubbard dendrite according to whether the orbit is finite or infinite . in the post - critically finite case , we get a hubbard tree . it is proved in @xcite ( see also ( * ? ? ? * prop.2.11 ) ) that a hubbard tree corresponds to a quadratic polynomial if and only if the `` arms '' are cyclically permuted at periodic branch points . as already mentioned at the beginning of this section , @xmath11 has this property , and this yields another proof of the conjugacy in the post - critically finite case . using the techniques of @xcite , one can derive various properties of the dendrites from their kneading sequence . in particular , in all our examples ( see below ) , but not for general kneading sequences , the following holds [ personal communication from henk bruin ] : * all the @xmath403-fixed point of @xmath10 have three arms , and there are no other periodic branch points ; * the critical value @xmath404 is an endpoint of @xmath10 . we do not know if this holds in general . the following questions are also open : * is it true that @xmath405 is a cantor set ? * characterize those @xmath297 in the quadratic mandelbrot set for which there exists a corresponding @xmath33 . let @xmath406 be the ( only ) zero of the power series with coefficients @xmath407 in the upper half - plane of modulus less than @xmath408 ; here @xmath51 and @xmath52 correspond to @xmath53 s and parentheses indicate a period . by ( [ eq - knead ] ) , we get the kneading sequence @xmath409 . the condition ( [ eq - admis ] ) holds since @xmath410 never occurs in @xmath391 except at the beginning . note that the angle with the corresponding ( 0 - 1 flipped ) binary expansion is @xmath411 . we conclude that @xmath14 is conjugate to @xmath15 where @xmath297 corresponds to the point on the quadratic mandelbrot set with the external angle @xmath412 . see figure 2 for the picture of these sets . in figure 3 we show the dendrite for @xmath16 with the power series @xmath413 , with the kneading sequence @xmath414 , and the julia set corresponding to the external angle @xmath415 . denote @xmath416 where @xmath417 is the set of power series whose coefficients agree with @xmath407 up to degree 14 , and after degree 14 consist of arbitrary concatenations of blocks @xmath418 , @xmath419 , @xmath420 , @xmath421 , with the blocks of @xmath52 s followed by the blocks of @xmath51 s and vice versa . in @xcite it is proved that @xmath45 , that is , @xmath10 with @xmath422 are all dendrites . the family @xmath49 is clearly uncountable ( of positive hausdorff dimension ) . more examples can be found numerically . we found ( without rigorous proof , but with a high degree of confidence ) that zeros of power series ranging from @xmath423 , a zero of power series expressed by @xmath424 , to @xmath425 , a zero of power series expressed by @xmath426 , are all in @xmath427 , as well as a great many in between ( they can be surmised from figure 9(a ) in @xcite ) . in all our examples ( including @xmath422 ) the kneading sequence is admissible . this follows from lemma [ lem - admis ] since the kneading sequence will start with @xmath410 , and @xmath410 will never occur again ( @xmath410 in a kneading sequence means two sign changes on a row , that is , @xmath428 or @xmath429 ) . it may be that ( [ eq - admis ] ) is always satisfied for @xmath336 , with @xmath33 , but we do not know how to prove this . it seems that most dendrite julia sets @xmath13 do not correspond to any @xmath10 with @xmath33 . for example , consider , perhaps , the best known dendrite julia set , with @xmath430 . the critical value @xmath302 gets into the periodic 2-orbit @xmath431 after one iteration , so we have the kneading sequence @xmath432 ( the external angle is @xmath433 ) . if there is @xmath33 with such a kneading sequence , then @xmath16 is a zero of a power series with the coefficients @xmath434 . this is a rational function of degree four , whose zeros are the zeros of @xmath435 . the zeros inside the unit disk are @xmath436 , which have modulus @xmath437 . in fact , the corresponding @xmath10 is known as `` tame twindragon , '' which tiles the plane by translations and has nonempty interior , so it is definitely not a dendrite . another example is given in ( * ? ? ? * remark 8.2 ) : for the julia set with external angle @xmath438 , the only candidate @xmath16 is @xmath439 , which gives the `` twindragon '' @xmath10 , again with nonempty interior . the kneading sequence is @xmath440 . other external angles , for instance , @xmath441 or @xmath442 , yield unremarkable @xmath16 of modulus greater than @xmath437 , so again not in @xmath427 . here we prove the bounded turning property for a large class of self - similar sets , which include p.c.f . @xmath10 , not necessarily of dendrite type . we consider iterated function systems of the form @xmath443 for @xmath118 and @xmath444 . the attractor @xmath445 has an explicit representation as the set of sums of power series in @xmath16 with coefficients in @xmath446 : @xmath447 we have @xmath448 , where @xmath449 are the `` pieces '' of of the attractor . our main assumption is that @xmath445 is connected and the `` overlap set '' @xmath450 is finite . ( note , however , that here we do not assume the ifs to be invertible , as in sections 2 and 3 . ) unfortunately , we do not know if this already implies the bounded turning property , so we have to impose an additional technical assumption . we will show , however , that this assumption is satisfied in many cases . let @xmath451 it is immediate that @xmath452 let @xmath453 since @xmath445 is connected , we have that @xmath454 . [ th - bt ] assume that the self - similar set @xmath445 of the form ( [ eq - attr ] ) is connected and has a finite overlap set . in addition , suppose that there exists @xmath461 such that for all @xmath462 sufficiently large , @xmath463 then @xmath445 has bounded turning . we say that @xmath445 is p.c.f . ( post - critically finite ) if the overlap set is finite and every sequence in @xmath464 is eventually periodic . this definition is consistent with the definition of p.c.f . used earlier , but it is more general , since here we do not assume that the ifs is invertible . [ cor - new ] let @xmath33 be such that the conditions of proposition [ prop - flaky ] are fulfilled , with some @xmath44 , and the self - similar set @xmath10 satisfies ( [ eq - tech ] ) . then @xmath14 and @xmath15 are quasisymmetrically conjugate . in particular , this holds for all @xmath422 . [ rem - john ] _ by ( * ? ? ? * th.4.5 ) , the bounded turning property of a compact set @xmath445 implies that the components of the complement of @xmath445 are john domains . in the paper @xcite it is shown that there exist self - similar sets in the plane ( not of the kind considered in this paper ) which satisfy the open set condition , but the complement does not have the john property . on the other hand , @xcite obtained some geometric conditions ( which do not hold in our case ) for the complement of a self - similar set to be a uniformly john domain . the picture of a self - similar set in ( * ? ? ? * fig.4 ) suggests that it does not have the bounded turning property , even though the overlap set is finite . the ifs , however , contains both orientation - preserving and orientation - reversing maps , so it is also of a different kind from those considered in our paper . _ [ lem - vspom ] suppose that the self - similar set @xmath445 from ( [ eq - attr ] ) has a finite overlap set @xmath0 . then the set @xmath466 from ( [ eq - fk ] ) is finite , and moreover , for each power series @xmath467 , the coefficients @xmath468 have a unique representation @xmath469 for @xmath470 , for all @xmath137 sufficiently large . _ proof of the lemma . _ bandt and rao @xcite proved that an ifs in the plane whose attractor @xmath1 is connected and the overlap set is finite , satisfies the open set condition . it is well - known that this implies that no point in @xmath1 can have infinitely many distinct addresses . if @xmath471 is such that @xmath472 ( that is , @xmath467 ) and @xmath469 for @xmath470 , then @xmath473 if a coefficient @xmath474 has non - unique representation as a difference of elements in @xmath475 , say @xmath476 , with @xmath477 , then we get a pair of distinct points in @xmath0 , with a difference of @xmath478 . if there are infinitely many such @xmath137 , we get that @xmath0 is infinite , a contradiction . this proves the second claim of the lemma . in order to see that @xmath466 is finite , note that every @xmath467 yields at least one point in @xmath0 . different functions in @xmath466 yield either distinct points in @xmath0 or one point in @xmath0 with multiple addresses . it follows that @xmath466 must be finite . _ proof of theorem [ th - bt ] . _ consider @xmath479 , with the goal of proving ( [ eq - bt ] ) . we can assume that @xmath480 and @xmath481 do not have addresses starting with the same symbol , that is , there is no @xmath482 such that @xmath483 . indeed , otherwise we could pass to @xmath484 , @xmath485 , keeping in mind that @xmath486 is a similitude . this process has to stop since @xmath487 is an expansion on @xmath488 . we will further assume that @xmath480 and @xmath481 are in @xmath489 and @xmath490 respectively , and both are very close to some @xmath491 , since otherwise ( [ eq - bt ] ) trivially holds with some constant @xmath62 . indeed , if the claim is false , we have a sequence of points @xmath495 and addresses @xmath496 such that @xmath497 and no address of @xmath159 agrees with any of @xmath498 in the first @xmath299 digits . passing to a subsequence we can assume that @xmath499 , whence @xmath500 , so @xmath325 is an address of @xmath159 and @xmath498 must eventually agree with @xmath325 in the first @xmath299 digits . this contradiction proves the claim . let @xmath504 . if either @xmath480 or @xmath481 are not in @xmath505 , then ( [ eq - bt ] ) holds with some constant @xmath62 . now suppose that @xmath506 . by the claim , there are addresses @xmath507 of @xmath508 , for @xmath485 , such that @xmath507 agrees with @xmath509 in at least @xmath502 first digits , where @xmath509 is an address of @xmath159 starting with @xmath302 . ( note that @xmath510 , since otherwise both points are in @xmath80 where @xmath511 . ) we have that @xmath512 let @xmath513 , @xmath485 , be the length of the common initial segment of @xmath509 and @xmath507 . assume that @xmath514 without loss of generality . observe that @xmath515 since @xmath516 , and @xmath517 has a unique representation as a difference of elements in @xmath475 by the assumption ( ii ) . it follows that @xmath518 . now , by ( [ eq - tech ] ) , @xmath519 , @xmath520 , be the common initial part of @xmath507 and @xmath509 . we know that @xmath521 is connected ( being a similar copy of @xmath1 ) , hence @xmath522 \le \diam a_{u^{(i)}}=\|{\lambda}\|^{k_i } \diam a.\ ] ] combining this with ( [ eq1 ] ) yields @xmath523 + \diam[z_2,x ] ) \ge ( c_1/2 ) ( \diam a)^{-1 } \diam[z_1,z_2],\ ] ] which concludes the proof of ( [ eq - bt ] ) . _ proof of corollary [ cor - pcf ] . _ in the p.c.f . case we have finitely many functions @xmath467 , and each of them has an eventually periodic sequence of taylor coefficients . let @xmath502 be so large that for any @xmath524 in @xmath466 we have @xmath525 and for every @xmath467 the coefficients are periodic for @xmath503 . for any given @xmath503 there exists @xmath526 such that we have @xmath527 indeed , otherwise , passing to a subsequence we obtain @xmath528 with @xmath529 and @xmath530 , that is , @xmath531 , contradicting the choice of @xmath137 and @xmath502 . let @xmath4 be a common period for all of the functions in @xmath466 . we claim that @xmath532 satisfies ( [ eq - tech ] ) . this is immediate from periodicity : if @xmath533 , then @xmath534 = { \lambda}^{kp}\wtil{g}({\lambda})\ ] ] for some @xmath535 with @xmath536 , hence we can take @xmath537 , and the claim follows . it remains to apply theorem [ th - bt ] .	we consider linear iterated function systems ( ifs ) with a constant contraction ratio in the plane for which the `` overlap set '' @xmath0 is finite , and which are `` invertible '' on the attractor @xmath1 , in the sense that there is a continuous surjection @xmath2 whose inverse branches are the contractions of the ifs . the overlap set is the critical set in the sense that @xmath3 is not a local homeomorphism precisely at @xmath0 . we suppose also that there is a rational function @xmath4 with the julia set @xmath5 such that @xmath6 and @xmath7 are conjugate . we prove that if @xmath1 has bounded turning and @xmath4 has no parabolic cycles , then the conjugacy is quasisymmetric . this result is applied to some specific examples including an uncountable family . our main focus is on the family of ifs @xmath8 where @xmath9 is a complex parameter in the unit disk , such that its attractor @xmath10 is a dendrite , which happens whenever @xmath0 is a singleton . c. bandt observed that a simple modification of such an ifs ( without changing the attractor ) is invertible and gives rise to a quadratic - like map @xmath11 on @xmath10 . if the ifs is post - critically finite , then a result of a. kameyama shows that there is a quadratic map @xmath12 , with the julia set @xmath13 such that @xmath14 and @xmath15 are conjugate . we prove that this conjugacy is quasisymmetric and obtain partial results in the general ( not post - critically finite ) case .
icecube is a cubic - kilometer - sized neutrino detector installed in the ice at the geographic south pole between depths of 1450 m and 2450 m @xcite . it consists of an array of photon sensors , called digital optical modules ( doms ) , attached to 86 cables , called strings . detector construction started in 2005 and finished in december 2010 . neutrino observation relies on the optical detection of cherenkov radiation emitted by secondary particles produced in neutrino interactions in ice or bedrock near icecube . due to the small neutrino interaction cross - section , the kilometer - scale detector has an effective area of only for muon neutrinos of energy @xcite . as part of the optical follow - up ( ofu ) program , the icecube neutrino observatory records high - energy ( @xmath1 ) neutrino events at a rate of about @xmath2 , about 250 per day . those events are mostly ( @xmath3 ) cosmic - ray induced neutrinos from the atmosphere , referred to as atmospheric neutrinos , with about a 10% contamination of cosmic - ray induced atmospheric muons . routine neutrino analyses in icecube , which are referred to as _ offline _ analyses , are performed after a certain amount of data , e.g. one or several years , has been collected . they benefit from events being put through computationally expensive reconstructions , as well as information on detector performance that become available only days or weeks after data acquisition . in contrast , neutrino analyses running _ online _ will not have access to such information , but have the advantage of being near real - time results are available with a latency of @xmath4 . with such a short latency neutrino analysis , multi - wavelength follow - up observations can be triggered by neutrino events . these follow - up data have the potential to reveal the electromagnetic counterpart of a transient neutrino source , which might otherwise be missed and thus be unavailable for further observations . in addition , the coincident detection of neutrino and electromagnetic emission can be statistically more significant and provide more information about the physics of the source than the neutrino detection alone . another advantage of an online analysis is the prompt availability of the reconstructed neutrino dataset and thus the possibility of fast response analyses . thus , icecube s online neutrino analysis efforts have also enabled fast @xmath5-ray burst ( grb ) searches like the one following grb 130427a , published in a gcn circular @xcite . the online search for short transient neutrino sources ( order of @xmath6 ) is mostly motivated by models of neutrinos from long duration grbs @xcite and from choked jet supernovae ( sne ) @xcite . the two source classes are related : both are thought to host a jet , which is highly relativistic in case of long grbs , but only mildly relativistic in case of choked jet sne . long grb progenitors are conceived to be wolf - rayet stars that have lost their outer hydrogen and helium envelope @xcite , while choked jet sne can still have those outer layers @xcite which are important for efficient high - energy ( he ) neutrino production . the choked jet is more baryon - rich and has a much lower lorentz factor @xmath7 than the grb jet with @xmath8 . it can not penetrate the stellar envelope and remains optically thick , making it invisible in @xmath5-rays . the neutrinos produced at energies can escape nevertheless and may trigger the discovery of the sn in other channels . in particular , for a bright nearby sn , a neutrino detection would enable the acquisition of optical data during the rise of the light curve , strengthening the time correlation between the neutrino burst and the optical supernova via the improved estimate of the explosion time @xcite . mildly relativistic jets may occur in a much larger fraction of core - collapse sne than highly relativistic jets , i.e. grbs @xcite . detections of neutrinos from grbs and choked jet sne would be a remarkable discovery and could provide important insight into the supernova - grb connection and the underlying jet physics @xcite . both sources are expected to emit a short , about long burst of neutrinos @xcite either before or at the time of the @xmath5-ray burst ( if detectable ) @xcite , setting the natural time scale of the neutrino search . after recording the neutrino burst , follow - up observations can be used to identify the counterpart of the transient neutrino source . a @xmath5-ray burst can be identified either via the prompt @xmath5-ray emission lasting up to about @xmath9 @xcite or via the optical and x - ray afterglow lasting up to several hours @xcite . the latter involves instruments of limited field of view and thus requires telescope slew , but has the advantage of much better angular resolution well below an arc minute @xcite . a fast response within minutes to hours is required for a grb afterglow follow - up . a choked jet sn is found by detecting a shock breakout or a sn light curve in the follow - up images , slowly rising and then declining within weeks after the neutrino burst . following this scientific motivation , an online neutrino analysis , targeted at sn and grb afterglow detection , was installed at icecube in 2008 ( see section [ sec : ofu ] ) . in addition to the transient neutrino emission within , as discussed above , sne can be promising sources of high - energy neutrinos over longer time scales . in this paper , the class of type iin sne @xcite is explored further . these are core - collapse sne embedded in a dense circumstellar medium ( csm ) that was ejected in a pre - explosion phase . following the explosion , the sn ejecta plow through the dense csm and collisionless shocks can form and accelerate particles , which may create high - energy neutrinos . this is comparable to a sn remnant , but on a much shorter time scale of @xcite . sne iin ( `` n '' for narrow ) are spectrally characterized by the presence of strong emission lines , most notably h@xmath10 , that have a narrow component , together with blue continuum emission @xcite . the narrow component is interpreted to originate from surrounding hii regions , and the generally slow spectral evolution is due to the presence of a high - density csm @xcite . since interaction of the sn ejecta with the dense csm can lead to the conversion of a large fraction of the ejecta s kinetic energy to radiation , sne iin are on average more luminous than other sne ii @xcite . they generally fade quite slowly and some belong to the most luminous sne @xcite . there is much diversity within the subclass of sne iin @xcite , both spectroscopically and photometrically , which can be explained by a diversity of progenitor stars and mass loss histories prior to explosion @xcite . recently , there have been observations of eruptions prior to sn iin explosions associated with mass loss which explain the existence of the dense csm shells ( e.g. * ? ? ? * ) . in this paper , we report the discovery of a type iin sn in the optical observations triggered by an icecube neutrino alert from march 2012 , and analyze the available neutrino and electromagnetic observations . the sn is already at a late stage at the time of the neutrino detection , which means that the neutrino - sn connection is presumably coincidental . the paper is organized as follows : section [ sec : ofu ] introduces the optical follow - up system of icecube . section [ sec : alert ] gives details about the neutrino alert triggering the follow - up observations that led to the sn discovery . section [ sec : highenergy ] reports limits from a complementary offline neutrino search and x - ray limits . the uv and optical data that were obtained are discussed in depth in section [ sec : lowenergy ] . we finally summarize the results and give a conclusion in section [ sec : conclusion ] . in late 2008 , an online neutrino event selection was set up at icecube , looking for muons produced by charged current interactions of muon neutrinos in or near the icecube detector . the analysis is running in real - time within the limited computing resources at the south pole , capable of reconstructing and filtering the neutrinos and sending alerts to follow - up instruments with a latency of only a few minutes @xcite . the optical ( ofu ) and x - ray ( xfu ) real - time follow - up programs currently encompass three follow - up instruments : the robotic optical transient search experiment ( rotse ) @xcite , the palomar transient factory ( ptf ) @xcite and the swift satellite @xcite . these triggered observations were supplemented with a retrospective search through the pan - starrs1 3@xmath11 survey data @xcite , which is discussed further in section [ sec : lowenergy ] . in addition , there is also a real - time @xmath5-ray follow - up program ( gfu ) targeting slower transients ( time scale of weeks ) , e.g. flaring active galactic nuclei , that is sending alerts to the @xmath5-ray telescopes magic and veritas @xcite . the background of cosmic - ray induced muons from the atmosphere above the detector amounts to @xmath12 muon events per neutrino event . in a first step , it is reduced by limiting the sample to the northern hemisphere , using the earth as a muon shield and selecting only muon tracks that are reconstructed as up - going in the detector . afterwards , cuts on quality parameters similar to those in @xcite are applied to reject mis - reconstructed muon events : after an initial removal of likely noise signals , a chain of reconstructions is performed , which utilizes the spatial and temporal distribution of recorded photons on the doms . starting with a simple linear track algorithm , more advanced reconstructions are employed , seeded with the respective preceding reconstruction . the advanced reconstructions maximize a likelihood that accounts for the optical properties of the ice for photon propagation @xcite . the final reconstruction in this chain is used to select events that are up - going in the detector . at this point , because of the vast amount of atmospheric muons , the data are still dominated by ( down - going ) muons that are mis - reconstructed as up - going . to reject this remaining background , high quality events are selected , where the selection parameters are derived from the value of the maximized track likelihood and from the number and geometry of recorded signals with a detection time compatible with unscattered photon propagation . alternatively , events with a large number of total recorded photon signals are selected . the resulting analysis sample has an event rate of about @xmath2 , of which @xmath3 are atmospheric muon neutrinos that have passed through the earth and @xmath13 are contamination of cosmic - ray induced muons from the atmosphere . in the analysis sample , the median angular resolution of the neutrino direction is about for a multi - tev muon neutrino charged current interaction event , and or less for and higher energies . the angular resolution of the sample is estimated using monte carlo ( mc ) simulation . additionally , an estimator of the directional uncertainty is computed for each event , which is based on the shape of the reconstruction likelihood close to the found maximum @xcite . it is calibrated such that its median matches the mc derived angular resolution . we define the _ bulk _ of a sample as events within the central 90% of the energy distribution . the bulk of the main background contained in the analysis sample , atmospheric neutrinos , has energies between and . in contrast , the bulk of signal events from an unbroken @xmath14 power - law spectrum would have energies @xmath15 . signal neutrinos from grbs are expected to follow a spectrum similar to @xmath14 , which has cut - off energies between @xmath16 to @xmath17 @xcite . choked jet sne are predicted to have lower cut - off energies around @xcite so that the overlap with atmospheric neutrinos is presumably much larger , while iin neutrinos may have higher cut - offs of ( see section [ sec : offline ] ) . in order to suppress background from atmospheric neutrinos , a multiplet of at least two neutrinos within 100 seconds and angular separation of or less is required to trigger an alert . in addition , since 2011 mid - september , a test statistic is used , providing a single parameter for selection of the most significant alerts . it was derived as the analytic maximization of a likelihood ratio following @xcite , for the special case of a neutrino doublet with rich signal content : @xmath18 where the time between the neutrinos in the doublet is denoted as @xmath19 , and their angular separation as @xmath20 . the quantities @xmath21 and @xmath22 depend on the event - by - event directional uncertainties @xmath23 and @xmath24 of the two neutrino events , typically @xmath25 . the angle @xmath26 corresponds to the circularized angular radius of the field of view ( fov ) of the follow - up telescope . it is set to @xmath27 for swift and @xmath28 for rotse and ptf . the test statistic @xmath29 is smaller for more signal - like alerts , which have small separation @xmath20 , small time difference @xmath19 and a high chance to lie in the fov of the telescope . thus , @xmath29 is a useful parameter to separate signal and background alerts . for each follow - up program , a specific cut on @xmath29 is applied in order to send the most significant alerts to the follow - up instruments . in the 2011/12 data acquisition ( daq ) season , which is discussed here , a cut of @xmath30 was used for the rotse follow - up , while cut values of and were used for ptf and swift . multiplets of multiplicity higher than two are passed directly to all follow - up instruments . since the expected background rate is low ( @xmath31 per year ) , each observation of a triplet or higher order multiplet is significant by itself . * rotse * @xcite is a network of four optical telescopes with aperture and @xmath32 fov , located in australia , texas , namibia and turkey . since late 2012 , only the two northern hemisphere telescopes continue operation . rotse is a completely automatic and autonomous system that can receive alerts , perform observations and send resulting data without requiring human interaction . the limiting magnitude of @xmath33 is however insufficient to discover faint or far sne . for instance , for a very bright sn with absolute magnitude , the detection radius is about , while a faint sn with is only visible within a radius of . icecube has been sending @xmath34 alerts per year to rotse , since 2008 december . the first 116 alerts , with a background expectation of @xmath35 alerts , were followed up with a median latency of hours between the neutrino alert and start of the first follow - up observation . * ptf * @xcite is a survey based at the palomar observatory in california , usa . it utilizes the @xmath36 oschin schmidt telescope on mount palomar . the focal plane is equipped with a mosaic of 11 ccds with field of view of . the typical @xmath37-band limiting magnitude of ptf during dark time is about . all the ptf data are reduced using the lbnl real - time pipeline responsible for transient identification and the ipac pipeline described in @xcite . the image photometric calibration is described in @xcite . ptf pursues a number of science goals , most notably the discovery and observation of sne . several other telescopes in palomar and at other locations can be used for photometric and spectroscopic follow - up observation . icecube has been sending @xmath38 alerts per year to ptf , since august 2010 . the first 23 alerts , with a background expectation of @xmath39 alerts , were followed up with a median latency of hours between the neutrino alert and start of the first follow - up observation . * swift * @xcite is a satellite operated by nasa and boards various instruments : a ultraviolet / optical telescope ( uvot ) , a x - ray telescope ( xrt ) and a hard x - ray burst alert telescope ( bat ) . swift s main goal is the discovery and study of grbs , of which it detects about per year @xcite ( one third of all grbs ) . icecube s x - ray follow - up program triggers swift s xrt , which can provide valuable information by observing a grb afterglow in x - rays . the xrt has a fov of only @xmath40 in diameter , hence swift performs seven pointings for each icecube follow - up , resulting in an effective fov of about in diameter . icecube has been sending @xmath41 alerts per year to swift , since february 2011 . the first 18 alerts , with a background expectation of @xmath42 alerts , were followed up with a median latency of hours between the neutrino alert and start of the first follow - up observation ( see * ? ? ? * ) . on 2012 march 30 ( mjd 56016 ) , the most significant alert since initiation of the follow - up program ( significance of @xmath43 , converting the @xmath29 cumulative distribution function value to single - sided gaussian std . deviations ) was recorded and sent to rotse and ptf simultaneously . the significance was also above the threshold for swift ( @xmath44 ) , however it was within swift s moon proximity constraint , which delayed the observations by three weeks . the two neutrino events causing the alert happened on 2012 march 30 at 01:06:58 ut ( mjd 56016.046505 ) and 1.79 seconds later , with an angular separation of . the combined average neutrino direction is at right ascension and declination in j2000 with an error radius of @xmath45 . this average is a weighted arithmetic mean , weighting the individual directions with their inverse squared error , given by the event - by - event directional uncertainty ( s.a . ) . the error @xmath46 is defined after eq . [ eq : alert - llh ] . this assumes that the two neutrino events were emitted by a point source at a single fixed position . a variance of the individual true neutrino directions does not need to be taken into account , since the assumed intrinsic variance is zero in case of a point source . this leads to a relatively small error on the average direction . the main event properties are summarized in table [ tab : properties ] : the occurrence time on 2012 march 30 , the reconstructed muon energy proxy @xmath47 ( see * ? ? ? * ) , and the estimated directional error @xmath48 . the quantity @xmath47 is a fit parameter and serves as a proxy for the muon energy , however it is not an estimator of the true muon energy . the energy @xmath49 of the neutrino that produced the muon is not directly observable , since only the muon crossing the detector is accessible . however , using monte carlo ( mc ) simulated neutrino events , one can use the muon energy proxy @xmath47 to compare with mc events having a similar @xmath47 value . from those mc events , a distribution of the true muon energy @xmath49 can be derived , which depends on the assumed underlying neutrino energy spectrum . @xmath49 probability density functions are filled with true energy values of mc events that have the reconstructed muon energy proxy @xmath47 not more than 10% and the reconstructed zenith angle @xmath50 not more than 0.1 away from the observed values of the two alert events . the median and the central 90% c.l . interval are calculated from the @xmath49 probability density functions . the results are listed in table [ tab : properties ] , where the assumed neutrino spectrum is given in parentheses . cccccc 01:06:58 & 0.96 & & @xmath51 & @xmath52 & @xmath53 + 01:07:00 & 0.66 & & @xmath54 & @xmath55 & @xmath56 follow - up observations at the direction of the neutrino alert were performed with multiple instruments ( see section [ sec : obs ] ) . in the ptf images , a core - collapse supernova , named _ ptf12csy _ , was discovered at right ascension and declination ( j2000 ) , only away from the average neutrino direction , see figs . [ fig : map ] and [ fig : subtraction ] . this was a promising candidate for the source of the neutrinos , but a search of the pan - starrs1 archive ( see section [ sec : obs ] ) revealed that it was already at least 169 observer frame days old , i.e. 158 days in host galaxy rest frame , at the time of the neutrino alert . therefore , it is highly unlikely that the neutrinos were produced by a jet at the sn site , as this is expected to happen immediately after core collapse in the choked jet scenario @xcite . however , steady neutrino emission on a time scale of several months is a possibility and explored in section [ sec : offline ] . the value of the test statistic @xmath29 for the neutrino doublet amounts to @xmath57 . the background distribution of @xmath29 is constructed from experimental data , containing mostly atmospheric neutrinos , by randomly permuting ( shuffling ) the event times and calculating equatorial coordinates , i.e. right ascension and declination , from local coordinates , i.e. zenith and azimuth angle , using the new times . that way , all detector effects are entirely preserved , e.g. the distribution of the azimuth angle , which has more events at angles where detector strings are aligned , and the time distribution , which is affected by seasonal variations . at the same time , all potential correlations between the events in time and space , and thus a potential signal , are destroyed . the false alarm rate ( far ) for an alert with @xmath58 is , calculated via integration of the @xmath29 distribution below . considering the ofu live time of 220.1 days in the data acquisition season of the alert , september 2011 to may 2012 , yields @xmath59 false alerts . hence , the probability , or p - value , for one or more alerts at least as signal - like to happen by chance in this period is @xmath60 . the ofu system had already been sending alerts to ptf for @xmath61 days at the time of the alert . scaling up the number of expected alerts with @xmath62 , one derives a probability of @xmath63 during 460 days . the estimated explosion time of sn ptf12csy does not fall within the _ a priori _ defined time window for a neutrino - sn coincidence of @xmath64 . it is thus not considered an a priori detection of the follow - up program . despite this fact , for illustrative purposes , we calculate the _ a posteriori _ probability that a random core - collapse sn ( ccsn ) of any type , at any stage after explosion , is found coincidentally within the error radius of this neutrino doublet and within the luminosity distance of ptf12csy , i.e. @xmath65 . the number of such random sn detections is @xmath66 where @xmath67 is the solid angle of the doublet error circle ( blue circle in fig . [ fig : map ] ) , which is @xmath68 . for the volumetric ccsn rate @xmath69 , a value of @xmath70 is used , ( see * ? ? ? the control time @xmath71 is the average time window in which a sn is detectable , i.e. brighter than the limiting magnitude . it depends on the distance to the source @xmath72 , the peak absolute magnitude @xmath73 of the sn , the limiting magnitude @xmath74 of the telescope , and the shape of the light curve which is adopted from a sn template web page by p. e. nugent . it is assumed that @xmath73 follows a gaussian distribution with mean @xmath75 and standard deviation @xmath76 , based on @xcite . for ptf , @xmath77 is assumed . the resulting expectation value for coincidental sn detections is @xmath78 , which results in a poisson probability of @xmath79 to detect any ccsn within the neutrino alert s error radius . combining this probability with the probability of @xmath80 for the neutrino alert , fisher s method @xcite delivers a combined p - value of @xmath81 , corresponding to a significance of . for the total live time of 460 days , the combined p - value is @xmath82 which corresponds to a significance . this means , even ignoring the a posteriori nature of the p - value , a chance coincidence of the neutrino doublet and the sn detection can not be ruled out and thus we consider the sn detection to be coincidental . the following section reports about the available high - energy follow - up data . limits on a possible long - term neutrino emission from ptf12csy are set using one year of icecube data . limits on the x - ray flux were obtained using the swift satellite . section [ sec : lowenergy ] deals with the analysis of the low - energy optical and uv data as the sn detection is significant and interesting by itself . type iin supernovae , such as ptf12csy , are a promising class of high - energy transients ( see @xcite ) . the expected duration of neutrino emission from sne iin is months , hence it is extremely unlikely that two neutrinos arrive within less than , so late after the sn explosion . however , to test the possibility of a long - term emission , a search for neutrinos from ptf12csy within a search window of roughly one year is conducted . after the core - collapse of a sn iin , the supernova ejecta are crashing into massive circumstellar medium ( csm ) shells , producing a pair of shocks : a forward and a reverse shock . cosmic rays ( crs ) may be accelerated and multi - tev neutrinos produced , potentially detectable with icecube . the collisionless shocks generating the neutrinos are expected to generate x - rays as well at late times ( see e.g.@xcite ) , but no x - rays were detected for ptf12csy , likely because of the large distance to the sn . following @xcite and @xcite ( see also * ? ? ? * ) , we model high - energy neutrino emission from ptf12csy . as a simplified approach , in order to get an order - of - magnitude estimate of the expected event rate , we perform the following calculation : the csm density profile is calculated using ( * ? ? ? * eq . 4 , eq . the proton spectrum is modeled with a power law index , as in ( * ? ? ? * eq . a7 ) , with a cut - off energy given by the maximum energy of accelerated protons . the latter is determined by comparing the proton acceleration time scale either with the dynamical time scale ( * ? ? ? * eq . 28 ) , or , if @xmath83 energy losses are relevant , with the cooling time scale ( * ? ? ? * eq . 30 ) . the lower of the two maximum energies is the one that needs to be considered . the proton spectrum is normalized to the total cr energy @xmath84 by assuming that a fraction of the kinetic energy of the ejecta @xmath85 is converted into crs , that is @xmath86 ( compare * ? ? ? * eq . 3 , eq . 25 , using csm shell mass @xmath87sn ejecta mass @xmath88 ) . other model parameters are the break - out radius @xmath89 and shock velocity @xmath90 . the expected neutrino spectrum from @xmath83 interaction is derived from the semi - analytical description in @xcite , taking into account the meson production efficiency ( * ? ? ? * eq . 35 , eq . 36 ) . it is distance - scaled and folded with icecube s effective area from @xcite to obtain the expected number of events . it is found that inserting commonly assumed values @xcite of @xmath91 ( with the bolometric energy found in section [ sec : sed ] ) , @xmath92 , @xmath93 , @xmath94 , and @xmath95 , on average only icecube neutrino detections are expected . despite the low expectation for the neutrino fluence , we search for a long - term neutrino signal from ptf12csy in the icecube data . as a more elaborate approach compared to the simplified approximation described above , we test the neutrino emission models a and b given in fig . 1 of @xcite , which are two representative cases of cr accelerating scenarios : model a corresponds to a csm shell with a high density of at a small radius of , while model b is the opposite with a density of at radius . for the ejecta , a kinetic energy of , a velocity of , and a mass of several solar masses , lower than the csm mass , are assumed . model a is close to a scenario explaining superluminous sne iin such as sn 2006gy , while model b is a good description for dimmer , but longer lasting sne like sn 2008iy . both models have a neutrino energy spectrum close to @xmath96 , with a cut - off energy around for model a , around for the forward shock ( fs ) in model b , and around for the reverse shock ( rs ) in model b. in model a , only the reverse shock is of importance for cr acceleration . the suggested emission time scales are @xmath97 for model a and @xmath98 for model b @xcite . figure [ fig : limits ] shows the model fluences scaled down to a luminosity distance of . we analyze about one year of icecube data : the entire icecube 86 strings data acquisition season 2011/12 , from 2011 may 13 to 2012 may 15 . the long search window is motivated by the large uncertainty on the explosion date ( between 2011 march 21 and 2011 october 13 ) as well as the long duration of neutrino emission for some scenarios like @xmath99 days for model b in @xcite . for simplicity , we assume that the entire fluence was emitted during the search window . we use the neutrino sample of the icecube optical follow - up system and perform a statistical point source analysis of neutrino events close to the position of the supernova , based on @xcite . each neutrino candidate event @xmath100 is given both a signal and background probability @xmath101 and @xmath102 which are combined in the likelihood function @xmath103 the variable @xmath104 is the number of signal events contained in the sample , which is fitted to maximize the likelihood , and @xmath105 is the total number of selected neutrino candidate events . the signal probability @xmath101 is the product of the spatial gaussian probability density function ( pdf ) and the energy pdf @xmath106 , i.e. the probability of a signal event having reconstructed energy @xmath107 given the neutrino spectrum @xmath108 of the source ( derived from monte carlo simulation ) : @xmath109 here , @xmath110 is the event s angular error estimate , @xmath111 is the event s reconstructed direction , and @xmath112 the sn or neutrino source position . the background probability @xmath102 contains the energy pdf and a normalization constant for the background from atmospheric neutrinos . we define the test statistic as likelihood ratio @xcite @xmath113 , \label{eq : teststat}\ ] ] serving as a powerful test for separating the null hypothesis from the hypothesis of signal event contribution . here , @xmath114 is the number of signal events that maximizes the likelihood and corresponds to the most likely description of the data . the result of the maximum likelihood fit is @xmath115 both for model a and b , i.e. we see no sign of signal contribution in our neutrino event sample . we set 90% c.l . neyman upper limits ( see @xcite , reprinted in @xcite ) on the tested neutrino fluence models , which amount to @xmath116 and @xmath117 times the fluences given for model a and b above , respectively . the limits are much higher than the fluence prediction because of icecube being insensitive to sne iin at such large distances . figure [ fig : limits ] shows a plot of the tested neutrino fluence and the limits set using of icecube data . this null result and the large distance to the sn further support our conclusion that the sn detection was coincidental . nevertheless , it is interesting to roughly estimate the hypothetical emitted neutrino fluence : we take the median neutrino energies of the two alert neutrinos from table [ tab : properties ] and look up the effective areas for the ofu neutrino sample at the respective energies . from this , we derive a hypothetical neutrino fluence of for a source spectrum @xmath118 ( @xmath119 ) , or for a source spectrum @xmath120 ( @xmath121 ) . assuming that this neutrino fluence was emitted by the sn , this would imply a radiated neutrino energy of @xmath122 or @xmath123 using the luminosity distance of @xmath124 , corresponding to about 15 or about 50 times the radiated electromagnetic energy of @xmath125 ( see section [ sec : sed ] ) . this is higher than what can be expected , since with reasonable assumptions that the explosion energy @xmath126 @xcite and a fraction @xmath127 of it going into crs @xcite , the energy in neutrinos should be on the same order or less than @xmath128 . thus , also with a simple energetic argument , isotropic neutrino emission from ptf12csy causing the neutrino alert is implausible , especially on a time scale of seconds . however , a beamed emission from a jet with a small opening angle of @xmath129 would in principle be possible . the swift satellite observed the supernova four times , on 2012 april 20 ( mjd 56037 ) and around 2012 november 15 ( mjd 56246 ) ( see table [ tab : xray ] ) . we perform source detection using the software developed for the 1sxps catalog @xcite on each of the four observations , and on a summed image made by combining all the datasets . no counterpart to ptf12csy is detected . upper limits are generated for each of these images , following @xcite . a radius circle centered on the optical position of ptf12csy is used to measure the detected x - ray counts @xmath130 at this location , and the expected number of background counts @xmath131 , predicted by the background map created in the source detection process . we then use the bayesian method of @xcite to calculate the upper limit @xmath132 on the x - ray count rate of ptf12csy , using the xrt exposure map to correct for any flux losses due to bad pixels on the xrt detector , and the finite size of the circular region . the upper limit count rate is converted to unabsorbed flux @xmath133 using the heasarc tool webpimms , assuming a black body model with @xmath134 as in @xcite , a galactic hydrogen column density of @xcite and a redshift of @xmath0 . the result is a x - ray flux @xmath135 for the most constraining upper limits , corresponding to a x - ray luminosity of @xmath136 with a luminosity distance of about . using a power - law @xmath118 instead of a black body as an alternative x - ray emission model , the unabsorbed flux upper limit becomes @xmath137 , and hence , @xmath138 . cccccc 56037.15 & 4.9 & 1 & 1.47 & 1.3 & + 56245.29 & 2.0 & 1 & 0.62 & 2.1 & + 56246.04 & 1.2 & 2 & 0.44 & 9.8 & + 56247.62 & 5.0 & 2 & 1.35 & 2.1 & + sum & 13.0 & 6 & 3.71 & 1.3 & comparing with other sne iin , e.g. sn 2008iy @xcite which had a measured x - ray luminosity of @xmath139 or sn 2010jl @xcite with @xmath140 , we can not exclude x - ray emission from ptf12csy with our measured upper limit . however , @xcite suggest that @xmath141 be about @xmath142 of the bolometric luminosity at the time of the shock breakout . with the estimated bolometric luminosity from section [ sec : sed ] around the time of the first swift observations , this implies @xmath143 , well below our x - ray limits . during the follow - up program of the neutrino alert , the first observations were done on 2012 april 03 , 05 , 07 and 09 ( mjd 56020 to 56026 ) by ptf with the palomar samuel oschin 48-inch telescope ( p48 ) @xcite , which is a wide - field schmidt telescope . the images ( see fig . [ fig : subtraction ] ) revealed a so far undiscovered supernova , named _ ptf12csy _ , at a magnitude of @xmath144 in the mould @xmath37-band . more photometric observations were carried out , with the p48 and the palomar 60-inch ( p60 ) telescopes @xcite at the palomar observatory in california , and the faulkes telescope north ( ftn ) @xcite at haleakala on maui , hawaii . spectroscopy was taken as well , with the gemini north multi - object spectrograph ( gmos ) @xcite on the gemini north telescope ( mauna kea , hawaii ) on 2012 april 17 ( mjd 56034 ) and with the low - resolution imaging spectrometer ( lris ) @xcite on the keck i telescope ( kamuela , hawaii ) on 2013 february 09 ( mjd 56332 ) , enabling the identification of the supernova as a type iin sn with narrow emission lines . the spectra are available from wiserep @xcite . p48 data were extracted using an aperture photometry pipeline and are calibrated with 21 close - by sdss stars @xcite . the faint host galaxy was subtracted and the upper limits are at the level . p48 magnitudes are in the ptf natural ab magnitude system , which is similar , but not identical to the sdss system . the difference is given by a color term , which is ignored in this work , except for the conversion of mould @xmath37 to sdss @xmath72 , explained in section [ sec : corr ] . the p60 photometry is tied to the same 21 sdss calibration stars . note that there might be host galaxy contamination in the late - epoch p60 photometry . p60 magnitudes are in the sdss ab magnitude system . the ftn data were processed by an automatic pipeline , without host subtraction , and agree very well with the host - subtracted p60 data taken in the same night . the pan - starrs1 ( ps1 ) telescope was not part of the real - time triggering and response system , but its wide - field coverage provides a useful archive to search retrospectively for detections . ps1 is a @xmath145 telescope located at haleakala on maui in the hawaiian islands , equipped with a fov and a 1.4 gigapixel camera @xcite . in the course of its 3@xmath11 steradian survey , the telescope observes each part of the sky typically times per year @xcite . ps1 first detected ptf12csy on mjd and archived it as object _ pso j104.6365 + 17.2622_. the magnitudes in all ps1 images were obtained with psf fitting within the pan - starrs image processing pipeline @xcite . they are calibrated to typically seven local sdss dr8 field stars . the magnitudes are in the natural ps1 ab system as defined in @xcite , which is similar , but not exactly the same as sdss ab magnitudes . particularly the @xmath146-band can differ . the swift uvot data were analyzed using the publicly available swift analysis tools @xcite , and a source was seen close to the detection threshold . rotse s limiting magnitude of about prevented a detection of the sn in rotse follow - up observations . the photometry is corrected for galactic extinction using @xmath147 and @xmath148 @xcite . the extinction coefficient is converted to the filters effective wavelengths using the algorithm from @xcite , i.e. for @xmath149 , for @xmath150 , for @xmath146 , for @xmath72 , for @xmath151 . the extinction within the host galaxy could not be determined . in fig . [ fig : spectrum_and_filters ] , the gemini north spectrum is overlaid with the applied photometric filters . the strong balmer lines contribute differently to the various filters . for the sed construction ( see section [ sec : sed ] ) , in order to approximate the black body continuum , the contribution of the strongest emission lines , h@xmath10 and h@xmath152 , is removed from the photometry using the gemini north spectrum and the filter curves . for figs . [ fig : lc_abs ] and [ fig : decline ] , the p48 mould @xmath37 magnitudes are converted to sdss @xmath72 by subtracting the h@xmath10 contribution ( as above ) , applying the formulae in @xcite valid for black body spectra , and then re - adding the h@xmath10 contribution to the @xmath72-band . after conversion , the p48 @xmath37 magnitudes are consistent with the p60 sdss @xmath72 magnitudes . the swift uvot data contain host contamination . since no galex data from a pre- or post - sn epoch are available for the host galaxy , no host subtraction can be done in the uv filters of uvot . for the @xmath149 , @xmath153 and @xmath154 filters , the host is subtracted by interpolating the host magnitudes from the sdss dr12 data @xcite to the effective wavelengths of the uvot filters . the earliest detection of ptf12csy was in the pan - starrs1 y - band on 2011 october 13 ( mjd ) , 169 days prior to the neutrino alert in observer frame , corresponding to 158 days in host galaxy rest frame using @xmath0 ( see section [ sec : spec ] ) . the latest non - detection , again in pan - starrs1 , was on 2011 march 21 ( mjd 55641.3 ) in a @xmath155 @xmath151-band frame , 206 days before the first detection ( 193 days in rest frame ) . hence , the explosion time is not well constrained and can be anytime between mjd 55641.3 and mjd . hereafter , we refer to the @xmath156-band detection at mjd as the first detection and use it as day 0 for the light curve . the uncorrected sn light curves with the data available through the icecube optical follow - up program are displayed in fig . [ fig : lc_app ] , including photometry acquired with the swift uvot filters @xmath157 , @xmath158 , @xmath159 , @xmath149 and @xmath153 ; the johnson @xmath150 filter on p60 ; the sdss filters @xmath146 , @xmath72 , @xmath100 with data from p60 , ps1 , and ftn ; the sdss @xmath151 filter on p60 ; mould @xmath37 filter on p48 ; and pan - starrs @xmath156 filter on ps1 . the entire uncorrected photometry in apparent magnitudes , as seen in fig . [ fig : lc_app ] , is also available in table [ tab : lc ] . the light curves are averaged within bins of 10 days width , for each filter and telescope separately . note that , in contrast to most of the other photometry , no host subtraction is performed for the swift uvot magnitudes presented in fig . [ fig : lc_app ] and table [ tab : lc ] . ccccccc & & & & & @xmath146 & p48 + & & & & & @xmath37 & p48 + & & & & & @xmath37 & p48 + & & & & & @xmath37 & p48 + & & & & & @xmath146 & p48 + & & & & & @xmath151 & ps1 + & & & & & @xmath156 & ps1 + & & & & & @xmath100 & ps1 + & & & & & @xmath151 & ps1 + & & & & & @xmath146 & ps1 + & & & & & @xmath72 & ps1 + & & & & & @xmath146 & ps1 + & & & & & @xmath100 & ps1 + & & & & & @xmath156 & ps1 + & & & & & @xmath37 & p48 + & & & & & @xmath151 & ps1 + & & & & & @xmath72 & p60 + & & & & & @xmath151 & p60 + & & & & & @xmath37 & p48 + & & & & & @xmath100 & p60 + & & & & & @xmath150 & p60 + & & & & & @xmath146 & p60 + & & & & & @xmath100 & ftn + & & & & & @xmath72 & ftn + & & & & & @xmath146 & ftn + & & & & & @xmath154 & uvot + & & & & & @xmath149 & uvot + & & & & & @xmath158 & uvot + & & & & & @xmath157 & uvot + & & & & & @xmath159 & uvot + & & & & & @xmath153 & uvot + & & & & & @xmath100 & p60 + & & & & & @xmath72 & p60 + & & & & & @xmath150 & p60 + & & & & & @xmath146 & p60 + & & & & & @xmath100 & ftn + & & & & & @xmath37 & p48 + & & & & & @xmath100 & p60 + & & & & & @xmath72 & p60 + & & & & & @xmath100 & p60 + & & & & & @xmath146 & p60 + & & & & & @xmath146 & p60 + & & & & & @xmath72 & p60 + & & & & & @xmath100 & p60 + & & & & & @xmath72 & p60 + & & & & & @xmath146 & p60 + & & & & & @xmath100 & p60 + & & & & & @xmath72 & p60 + & & & & & @xmath72 & p60 + & & & & & @xmath100 & p60 + & & & & & @xmath146 & p60 + & & & & & @xmath146 & p60 + & & & & & @xmath72 & p60 + & & & & & @xmath100 & p60 + & & & & & @xmath100 & p60 + & & & & & @xmath157 & uvot + & & & & & @xmath159 & uvot + & & & & & @xmath158 & uvot + figure [ fig : lc_abs ] shows the light curve of selected filters in absolute magnitudes , after the photometric corrections . light curves of other exceptional type ii sne are overlaid for comparison : sn iin 2006gy @xcite , one of the most luminous sne ever recorded , and sn 2010jl @xcite , a sn iin that is spectroscopically similar to ptf12csy ( see section [ sec : spec ] ) and shows signs of a collisionless shock in an optically thick circumstellar medium ( csm ) , hinting towards potential high - energy neutrino production @xcite . note that the sn 2010jl light curve is not extinction corrected and the comparison light curves have different reference dates : sn 2010jl is relative to maximum light , 2006gy is relative to the explosion time . a theoretical light curve from pure radioactive decay of @xmath160 ( black dashed line ) is added to the figure as well , scaled to match the observed absolute magnitude of ptf12csy . the brightest observed absolute magnitudes after application of photometric corrections ( see section [ sec : corr ] ) and converting to absolute magnitudes with a distance modulus of @xmath161 ( @xmath0 ) are @xmath162 , @xmath163 , @xmath164 , @xmath165 , and @xmath166 , assuming standard cosmology with hubble parameter @xmath167 , matter density @xmath168 , and dark energy density @xmath169 . while these are lower limits to the peak magnitude due to the sparse sampling , these absolute magnitudes are relatively modest compared to the most luminous sne iin , e.g. sn 2006gy ( @xmath170 ) @xcite or sn 2008fz ( @xmath171 ) @xcite . they are however comparable to the sne iin 2008iy ( @xmath172 ) @xcite , 1988z ( @xmath173 ) @xcite and sn 2010jl ( @xmath174 ) @xcite . the light curves of ptf12csy indicate a plateau within @xmath175 after first detection , and a slow fading afterwards . the corrected absolute magnitude light curves are fitted to obtain the linear decline rates in different photometric filters , during different epochs ( see fig . [ fig : decline ] and table [ tab : decline ] ) . for some epochs and filters , especially @xmath146 and @xmath72 , the decline rates are close to , the decline rate expected for radioactive @xmath176co decay @xcite , while in general decline rates are slower , indicating that at least part of the radiated energy is powered by interaction of the sn ejecta with a dense csm @xcite . cccc @xmath157 & & & + @xmath146 & & & + @xmath72 & & & + @xmath100 & & & + @xmath151 & & & + @xmath156 & & & additionally , radioactive decay of @xmath176co at a still relatively high absolute magnitude of @xmath177 implies a preceding @xmath176ni decay with an extremely bright peak , which was not observed , although the data are quite sparse . assuming that the luminosity is generated by radioactive decay alone and following @xcite , we estimate that @xmath178 of @xmath176ni would be required to provide the bolometric luminosity of at in rest frame ( see section [ sec : sed ] ) . the lower limit on the @xmath176ni mass is set by assuming that the explosion and thus generation of @xmath176ni was at the latest possible time , directly before the first detection . figure [ fig : lc_abs ] shows the corresponding theoretical light curve resulting from the radioactive decay of nickel and cobalt ( black dashed line ) . adopting an earlier explosion time results in an even larger @xmath176ni mass . this is much more than the usual amount of @xmath176ni of , often ( see e.g.@xcite , also @xcite ) . however , extremely superluminous sne might have @xmath176ni masses of that order of magnitude @xcite . we note that in addition to the @xmath176ni mass and luminosity arguments , the spectrum showing intermediate width balmer lines and a continuum appears inconsistent with radioactive decay as well ( see section [ sec : spec ] ) . here we assume that the light curve is powered by conversion of the ejecta s kinetic energy to luminosity through interaction of the ejecta with the csm . following ( * ? ? ? * see also : e.g. @xcite ) , we model the light curve as a power - law of the form @xmath179 . after shock breakout , there is a phase of power - law decline of the luminosity , with an index of typically @xmath180 . this lasts until the shock runs over a csm mass equivalent to the ejecta mass and the shock enters a new phase of either conservation of energy if the density is low enough and the gas can not cool quickly ( the sedov - taylor phase ) , or conservation of momentum if the gas radiates its energy via fast cooling ( the snow - plow phase ) . during the late stage , the light curve will be declining more steeply , in both cases @xcite . since we assume ptf12csy to be powered by interaction , we try to fit the interaction model from @xcite to the light curve data with the least - squares method . we perform the fit within the range of rest frame days , starting at the first r - band detection , and use the r - band light curve scaled with the bolometric luminosity from section [ sec : sed ] . it is found that the power - law index @xmath10 needs to be significantly steeper than in order to reasonably describe the data . it lies in the range of , using the constraint on the explosion time ( see section [ sec : lc ] ) for the temporal zero point of the power - law . the best fit is at @xmath181 , with the explosion time at the lowest allowed value which is the date of the last non - detection . this suggests a very steep csm density profile @xmath182 ( see * ? ? ? * eq . 12 ) , compared to the profile @xmath183 resulting from a wind with steady mass loss . however , the self - similar solutions of the hydrodynamical equations @xcite that are used in ( * ? ? ? * eq . 12 ) are invalid if the csm density profile is steeper than @xmath184 . but nevertheless , as discussed for the late - time light curve in @xcite , probably the profile is steeper than @xmath184 . this leaves us with several possible explanations : 1 . already between rest frame days 93 and 200 , the sn was in the late , e.g. snow - plow , phase . this is consistent with sn 2010jl , where the late - time light curve also shows a power - law index @xmath185 @xcite . assuming that the break in the light curve between power - law phase and late phase occurred just before the first r - band detection at day 93 , and comparing with sn 2010jl @xcite , then this means that the sn was likely already a few hundred days old , and the r - band maximum was about brighter than the observed one ( cf . * fig . 1 ) , consistent with sn 2010jl s r - band maximum . it follows that the power - law phase ended @xmath186 rest frame days after explosion , from which we can derive a swept csm mass of @xmath187 , using ( * ? ? ? * eq . 22 ) and adopting the standard values given for sn 2010jl . the sn is powered by ejecta - csm interaction , but its light curve is declining steeper than a @xmath188 power - law . this is possible , e.g. if spherical symmetry , assumed in @xcite , is broken , if the optical depth is lower than in sn 2010jl or if the csm density profile falls steeper than @xmath189 ( s.a . ) . the sn is not powered by interaction , but by radioactive decay , leading to an exponential light curve decline . however , this appears unlikely , as noted above in section [ sec : decline ] . since the spectra are only roughly calibrated , the spectral energy distribution ( sed ) is approximated from photometric data . for the highest spectral range and number of observations , a window of 10 observer frame days around day 189 , from day 184 to 194 , is used to select data ( day 172.2 to day 181.6 in rest frame ) . photometric corrections are applied , e.g. h@xmath10 and h@xmath152 removed ( see section [ sec : corr ] ) . the data are plotted in fig . [ fig : sed ] as function of the filters effective wavelengths . a black body spectrum is assumed to describe the sed and is fitted to the data . for each filter , a model data point corresponding to the black body spectrum is calculated via integration of the black body spectrum and the filter function following the sdss definition of ab magnitude in ( * ? ? ? * eq . 7 ) . a @xmath190 fit minimizes the difference between the model data and the measured data . the fit results in a reduced @xmath191 and delivers estimates for both the rest frame temperature @xmath192 and the absolute bolometric luminosity @xmath193 of the photosphere emitting the black body radiation : @xmath194 and @xmath195 , where the errors correspond to . applying the stefan - boltzmann law , we can calculate the radius of the black body photosphere from the bolometric luminosity . we estimate it to be @xmath196 . finally , to obtain an estimate on the total radiated energy , the lines contributions to the luminosity have to be added to the continuum luminosity . using the gemini north spectrum , the contribution of the h@xmath10 and h@xmath152 line to the total luminosity is computed and added to the continuum luminosity from the black body fit . this results in an estimated total radiated luminosity of at day 189 in the observer frame , i.e. day 177 in the rest frame . we use the fitted shape of the @xmath100-band light curve ( cf . section [ sec : phot ] , fig . [ fig : decline ] , table [ tab : decline ] ) to extrapolate this value and find a total radiated luminosity of at ( rest frame ) , as used in section [ sec : decline ] , and a total energy of @xmath197 radiated within 400 rest frame days after first detection , comparable to sn 2008iy which had ) @xcite and sn 2010jl with @xcite . this is a lower limit on the total radiated energy , since we lack photometric data between explosion and first detection and do not extrapolate before the first detection . additionally , as discussed below , we are neglecting a possible contribution of x - ray and @xmath5-ray emission to the total radiated energy , which is not considered by the black body spectrum based on the uv and optical data . we recommend to treat these results with caution , since @xcite pointed out that at late times the fraction of energy released from sne iin in x - rays can increase , causing the optical spectrum to deviate from a black body as fewer photons are available in the optical . this can lead to an effective decrease of the estimated photospheric radius . in this context , our estimates of @xmath198 , @xmath199 , and @xmath128 must be treated as lower limits . unfortunately , the x - ray flux from ptf12csy was not detected ( see section [ sec : x - ray ] ) . two spectra were acquired ( table [ tab : speclog ] , fig . [ fig : spectra ] ) . they are dominated by narrow emission lines , characteristic for type iin sne , with a very weak blue continuum , which indicates the old age of the sn . no continuum is visible in the late spectrum . the sn emission lines are primarily hydrogen , the balmer series is visible from h@xmath10 up to h@xmath200 . the oxygen lines oi @xmath201 , oii @xmath202 with fwhm @xmath203 , and oiii @xmath204 with fwhm @xmath205 are very narrow and were most likely produced by circumstellar gas released by the progenitor prior to explosion and then photoionized by uv radiation @xcite . ccccc 56034 & 11 & 175 & 80 & gemini north gmos + 56332 & 290 & 454 & 100 & keck i lris figure [ fig : halpha ] shows a close - up on the h@xmath10 line from both spectra , plotted vs. doppler velocity relative to the rest frame line center . in the early spectrum , the h@xmath10 line peaks at the line center and the line is composed of a narrow , intermediate and broad component with fwhm of @xmath206 , @xmath207 , and @xmath208 respectively , found by fitting a superposition of three gaussian functions to the h@xmath10 profile . this is similar to other sne iin , e.g. sn 1988z and 2008iy @xcite . the h@xmath10 profile with broad and intermediate - width component can be explained as a result of the interaction of the sn ejecta with a two - component wind @xcite . in this model , the broad line component is emitted from shocked sn ejecta expanding in a relatively rarefied wind , while the intermediate component arises from a shocked dense part of the wind , which can either consist of dense clumps or be a dense equatorial wind @xcite . this is an indication of ejecta - csm interaction . the intermediate and broad component of the early spectrum s h@xmath10 line are blueshifted which may indicate formation of dust , as explored by @xcite for sn 2010jl . a more recent study by @xcite , the most comprehensive work on the sn 2010jl emission line blueshifts to date , finds very strong evidence for a wavelength dependence of the blueshift . therefore , the authors conclude that the origin of the blueshifts is most likely the rapid formation of large dust grains , confirming @xcite and having implications on the origin of dust in galaxies . alternatively , @xcite explain the line blueshift in sn 2010jl with a bulk velocity of the emitting gas towards the observer . this is more consistent with observations if the spectral lines are symmetric about a center and if there is no wavelength dependence of the blueshift . the bulk velocity is believed to be the result of radiative acceleration of the gas by flux from the sn . presumably , there are also other possible explanations for the blueshift , e.g. the geometry or density structure of the csm . in case of sn ptf12csy , spectral line blueshift is only clearly visible in the h@xmath10 line , prohibiting the interpretation of the blueshift in favor of any scenario . the late spectrum s h@xmath10 line has a much more complicated structure than the early spectrum s . it does not peak at zero velocity anymore , but the peak is blueshifted and there are many sub - peaks . again , the blueshift might be connected to dust formation or radiative gas acceleration , but other reasons are conceivable as well . at least part of the late h@xmath10 line s complex appearance might be due to superposition of other spectral lines , e.g. nii . apart from that , it is an indication of an inhomogeneous , maybe clumpy , csm structure , and perhaps asymmetric sn explosion . the spectra are compared to template spectra from the padova - asiago supernova archive ( asa ) @xcite using the online tool gelato . the algorithm @xcite divides a spectrum into 11 relevant bins and averages within the bins to classify and compare with the archived spectra . the ptf12csy spectra are de - reddened with @xmath209 and compared to other type ii sne . gelato returns the best 30 matching spectra together with their phases , ordered by quality of fit . for both the gemini north spectrum taken at and the keck i spectrum from , the majority of best matching template spectra come from sn 2010jl . the mean phase of the matching spectra is significantly higher for the keck i spectrum : , versus for the gemini n spectrum . however , the reference dates for the spectra are mostly the discovery date , only rarely the date of maximum light or explosion date . the galaxy hosting ptf12csy is a faint dwarf galaxy designated sdss j065832.82 + 171541.6 , barely visible in the sdss dr12 images . since it has no cataloged redshift , we measure its redshift via oii @xmath202 and oiii @xmath210 lines in the keck i spectrum . the redshift is , corresponding to a luminosity distance of @xmath211 assuming standard cosmology with hubble parameter @xmath212 , matter density @xmath168 , and dark energy density @xmath169 . adopting the luminosity distance from above , the host galaxy has absolute magnitudes of @xmath213 , @xmath214 and @xmath215(corrected for galactic extinction ) . using the luminosity - metallicity relation from ( * ? ? ? * eq . 1 ) , we find a metallicity of @xmath216 , indicating that the host galaxy is quite metal - poor . overluminous sne iin , such as ptf12csy , have been preferentially found to occur in subluminous , low - metallicity galaxies @xcite , such as the host of ptf12csy . this is a trend , which is also observed for long grbs @xcite . statistics are still low and @xcite cautioned that there could be some selection bias due to intrinsically bright sne in faint host galaxies being more easily discovered during surveys doing aperture photometry . however , new surveys performing image subtraction and observing large untargeted fields , e.g. ptf and pan - starrs , provide increasing evidence for this trend , as most of the discovered bright objects would have been luminous enough to be detected in bright galaxies and in searches that are targeted to bright galaxies @xcite . ptf12csy , probably discovered by coincidence in an unbiased way , confirms this emerging trend as well . the sdss dr12has the host galaxy s cataloged center position about away from the found sn position ( see also fig . [ fig : subtraction ] ) . with an apparent diameter of about , corresponding to @xmath217 , this is quite far from the center of the galaxy , i.e.about off - center . the highest significance alert from the icecube optical follow - up program led to a coincidental discovery of the interesting and unusual type iin sn ptf12csy , which was already at least 169 days old . the combined _ a posteriori _ significance of the neutrino doublet alert and the coincident detection of any core - collapse sn within the error radius of the neutrino events ( ) and within the luminosity distance of the sn ( ) is , for the time interval of the icecube data acquisition season 2011/12 . ptf12csy is rare and unusual : with peak absolute magnitudes of @xmath218 , perhaps about -20 , it belongs to the most luminous sne . the sn is most likely powered by interaction of the ejecta with a dense circumstellar medium ( csm ) . the spectrum indicates a complicated structure of the csm . its host galaxy is a faint and metal - poor dwarf galaxy , confirming an observed trend for luminous sne iin . ptf12csy is similar in photometry and spectroscopy to other rare luminous sne iin , e.g. sne 2008iy and 2010jl . the total radiated energy is within the first 400 rest frame days after detection . given the ejecta - csm interaction , high - energy ( he ) cosmic ray production and neutrino emission may be expected on a time scale of , according to @xcite and @xcite . however , the sn is too far away for icecube to detect this emission . a complementary neutrino analysis performed offline , using one year of icecube data which cover most of the optical sn fluence , does not reveal a signal - like accumulation of neutrino events from the sn s position , leading to a very high upper limit of more than 1000 times the tested model fluence , owing to the large distance . due to the long delay of several months between explosion date and neutrinos , the doublet of neutrinos within less than two seconds can not be explained by the formation of a jet shortly after core - collapse according to the choked jet model @xcite . nor can it be explained by the expected he neutrino production from ejecta - csm interaction of sne iin ( s.a . ) . we therefore conclude that the only reasonable explanation is that the sn detection was coincidental and the neutrino doublet was produced by uncorrelated background events of atmospheric neutrinos and/or mis - reconstructed atmospheric muons . however , if there was a delaying mechanism at work , such as the spin - down of a supramassive neutron star that delays its collapse to a black hole , as in the blitzar model @xcite , then the neutrino doublet may have originated from a jet forming in the course of this delayed collapse few hundred days after sn explosion . the coincidental detection of a type iin sn following an icecube neutrino alert demonstrates the capability of the follow - up system to reveal transient he neutrino sources . an advantage of the follow - up paradigm is the prompt availability of multi - messenger information for the identification of the source , as well as the mere statistical significance of a coincidence between a neutrino burst and an electromagnetic transient detection . in this case , the significance is very low due to the delay of several months between explosion and neutrinos . however , this coincidence motivates the continuation of the follow - up program as well as further stacked neutrino analyses of type iin supernovae . fisher , r. a. 1990 , statistical methods , experimental design , and scientific inference : a re - issue of statistical methods for research workers , the design of experiments , and statistical methods and scientific inference ( oxford england new york : oxford university press ) , b. , sapir , n. , & waxman , e. 2012 , in http://dx.doi.org/10.1017/s174392131201304x [ ] , http://adsabs.harvard.edu/abs/2012iaus..279..274k[vol . 279 , death of massive stars : supernovae and gamma - ray bursts , 274281 ]	the icecube neutrino observatory pursues a follow - up program selecting interesting neutrino events in real - time and issuing alerts for electromagnetic follow - up observations . in march 2012 , the most significant neutrino alert during the first three years of operation was issued by icecube . in the follow - up observations performed by the palomar transient factory ( ptf ) , a type iin supernova ( sn ) ptf12csy was found away from the neutrino alert direction , with an error radius of . it has a redshift of @xmath0 , corresponding to a luminosity distance of about and the pan - starrs1 survey shows that its explosion time was at least 158 days ( in host galaxy rest frame ) before the neutrino alert , so that a causal connection is unlikely . the _ a posteriori _ significance of the chance detection of both the neutrinos and the sn at any epoch is within icecube s 2011/12 data acquisition season . also , a complementary neutrino analysis reveals no long - term signal over the course of one year . therefore , we consider the sn detection coincidental and the neutrinos uncorrelated to the sn . however , the sn is unusual and interesting by itself : it is luminous and energetic , bearing strong resemblance to the sn iin 2010jl , and shows signs of interaction of the sn ejecta with a dense circumstellar medium . high - energy neutrino emission is expected in models of diffusive shock acceleration , but at a low , non - detectable level for this specific sn . in this paper , we describe the sn ptf12csy and present both the neutrino and electromagnetic data , as well as their analysis .
the presence of compact star - forming regions in the tidal tails of colliding galaxies is commonly observed . the formation , in that environment , of super - star clusters with masses up to those of globular clusters has also often been reported ( e.g. , knierman et al . , 2003 ) . the existence of even more massive tidal objects , with global properties characteristics of dwarf galaxies , has been claimed for more than a decade ( see the review by duc & mirabel , 1999 ) . but whether such tidal dwarf galaxies " are genuine galaxies i.e. they are gravitationally bound entities is still strongly debated ( see in this volume the contribution by hibbard & barnes and that of amram et al . and braine et al . for an alternative view ) . but perhaps the more important issue of whether tdgs survive , escape from their parent galaxies , and significantly contribute to the overall population of dwarf galaxies , is a matter of speculation . indeed , a more plausible fate is their tidal destruction or their falling back onto their progenitors . there are two ways to tackle the problem : using numerical simulations of interacting galaxies and following in them the buildup and evolution of tidal objects , or trying to identify in catalogs of dwarf galaxies those which might have a tidal origin . we have explored both approaches . since the seminal work by toomre & toomre ( 1972 ) , numerous numerical simulations have been made to study the evolution of colliding galaxies . if the reliability of the computer calculations has often been checked from their ability to produce realistic tidal tails , most of these efforts were focussed in understanding what happens in the most central regions . the simulations by barnes & hernquist ( 1992 ) were the first ones to show the formation , out of tidal material , of bound clumps with masses of up to @xmath1 , distributed all along the tails . as stressed later on , among others by hibbard & mihos ( 1995 ) , the tidal material will quickly fall back unless it was originally sent to large distances where it may survive for more than a hubble time . in real interacting systems , however , this is precisely where the most massive tdg candidates , with apparent masses of a few @xmath2 , are found ( unless these are just the result of projection effects ; see arguments against that idea in the contribution by amram et al . in this volume ) . very recently , we were able to reproduce for the first time in n - body simulations the formation of such massive objects near the tip of long tidal tails ( bournaud , duc & masset , 2003 ) . this could be achieved when we adopted very extended dark matter halos for both parent galaxies . such large halos are actually consistent with theoretical cosmological models ; they are however usually disregarded and truncated in numerical simulations because they consume a large number of particles ( and cpu time ) but have only a minor impact on the merger remnant where most of the attention is . figure 1 presents our simulations ; it shows how the gas of one galaxy with an extended dark matter halo reacts to the perturbation of another one of the same mass . note in particular how the gas piles up at the tip of one of the tidal tails . self - gravity makes it become a compact object with a mass of @xmath2 . two gyr after the first encounter and the dissolution of the tails , it is still visible on a quasi - circular orbit at large radii and appears as a classical companion galaxy . in simulations where long tidal tails are able to form ( and hence for which the colliding galaxies have the adequate orbital parameters and relative velocities ) , the production of massive , long - lived , tidal objects is not exceptional . however , we still need to explore more systematically the parameter space to obtain quantitative predictions on the production rate of tdgs . in any case , our preliminary study indicates that massive tidal objects may be created during tidal collisions , and have a life time long enough to become old tidal dwarf galaxies , as previously defined . therefore searching for such objects in the real universe makes sense . how to identify old tdgs when they have no obvious connection an optical stellar bridge or an hi gaseous bridge to any nearby merging galaxy ? hunter et al . ( 2000 ) proposed to look at their dark - matter content , expected to be small if tdgs are made of disk material and if most of the dark matter is distributed in a halo . such objects should also have a special location on the tully - fisher diagram . however , by far the easiest way is an unusually high metallicity . standard isolated galaxies follow a fair correlation between their luminosity ( mass ) and their metallicity . the less massive ones tend to be less metal - rich because they are less able to retain their heavy elements that the most massive ones . this correlation is shown in figure 2 . galaxies departing form that relation and having an oxygen abundance which is too high for their luminosity can be considered as tdg candidates . indeed tidal dwarfs are made of recycled , pre - enriched , material . in fact , the oxygen abundances of young tidal objects ( observed near confirmed mergers , and shown in fig . 2 ) have an oxygen abundance of about 1/2 - 1/4 solar , which is independent of their luminosity , and typical of that of the external regions of spiral disks from where their building material originally comes . one should note that other phenomena can account for a high metallicity in a dwarf galaxy . an external confinement effect by a dense intergalactic medium may prevent it from expelling its heavy elements . some dwarfs could also be the remnants of more massive and metal rich galaxies that have been disrupted by collisions ( see the contribution of k. bekki in this volume ) . those effects should principally apply in dense environments such as in clusters of galaxies . as a side effect , molecular gas as traced by the co emission should be more easily detectable in pre - enriched dwarfs whereas it is not observed in the metal poor classical dwarfs . strong co emission has been detected in most massive tdgs ( braine et al . 2001 , and his contribution in this volume ) . is there any observational evidence that tdgs may survive for at least 500 myr and may hence already be considered as old ? first of all , a few confirmed tdgs are observed in the vicinity of advanced mergers . in these systems , the nuclei of the colliding galaxies have already merged . the associated time scale given by numerical simulations is typically 0.5 - 1 gyr . examples of such objects are shown in figure 3 . in those , the tidal tails are already quite faint whereas the most massive and distant condensations are still clearly visible . optical spectroscopy indicates that all of them have oxygen abundances consistent with those expected for a tidal origin . we explore in the following whether even older , completely detached , tdgs are present in catalogs of classical dwarf galaxies , starting with the local group . our closest companion , the sagittarius dwarf galaxy , is currently being eaten by the milky way . measuring with high - resolution uves spectra element abundances of a sample of individual stars , bonifacio et al ( 2004 ) estimate that a substantial metal rich ( [ fe / h]=-0.25 ) population exists in sgr . such a high degree of chemical enrichment either suggests that sgr was a much larger galaxy in the past , or that it was detached from the lmc or the mw during their interaction . the distribution of the other dwarf spheroidals in the local group along a great circle suggests that they also may have been involved in this collision ( lynden - bell , 1982 ) . however , their metallicities , usually considered to be low and their high dark matter content ( although challenged by kroupa , 1997 ) make them unlikely tdg candidates . one of the closest groups of galaxies , the m81 group , is well known for the spectacular tidal interaction in which three of its members are involved : m81 , m82 and ngc 3077 . numerous tdg candidates or intergalactic hii regions were identified in that environment ( makarova et al . , 2002 ; durrell et al . in this volume ) ; among them the garland object ( walter et al . , in this volume ) , and the protogalactic molecular cloud near holm ix are best studied . whereas , on optical images , the latter intergalactic objects appear `` detached '' , hi maps show that they actually lie towards the prominent gaseous tidal tail linking the three main galaxies and in which they were probably born . they can not hence be considered as old tdgs . the closest cluster of galaxies , virgo , has its own tdg candidate : a hi , co and oxygen rich low - surface brightness galaxy , not too far from the lenticular ngc 4694 ( braine et al . , in this volume ) . studying the early - type dwarf galaxies in the nearby clusters of coma and fornax , resp . rakos et al . ( 2000 ) and poggianti et al . ( 2001 ; see also her contribution in this volume ) , found that a significant fraction of them deviate from the metallicity - luminosity relation ; the youngest ones being the more metal - rich ones . we reached the same conclusion in our own complete survey of hi - rich , star forming dwarf galaxies in the hydra ( duc et al . , 2001 ) and hercules ( iglesias - paramo et al . , 2003 ) clusters : about 30% of them have unusually high oxygen abundances in their hii regions . it is still premature to conclude from these surveys whether old tdgs contribute significantly to the population of dwarf galaxies . as already mentioned , other phenomena may account for the deviating metallicities . on the other hand , selection criteria only based on metallicities disregard tdgs that would have formed out of the most external and metal poor disk material , or those formed in the early universe when collisions were more frequent and galaxies less metal rich . likely the frequency of tdgs will depend on the environment . compact groups may be a particularly favorable one ( hunsberger et al . , 1996 ) . our numerical simulations indicate that long - lived ( i.e. , formed at least 1 - 2 gyr ago ) tdgs should exist . whether they will survive for a hubble time is still unknown . one may however note that the falling back onto a giant galaxy is also the fate of numerous classical dwarfs . barnes , j. e. and hernquist , l. , 1992 , nature 715 , 360 bonifacio , p. , sbordone , l. , marconi , g. , pasquani , l. and hill , v. , 2004 , a&a 414 , 503 bournaud , f. , duc p .- a . and masset , f. , 2003 , a&a 411 , l469 braine , j. , duc , p .- a . , lisenfeld , u. , charmandaris , v. , vallejo , o. , leon , s. and brinks , e. , 2001 , a&a 378 , 51 duc , p.a . and mirabel , i. f. , 1999 , in iaus 186 : galaxy interactions at low and high redshift , barnes , j. e. and sanders , d. b. eds duc , p .- a . , cayatte , v. , balkowski , c. , thuan , t. x. , papaderos , p. , van driel , w. , 2001 , a&a 369 , 763 hibbard , j. e. and mihos , j. c. , 1995 , aj 140 , 110 hunsberger , s. d. , charlton , j. c. and zaritsky , d. 1996 , apj , 462 , 50 hunter , d. a. , hunsberger , s. d. , roye , e. w. , 2000 , apj 542 , 137 iglesias - p ' aramo , j. , van driel , w. , duc , p .- a . , papaderos , p. , vlchez , j.m . , et al . , 2003 , a&a 406 , 453 knierman , k. a. , gallagher , s. c. , charlton , j. c. , hunsberger , s. d. and whitmore , b. , 2003 , aj 126 , 1227 kroupa , p. , new astronomy , 2 , 139 lynden - bell , d. , 1982 , the observatory , 102 , 7 makarova , l. n. , grebel , e. k. , karachentsev , i. d. , dolphin , a. e. , karachentseva , v. e. et al . , 2002 , a&a 396 , 473 poggianti , b. m. , bridges , t. j. , mobasher , b. , carter , d. and doi , m. , 2001 , apj 562 , 689 rakos , k. d. , schombert , j. m. , odell , a. p. and steindling , s. , 2000 , apj 540 , 715 richer , m. and mccall , m. , 1995 , apj 642 , 445 toomre , a. and toomre , j. , 1972 , apj 623 , 178 weilbacher , p. m. , duc , p .- a . and fritze - v . alvensleben , u. , 2003 , a&a 397 , 545	most tidal dwarf galaxies ( tdgs ) so - far discussed in the literature may be considered as young ones or even newborns , as they are still physically linked to their parent galaxies by an umbilical cord : the tidal tail at the tip of which they are usually observed . old tidal dwarf galaxies , completely detached from their progenitors , are still to be found . using n body numerical simulations , we have shown that tidal objects as massive as @xmath0 solar masses may be formed in interacting systems and survive for more than one gyr . old tdgs should hence exist in the universe . they may be identified looking at a peculiarity of their genetic identity card " : a relatively high abundance in heavy elements , inherited from their parent galaxies . finally , using this technique , we revisit the dwarf galaxies in the local universe trying to find arguments pro and con a tidal origin . # 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in
weak absorbers , those with @xmath8 , constitute @xmath9% of the total absorber population @xcite at @xmath10 . they account for a fair fraction of the @xmath11 forest @xcite . unlike strong absorbers ( which almost always are associated with a @xmath1 galaxy @xcite ) , weak absorbers can usually not be associated with a @xmath1 galaxy within impact parameter @xmath12 kpc of the qso ( @xcite ; c. steidel , private communication ) . this lack of an association with bright galaxies suggests that weak absorbers may be a physically different population than strong absorbers . studying the physical conditions of weak absorbers ( eg . , metallicity , ionization conditions , total column density , and size ) is useful for two reasons : 1 ) physical conditions provide clues as to the nature of these absorbers , whose physical origin is not known ; 2 ) weak clouds provide an opportunity to study metal enriched environments over a range of redshifts . they trace metal production either in intergalactic space or in dwarf or low surface brightness galaxies . a study of weak absorbers at @xmath13 requires spectra in both the optical and in the uv in order that the several key low and high ionization transitions ( especially , , ) , as well as the lyman series , are covered . using the keck / high resolution spectrograph ( hires ) @xcite at high resolution ( @xmath14 ) and the faint object spectrograph ( fos)/_hubble space telescope _ ( hst ) at low resolution ( @xmath15 ) , @xcite applied photoionization models to @xmath16 single cloud weak absorbers . they argued that many multiple cloud weak absorbers ( which comprise @xmath17% of the weak absorbers ) are likely to be part of the same population as the strong absorbers , but that single cloud weak systems are likely to be a different population . at least half of the weak , single cloud systems of @xcite require a second phase of gas in order to reproduce the absorption and/or to fit the profile without exceeding the column density derived from the lyman limit break . while the clouds typically have doppler parameters of @xmath18@xmath19 , the second phase must have a larger effective doppler parameter ( @xmath20@xmath21 ) . with only the low resolution uv data available to @xcite , it was not clear if this implies a single broad component or multiple , blended clouds . also , in the systems for which the second phase is not required ( such as when only a limit for can be derived ) , it was not clear if the phase is just less extreme in its properties or if it is absent . by comparing the column densities to the profiles , @xcite inferred that weak , single cloud absorbers have metallicities of at least one tenth solar in the phase of gas in which the absorption arises . the sizes and densities of most of the weak , single cloud systems were unconstrained ; however , three of them had @xmath22 , which implies low ionization conditions , relatively high densities , @xmath23 , and small sizes , @xmath24 pc . those systems with limits on may be part of a `` continuum '' of single cloud weak absorbers , with some having just below the detection threshold , implying a continuous distribution of ionization conditions . alternatively , some of the systems without detected could be part of a different population , with ionization conditions , metallicities , and/or environments distinctly different from systems with detected . in may and june 2000 high resolution uv spectra ( @xmath25 ) of the @xmath26 quasar , pg @xmath0 , became public in the hst archive . we have high resolution optical ( hires / keck ) spectra of this same quasar . for the first time , weak absorbers can be studied through simultaneous high resolution coverage of numerous transitions . this allows more direct inferences of the physical properties of the gas , such as metallicity and phase structure . in particular , it should enable a determination of the nature of the second , broad phase required for some absorbers by the large relative strength of absorption in lower resolution data . is this higher ionization , broad phase centered on the cloud or is it offset ? is it produced by multiple clouds or by a single , smooth structure ? are systems for which was not detected at low resolution fundamentally different , or do they merely have a weaker second phase ? the kinematics and physical properties of the second phase , and its relationship to the cloud phase , are important diagnostics of the type of structure responsible for the weak absorption . in anticipation of the release of the high resolution stis / hst spectra , we previously pursued an in depth study of four absorbers along the pg @xmath0 line of sight @xcite using the available low resolution fos / hst spectra and the hires / keck spectra . since only the brightest quasars can , in the near future , feasibly be studied at high resolution in the uv , we aimed to compare inferences drawn on the basis of low resolution spectra to those that would be obtained once the higher resolution spectra were released . confirmation of our conclusions would lend credibility to larger statistical studies that rely on a combination of high and low resolution spectra @xcite . the present paper focuses on three single cloud weak absorbers along the pg @xmath0 line of sight . two of these absorbers , at @xmath27 and @xmath28 , were detected in the hires / keck spectra at the sensitivity of our original weak survey ( @xmath29 for @xmath30 ) @xcite . these two systems are dramatically different from each other in that the @xmath27 system has a stringent limit on ( @xmath31 at @xmath32 ) while the @xmath28 absorber has detected at @xmath33 . in the latter system , the arises not in the cloud phase , but in a second phase of gas . this provides an excellent opportunity both to study a system with very weak and to address the nature of the second phase . the metallicity could not be derived for the @xmath27 system because the fos / hst spectrum did not cover any of the lyman transitions ; however , the new stis / hst spectra can address metallicity for this system . models of the @xmath28 absorber did , however , imply a super solar metallicity and/or a depleted or @xmath34enhanced abundance pattern . this conclusion will be re assessed in this paper . also , as we will describe in [ sec : data ] , we have now identified a third single cloud weak system at @xmath35 that was just below the detection threshold of our previous study . that system was not modeled in @xcite , but will be considered in the present paper . we begin in [ sec : data ] by presenting , for the three single cloud weak systems toward pg @xmath0 , high resolution profiles of all relevant transitions covered by stis / hst or hires / keck spectra . in [ sec : techniques ] , we discuss our strategy to infer the physical conditions of these systems by applying cloudy photoionization models . we also consider the possibility that collisional ionization dominates . the results inferred for the phase structure , ionization parameters / densities , metallicities , kinematics , and abundance patterns of the three systems are presented in [ sec : results ] . in [ sec : caveats ] , the effect of relaxing assumptions of abundance pattern and the shape of the ionizing spectrum are considered . the three weak systems are compared in [ sec : discussion ] . in that discussion , we particularly focus on the nature of the second phase and compare the results to those obtained on the basis of just the lower resolution fos / hst spectrum . finally , we summarize our conclusions in [ sec : conclude ] . we briefly describe the observations with hires / keck , stis / hst , and fos / hst of pg @xmath0 , and then present the three single cloud weak absorbers along this line of sight at @xmath36 , @xmath28 , and @xmath35 . a wfpc2/hst image of the quasar field exists ( program 6740 , s. oliver , p. i. ) but the quasar is quite bright so it is not possible to perform adequate psf subtraction in order to detect galaxies close to the line of sight . also , without redshifts of candidate galaxies in the field , we could not separate out the identities of the three weak and two strong absorbers . the hires / keck observations covered @xmath37 to @xmath38 at @xmath14 ( fwhm @xmath39 ) ( @xmath40 pixels per resolution element ) . , , and transitions are covered for the three systems , but in all cases only is detected ( at @xmath41 ) . equivalent widths and limits are listed in table [ tab : ewtab ] . the spectra were obtained on 1995 july 4 and 5 , and the combined spectrum has @xmath42 over most of the wavelength range , but gradually falling toward the lowest wavelengths . the most stringent limits on come from the reddest of the transitions , at @xmath43 . the reduction of the spectrum , continuum fits , line identification , and procedure for voigt profile fitting were described in @xcite . the stis / hst observations provided useful coverage from @xmath44 to @xmath45 at @xmath25 ( fwhm @xmath24 ) ( @xmath18 pixels per resolution element ) . two sets of echelle spectra were obtained with two different tilts of the e230 m grating , using the @xmath46 @xmath47 @xmath46 slit . the first , with central wavelength @xmath48 , was obtained by burles et al . in 1999 may and june ( proposal i d 7292 ) . the total exposure time was @xmath49 s. the second , with central wavelength @xmath50 , was obtained by jannuzi et al . in 1999 june ( proposal i d 8312 ) , with a total exposure time @xmath51 s. the two spectra overlap in the region @xmath52@xmath53 . we co added the spectra , weighted by the exposure times , and also combined multiple order coverage in the same spectrum . however , we also considered the differences between the two realizations as an indication of systematic errors ( due to continuum fitting , correlated noise , and unknown factors ) when comparing model profiles to the data . reductions were done with the standard stis pipeline and continuum fits using the sfit task in iraf . based upon our original analysis of the hires / keck spectrum there were two single cloud , weak systems along the pg @xmath0 line of sight , at @xmath36 and @xmath28 . after analyzing the stis / hst spectra , another slightly weaker system was found in the hires / keck spectra , just below the threshold of our previous survey @xcite . first , the stis spectra were searched for doublets . only one doublet was found with comparable equivalent width to those associated with the two known absorption systems . the new system , at @xmath35 , also had detected , a doublet , , and in the stis spectrum . based upon the equivalent width of we expected that should be detected in the hires / keck spectrum . we then searched that location in the hires spectrum and found the doublet . the @xmath54 transition was detected at the @xmath55 level , and the @xmath56 transition at the @xmath57 level . ( our previous survey had a @xmath32 detection threshold for the @xmath56 transition . ) figure [ fig : data81 ] presents detected transitions and limits of interest in constraining the conditions in the @xmath36 system . figure [ fig : data90 ] displays the same for the @xmath28 system , and figure [ fig : data65 ] for the @xmath35 system . table [ tab : ewtab ] lists the equivalent widths for detected transitions and selected equivalent width limits ( @xmath32 ) for the three systems . we fit each transition with the minimum number of voigt profile components consistent with the errors @xcite . in all three systems , the and the other low and intermediate ionization tranistions could be fit with one component . however , the profiles required two components for the @xmath28 system , and three for the @xmath35 system . for selected transitions , column densities and doppler parameters of the voigt profile fits , performed separately for each transition , are given in table [ tab : vptab ] . for the @xmath35 system , numbers are estimated only for and because fits are ambiguous due to blending . there is no flux detected from pg @xmath0 shortward of a lyman limit break at @xmath58 , which is due to a strong absorber at @xmath59 @xcite . also , a partial lyman limit break at @xmath60 reduces the flux to about @xmath21% of the original continuum level due to a multiple cloud weak absorber at @xmath61 @xcite . although for the @xmath36 system is covered in a @xmath62 stis g230 m spectrum of pg @xmath0 , this reduced flux and crowding with lyman series lines from the @xmath59 system prevent us from deriving a useful constraint . we therefore have not used the g230 m spectrum as a constraint in the analysis . several descriptions of the basic technique used for modeling were given in previous papers @xcite . for those efforts , only , , and were covered at high resolution , and all other transitions were observed with fos / hst at only @xmath15 . even with the availability of numerous transitions at high resolution , the modeling technique used in the present paper is very similar to those previous efforts . we begin with the phase of gas that produces the dominant absorption , `` the low ionization phase '' . then we add in other phases of gas as needed to reproduce the observed absorption profiles for all transitions . for each single cloud absorber we begin with the column density , @xmath63 , and the doppler parameter , @xmath64 derived from a voigt profile fit to the doublet . cloudy photoionization models ( version 94.00 ; @xcite ) were applied to determine column densities of the various transitions that would result from the same phase of gas that produces this , assuming a slab geometry . the parameters for these models are the metallicity @xmath65 ( expressed in units of the solar value ) , the abundance pattern ( initially assumed to be solar ) , and the ionization parameter , @xmath66 . the ionization parameter is defined as the ratio of ionizing photons to the number density of hydrogen in the absorbing gas , @xmath67 . we assume that the ionizing spectrum is of the form specified by @xcite for @xmath68 . the normalization of the haardt and madau background at @xmath68 is fixed so that @xmath69 . the assumption that stellar sources do not make a substantial contribution is likely to be valid since weak absorbers typically do not have a nearby high luminosity galaxy , such as a starburst . the effects of alternative spectral shapes are discussed in [ sec : specshape ] . the doppler parameters of other elements are derived from @xmath64 , using the temperature output from cloudy to derive the thermal and turbulent contributions for each element . for each choice of parameters , the cloudy output column densities and doppler parameters are used to synthesize noiseless spectra , convolving with the instrumental profile characteristic of stis / e230 m . the synthetic model spectra are compared with the observed profiles to identify permitted regions of the parameter space . the model column densities and doppler parameters for permitted models are within @xmath70 of the values measured from the data . from the three systems studied in this paper , @xmath71 is covered , but not detected . for @xmath72 , the ratio @xmath73 is strongly dependent on the ionization parameter in the optically thin regime ( see @xcite ) and the limit can be used to place a lower limit on @xmath66 of the low ionization phase . to obtain an upper limit on the ionization parameter of this phase , and provide constraints . for optically thin gas , the constraints on @xmath66 do not depend significantly on the metallicity . the metallicity is constrained by fitting the and any higher lyman series lines that are covered for the system . low metallicities will overproduce the in the wings . high metallicities will underproduce the , but can not be excluded since the additional absorption can arise in a different phase . we assumed a solar abundance pattern , but note that there is a simple tradeoff between metallicity and the abundance of the metal line transition that is compared to the hydrogen , discussed further in [ sec : abun ] . in general , once the basic constraints on @xmath66 are derived , we consider whether a single cloud model can fully reproduce profiles of all of the observed transitions . _ in all three of the systems modeled in this paper , we will conclude that the absorption can not be fully produced in the same phase with the absorption . _ for the range of permitted values of @xmath66 for the low ionization phase , we constrain the properties of the high ionization phase that is required to fit the profile . first , we consider whether a single , relatively broad component is sufficient . if the profile is asymmetric ( as in the @xmath28 system ) or shows velocity structure ( as in the @xmath35 system ) , then it is clear that more than one component is needed . a voigt profile fit to the serves as a starting point for deriving the number of clouds needed in the high ionization phase , and their @xmath74 and @xmath75 . results from these fits are listed in table [ tab : vptab ] . the @xmath75 of these clouds is larger than for the clouds . cloudy models of this additional phase are constrained to match the observed clouds and @xmath66 is adjusted to determine what range of values are consistent with other transitions . an upper limit is set so that and ( if covered ) are not overproduced . a lower limit is set so that the broader components do not produce observable and . in some cases the intermediate ionization transitions , such as , , and especially , could not be fully produced in the cloud phase . in these cases , an intermediate value of @xmath66 is sought for the clouds in order to also account for the remainder of these transitions , without overproducing the lower ionization transitions . a lower limit on the metallicity of a cloud can be derived in order that absorption is not overproduced . more exactly , it is the combination of the and clouds that are constrained not to overproduce absorption , so there is a trade off between their metallicity constraints . if all transitions of a certain element ( such as , , and ) are underproduced relative to other elements , then simple abundance pattern variations ( @xmath34enhancement and depletion ) are considered . for the high ionization phase , we consider collisional ionization models as alternatives to cloudy photoionization solutions . in this case , an alternative source of heating ( eg . , shocks ) must be responsible for heating the gas to the assumed , higher @xmath76 . the measured doppler parameter of the , @xmath75 , is used to place an upper limit on the temperature , @xmath77 . for an assumed @xmath76 and @xmath65 , and with the measured @xmath74 , the collisional equilibrium tables of @xcite were used to determine the column densities of all other covered transitions . the doppler parameters of these other transitions , @xmath78 were calculated from @xmath79 , where the turbulent component of the doppler parameter is given by @xmath80 . in order to constrain @xmath76 , synthetic spectra were generated and superimposed upon the data to facilitate comparison . also , the model column densities of the various transitions were compared to the measured values . the profiles provided a constraint on the metallicity of any collisionally ionized phase . `` wings '' on these profiles can be produced by such a phase , which would be characterized by a relatively large @xmath81 parameter . the contribution of a collisionally ionized phase to is minimized at low @xmath76 and high metallicity . a lower limit on @xmath76 is therefore set so as not to overproduce at the highest reasonable metallicity ( usually taken to be solar ) . as for photoionization models , for the high ionization phase , the lower limit on the metallicity depends on the metallicity of the low ionization phases , which determines its contribution to . first , we describe the model results for the two weak , single cloud absorbers detected in the original hires / keck survey @xcite . then the results for the newly discovered @xmath35 system are presented . a range of parameters for satisfactory models are summarized in table [ tab : tabmod ] . model profiles for an example of an acceptable model are superimposed on the data for each of the three systems in figures [ fig : data81 ] , [ fig : data90 ] , and [ fig : data65 ] . parameters for these sample models are listed in table [ tab : tab4 ] . results in this section use the simplest set of assumptions that produce models consistent with the data , i.e. haardt and madau spectrum and a solar abundance pattern . the effects of alternate spectra and abundance patterns are discussed in [ sec : caveats ] . figure [ fig : data81 ] shows detections of , , , , , , and . there are useful limits for and . an obvious , but important , first result is that is now clearly detected in this system , despite the strong limit from the earlier low resolution data . we begin with the voigt profile fit to the doublet , and adjust the ionization parameter to fit as many of the other transitions as possible . we find that the cloud is constrained to have @xmath82 . there is no strict lower limit , because the @xmath83 vs. @xmath66 curve is flat for @xmath84 . very small values of @xmath66 ( as low as @xmath85 ) are permitted , but cloud sizes would be extremely small . for @xmath86 , is underproduced . to find an upper limit on ionization parameter , we first determined that , for @xmath87 , the model would produce minimum fluxes at the positions of and that are consistent with the observed profiles . however , the model @xmath88 and @xmath75 are small compared to the observed values , i.e. the model profiles are narrow compared to the observed profiles . we therefore conclude that the and are produced in a separate higher ionization phase , and that @xmath89 for the cloud phase . the contribution of the cloud phase to the absorption profiles ( negligible for all but the singly ionized transitions ) is denoted by a dotted line in figure [ fig : data81 ] . the metallicity of the cloud with detected is constrained to be @xmath90 dex greater than solar ( for a solar abundance pattern ) . for lower metallicities , absorption in the wings of the profile will exceed that observed . also , unless the metallicity is even higher for the cloud , the additional high ionization phase ( required to fit ) is constrained not to give rise to significant absorption . for @xmath91 and @xmath92 , the cloud size would @xmath93 pc . for @xmath94 and @xmath95 , the cloud size would be only @xmath96 au , and the observed profile would be somewhat underproduced by the model . primarily because of the breadth of the profiles , we concluded that a second phase is required to fit the , even in this system for which the absorption is relatively weak . a voigt profile fit to the profile yields an adequate fit for a single cloud with @xmath97 that is centered on the cloud . details of the fit are listed in table [ tab : vptab ] . we consider whether the can be produced by photoionization , and/or whether it can be produced by collisional ionization in gas that has been heated above the equilibrium value . for photoionization models of the high ionization phase , we optimized on the voigt profile fit values , and the ionization parameter was constrained by data for other intermediate and high ionization transitions . to produce the observed absorption in this phase , @xmath98 is the optimal range . the cloud size would be @xmath99 pc , considerably larger than the cloud . if @xmath100 , the system would have too much and absorption relative to . if @xmath101 , @xmath102 would be overproduced by the model . the metallicity of the high ionization phase is constrained , by the profile , to be solar or higher . for larger values of the ionization parameter , within the constrained range of @xmath98 , the lower limit on metallicity would be raised to a supersolar value . considering collisional ionization , for @xmath103 , pure thermal broadening gives an upper limit on the temperature of @xmath104 . below this limiting temperature , would be severely overproduced @xcite if we optimize on . however , raising the temperature to @xmath105 ( which is just consistent with the voigt profile fit @xmath75 within @xmath70 errors ) reproduces the observed @xmath106 . the metallicity of this collisionally ionized phase would have to be solar or greater in order that would not be overproduced . though it requires fine tuning of the temperature , a collisionally ionized phase with solar metallicity and with @xmath105 provides an adequate fit to the data . we conclude that the @xmath36 system has two phases , both with a metallicity solar or higher . the low ionization phase , with @xmath107 , has @xmath84 , while the high ionization phase , with @xmath108 , could be photoionized with @xmath109 or collisionally ionized with @xmath105 . the broader high ionization phase is centered on the cloud and is consistent with a single cloud producing the absorption . constraints are summarized in table [ tab : tabmod ] . an acceptable model , including two photoionized phases , is superimposed on the data in figure [ fig : data81 ] . column densities of selected transitions , produced by this model , are given in table [ tab : tab4 ] . the @xmath28 system is detected in , , , , , , , 1403 , , and . 1239 is not confirmed by a detection of 1242 , so it is viewed as a tentative detection . is also a likely detection , although 1038 is in a confused region of the spectrum that required an uncertain continuum fit . limits are available for and 2600 in the hires / keck spectrum . all of these transitions are shown in figure [ fig : data90 ] . in this system , the is quite strong , and shows an asymmetry in both members of the doublet . we first optimize on the @xmath63 and @xmath64 given by a voigt profile fit to the mg0.1emii @xmath110 profiles ( see table [ tab : vptab ] ) . comparing to the other transitions , the ionization parameter for this cloud is constrained to be @xmath111 . a higher value overproduces @xmath112 absorption at @xmath113 , which is well fit for @xmath66 at the upper end of this range and underproduced for @xmath114 . ( @xmath115 is blended with a stronger transition from another system and can not be used as a constraint . ) a value of @xmath116 would result in a model that exceeds the limit on @xmath43 . if @xmath117 , then there would have to be a significant contribution to the absorption from another phase . even for the maximum consistent @xmath118 , the absorption is not fully produced in this phase , as shown by the dotted curve in figure [ fig : data90 ] . the cloud giving rise to the , with @xmath119 , would overproduce in its blue wing unless it has solar or super solar metallicity . the metallicity constraint could be slightly relaxed if magnesium is enhanced due to an @xmath34enhanced abundance pattern . on the other hand , in the red wing , the observed is not exceeded by any model with @xmath120 . this difference between the metallicity constraints from the red and blue wings of the profile is a consequence of an asymmetry in the distribution of relative to the . the asymmetry in , along with the need to fully produce the absorption , requires an additional phase of gas . in addition to the cloud , two more clouds are needed in order to fit the profile . from a simultaneous voigt profile fit of the @xmath121 and @xmath122 transitions , the first , centered on the velocity of the absorption , would have @xmath123 and @xmath124 . a small amount of the absorption is produced in the same phase with the for @xmath118 , so the column density of the cloud would be reduced slightly in that case . the second cloud , offset by @xmath16 to the red , is fit with @xmath125 and @xmath126 . first , we consider the @xmath127 cloud centered on the profile . it is too narrow ( @xmath128 ) for the to be produced by collisional ionization . from cloudy photoionization models , to simultaneously fit the and the absorption , we derive the constraint , @xmath129 , for this high ionization phase . if so , little arises in this phase , which implies that absorption should be produced in the low ionization phase with the . revisiting the constraints for the cloud , we now find the constraint @xmath130 for the low ionization cloud , at the upper end of the previous constrained range discussed above . the metallicity of the high ionization gas that produces the absorption must be high enough not to overproduce in its blue wing . in conjunction with a solar metallicity cloud phase , this yields a solar metallicity for this @xmath129 phase as well . next , we consider the @xmath131 cloud , offset by @xmath16 from the cloud , that fills in the red wing of the profile . the @xmath132 cloud , at @xmath133 , can be fit with a simple , single cloud photoionization model with @xmath134 . this ionization parameter produces a consistent fit of the , , and , without overproducing the . the and data are very slightly underproduced by the model at this velocity , suggesting that a small ( @xmath135 dex ) abundance pattern enhancement of silicon might be needed . a simple model with the minimum number of phases would call for the asymmetry in the profile to be produced in this cloud along with the red wing of the . the metallicity is thus constrained to be at the solar value in order to fit the red wing of the line . our philosophy of fitting with the minimum number of phases argues against collisional ionization as the mechanism for producing the observed absorption at @xmath136 . to be consistent with this philosophy , both and should arise in the same phase . the limit on @xmath137 for this component with @xmath132 is consistent with production of absorption by collisional ionization . however , for the maximum permitted @xmath76 there would not be enough absorption in this model component unless the metallicity was well over supersolar . with the relatively low @xmath76 needed to produce the much absorption , absorption would be overproduced . we conclude that the @xmath28 system must have solar metallicity or higher in its low ionization phase . the @xmath138 cloud that produces absorption must be of a relatively high ionization state , with @xmath130 . the range of derived cloud sizes for this constrained range is @xmath21@xmath139 pc . two higher ionization , broader clouds are also required : one , with @xmath140 , is centered on the cloud , while the other @xmath141 cloud is offset by @xmath16 . both are consistent with photoionization with @xmath142 , and have sizes of @xmath143@xmath144 kpc . ranges of acceptable model parameters are presented in table [ tab : tabmod ] and model column densities contributed by the three clouds for an adequate model are listed in table [ tab : tab4 ] . the predictions for this same model are superimposed on the data in figure [ fig : data90 ] . the profile for this newly discovered system has a complex structure and the is extremely weak and narrow . the profile requires at least a three component fit . the is not symmetric about the , but it is approximately symmetric about the . these profiles as well as those of the other detected transitions , , , , are displayed in figure [ fig : data65 ] , along with the region of the spectrum that provides a limit on 2600 . is in a noisy region of the spectrum . if the detected feature is really , it is present only in the redward component , at @xmath145 . we begin by considering the cloud . since the 2600 spectrum is noisy , the low ionization parameter constraint on @xmath66 is not strong ; for @xmath146 , @xmath73 is relatively constant . however , assuming solar abundance pattern , @xmath147 provides a marginally better fit to the and . if @xmath148 , then the depth of the and absorption at @xmath149 can be reproduced in the cloud , but the model profile is too narrow and does not provide a good fit to the data . a separate higher ionization phase is needed to fit these higher ionization transitions . considering all transitions , and assuming solar abundance pattern , @xmath150 is favored for the cloud . the metallicity of the cloud would be @xmath151 if all of the in the red wing were to arise in this cloud . however , it could also be significantly higher if there were contribution from the clouds . for @xmath151 and @xmath152 , the cloud size ranges from @xmath19 pc to @xmath153 kpc . sizes scale with @xmath154 so that @xmath155 clouds would be a factor of @xmath156 smaller than @xmath151 clouds . consider the case of @xmath91 for the cloud . three more clouds are required to fit . the column densities and doppler parameters for these three clouds were obtained with a voigt profile fit . this fit is not unique because clouds can not be well separated due to blending , so there are no errorbars listed in table [ tab : vptab ] . the cloud centered on the was fit with @xmath157 , @xmath158 . the other two clouds , at @xmath159 and @xmath160 , were fit with @xmath161 , @xmath162 , and @xmath163 , @xmath164 , respectively . for photoionization models , the ionization parameters for these three clouds are constrained by the requirement that they produce the observed without producing significant lower ionization transitions . the @xmath165 cloud , centered on the cloud , is tightly constrained to have @xmath166 , by the profiles . the range already takes into account the fact that the cloud would make a small contribution to if its @xmath167 . the lower limit on @xmath66 of the cloud arises in order that the is not overproduced in its wings by this relatively broad component . the upper limit applies in order to produce sufficient absorption . the @xmath162 cloud at @xmath159 , has @xmath168 , in order to produce the optimal fit to , , and . for @xmath169 , is overproduced , and for @xmath170 , , , and are overproduced . for the @xmath132 cloud at @xmath160 , @xmath171 provides an optimal fit to the , and is consistent with and with the limit on . a value of @xmath172 is only marginally consistent , with the model slightly underproducing the . this inferred properties of this cloud are independent of the cloud since no is detected at this velocity . as with the cloud , the metallicities of the three additional clouds can not be extremely low . only @xmath173 or higher is consistent with the data . if the red side of the is to be produced by the @xmath160 cloud , @xmath155 would apply for this cloud . similarly , @xmath155 for the @xmath113 cloud would match the blue wing of . in the case of the blue wing , the could in principle be produced in a @xmath174 cloud and the metallicity of the cloud could be higher . constraints on the metallicities of these clouds , relative to the cloud that produced the narrower profile , are limited by the lack of coverage of higher order lyman series lines . however , the data are consistent with @xmath173 in the cloud and in the three clouds . for this metallicity , the sizes of the three clouds would be @xmath175@xmath176 kpc . we also consider whether collisionally ionized gas can be consistent with the observed profiles of , , and . for the @xmath113 cloud , @xmath177 does not overproduce , and @xmath178 is consistent with producing the blue side of the profile for @xmath173 . if the metallicity is higher , a higher ionization parameter is needed to fit the line , but in this case the model profile shape is not consistent with the data . the @xmath159 cloud is too narrow to be consistent with production of absorption through collisional ionization assuming the particular voigt profile fit that we have adopted . however , because this fit is not unique , there could possibly be a broader cloud component that could be reconciled with collisional ionization . finally , the @xmath160 cloud could be collisionally ionized with @xmath179 and @xmath173 , in which case it would match the red side of the profile . in summary , the @xmath35 system profile is quite weak and narrow . however , the profile can be fit with several components spread in velocity over @xmath180 . the three components are similar to each other , having @xmath181 and @xmath66 ranging from @xmath182 to @xmath183 if they are photoionized . collisional ionization with temperatures of @xmath184 could be consistent with the data , but requires fine tuning of the temperature and seems less likely . the bluest component has the detected at the same velocity , but in a narrower component ( @xmath185 ) produced in gas with a somewhat lower ionization parameter ( @xmath186 ) . the clouds have a metallicity of at least @xmath173 and must be at least a couple of kiloparsecs in size . abundance ratios measured in galactic stars show a clear range of @xmath187 \leq + 0.5 $ ] @xcite . in our presentation of model results ( [ sec : results ] ) solar abundance pattern was assumed unless the data require an alternative pattern . however , it should be noted that the inferred metallicity depends directly on the assumed abundance pattern . if , instead of solar , the abundance pattern is @xmath34group enhanced with @xmath188 } = + 0.5 $ ] , the inferred metallicity of a phase constrained by @xmath63 ( an @xmath34group element ) would proportionally drop by @xmath189 dex . constraints on the ionization parameter , @xmath66 , are also affected by changes in the assumed abundance pattern . for example , if the the abundance pattern is @xmath34group enhanced , the ratio @xmath190 would constrain @xmath66 to be larger than if the abundance pattern is solar . however , for @xmath188 } \le + 0.5 $ ] , the change in the constraint on @xmath66 ( and therefore on @xmath191 ) is less than @xmath192 dex . we have used the relative absorption strengths in several other transitions to determine @xmath66 , but @xmath192 dex is typical of the level of uncertainty due to variations in abundance pattern . based on results for other similar absorption line systems , there is likely to be no bright galaxy associated with these single cloud weak absorbers @xcite . therefore , we used the @xcite extragalactic background spectrum for @xmath68 in our detailed models presented above ( [ sec : results ] ) . however , because it is not strictly excluded , we do consider here the effect of changing the spectral shape . two alternative galaxy spectra , @xmath193 gyr and @xmath194 gyr instantaneous starburst models , were superimposed on the haardt and madau extragalactic background spectrum . both models , with solar metallicity and a salpeter imf were taken from @xcite . the normalizations of the starburst spectra are defined relative to the haardt and madau spectrum at @xmath144 rydberg , and in all cases the extragalactic and galactic contributions were superimposed . the largest reasonable value for the photon flux from even the most extreme starburst galaxy is @xmath195 , and of order @xmath144% of the photons escape @xcite . using these numbers , to have a flux ten times that of haardt and madau at @xmath68 , the absorber would have to be within @xmath19 kpc of the starburst . within this distance , we would expect stronger absorption than observed in these weak absorbers . therefore , we consider it most likely that the haardt and madau spectral shape is the appropriate assumption . nonetheless , in this section , we briefly examine the consequences of the alternatives . we consider the effect of spectral shape on the metallicity of the low ionization phase , on whether the absorption can arise from the same phase as the absorption , and on the ionization conditions of the high ionization phase . the largest impact on the metallicity would be from an ionizing spectrum with a large feature at the lyman edge , such as the bruzual and charlot @xmath194 gyr instantaneous burst model . in this case , the cloud would be more neutral so that not as much hydrogen would be required to fit the profile . in such a case , the metallicity would be even higher than we inferred assuming a pure haardt and madau spectrum . in general , we found that the starburst spectrum normalization needed to be at least ten times that of the haardt and madau spectrum in order to detect a difference in the models . even with a normalization of twenty five the difference in the inferred metallicity is less than @xmath196 dex . most importantly , even if the spectrum of ionizing radiation has a substantial lyman edge , the metallicity of the weak clouds would be even higher than the solar value an even more surprising result . a @xmath194 gyr instantaneous burst model , if it dominates over the haardt and madau spectrum by a factor of twenty five , can make significantly more relative to . however , only the column density can be made consistent with the observed profiles in the three systems studied here . the doppler parameters of these lines are still too large for them to be produced in the same phase with the . a spectrum with a sharp edge will have the largest effect on the inferred @xmath66 of a photoionized high ionization phase . a @xmath193 gyr starburst model is extreme in this respect , yet it takes a normalization of ten times haardt and madau to see even a small change . with a normalization of twenty five , the is significantly overproduced relative to for the same @xmath66 . however , the qualitative result of a relatively low density phase producing the bulk of the is unchanged . for example , for the @xmath36 system , @xmath66 must be increased from @xmath197 to @xmath198 in order that is not overproduced . we conclude , that at @xmath68 an extreme starburst spectrum would have to dominate in order to affect the conclusions of our models . even if such conditions prevailed , the conclusions would be qualitatively unchanged , only changing constraints by @xmath199 dex in parameter spaces ranging over a few orders of magnitude . this study of the multiple phases of gas and the physical conditions of the three weak systems along the pg @xmath0 line of sight indicates a heterogeneous population of objects selected by a weak doublet . figure [ fig : allsys ] is a comparison between the three systems showing the range of equivalent widths and kinematic structures , ranging from a profile consistent with a single cloud for the @xmath36 system , to the stronger asymmetric profile for the @xmath28 system , to the multiple component profile for the @xmath35 system . the profile strength does not systematically increase in proportion with the absorption . it could , however , be increasing with the kinematic spread of the . table [ tab : tabmod ] gives a summary of the range of `` acceptable models '' for each of the three systems , while table [ tab : tab4 ] gives more detailed information about a sample model that is within the acceptable range . that typical model was also superimposed on the data in figures [ fig : data81][fig : data65 ] . the @xmath36 system and the @xmath28 system provide contrasting examples of a weak absorber that clearly requires a separate broad phase to fit a strong profile and one with much weaker , not even detected in the previous low resolution fos spectrum @xcite . for both systems , in the present study we find that a separate broader phase ( @xmath7 and @xmath19 ) , with an ionization parameter @xmath200 , is needed to fit the observed stis profile . this phase is centered at the same velocity as the phase that gives rise to the weak absorption . however , in the @xmath28 system there is an asymmetry to the profile that indicates an additional offset component , @xmath201 redward of the cloud and the first broad phase . the profile can be fit with a two cloud model , though it is also possible that there is a more complex distribution of material , spread over a range of velocities , that gives rise to such an asymmetric profile shape . a general point is suggested by the comparison of the @xmath36 and @xmath28 systems . the strong absorption that is observed in many weak absorbers @xcite could be due to the presence of separate clouds that are of higher ionization and do not give rise to detectable absorption . the same would apply for the stronger absorption apparent in some weak absorbers . the @xmath35 system is a more extreme example in which three separate components are clearly apparent in the profile . a narrow , weak absorber is centered on one of the higher ionization clouds . the separate components at @xmath202 and @xmath203 are quite distinct from the centered on the cloud . like the higher ionization components of the other two systems , these two offset clouds have ionization parameters , @xmath200 . the profile is consistent with being centered around the three component , but not around the cloud . the fact that absorption is strongest in this system suggests that the spread in velocity of absorbing gas along the line of sight is an important factor in determining the equivalent width of absorption . the metallicities of the @xmath36 and @xmath28 systems are constrained to be at least as high as solar . the @xmath35 system is likely to have a metallicity of at least @xmath204th solar . these relatively high metallicities appear to be common for weak absorbers . rigby et al . ( 2001 ) inferred a high metallicity for several systems based upon low resolution fos data , and in no case did they find that a metallicity less than @xmath204th solar was required . for the @xmath36 system and the @xmath28 system , we can compare the results from this study , incorporating high resolution stis data , to our previous models , based upon the lower resolution fos data @xcite . for the @xmath36 system , was not covered in the earlier fos spectrum , so no constraints on metallicity were available . the was not detected in the fos spectrum , but we showed that it might be detected in the higher resolution stis spectrum either from a cloud with a high ionization parameter or from a broader , separate phase . no specific predictions could be made as to which possibility was more likely . now , with the high resolution stis spectrum coverage of we are able to place specific constraints as was outlined in [ sec : res81 ] . for the @xmath28 system the main conclusions of our previous study , based on the fos spectrum @xcite , were that the is present in a separate broader phase , and that the metallicity of the cloud is solar or higher . these conclusions are confirmed by the present study . the detailed properties of the broader phase were more difficult to determine based upon low resolution spectra . we suggested that the broad phase has @xmath205 because the doublet ratio of the was large compared to observations , if @xmath108 was assumed . the high resolution stis spectrum shows that the profile is asymmetric , so that it must be composed of at least two separate `` clouds '' , the broadest having @xmath206 . the `` wings '' of the profile in the low resolution spectrum were apparently due to fos fixed pattern noise rather than to an extremely broad phase , since these features are not apparent in the high resolution stis spectrum . from this very limited comparison we tentatively conclude that it should be possible to draw inferences about the presence of a separate phase based on a combination of high and low resolution spectra , i.e. drawing on the large existing fos database . also , our conclusion of solar metallicity for the @xmath28 system was confirmed by modeling of the high resolution stis profile of . this is important , since modeling of fos data was used to infer that many weak absorbers have close to solar metallicity @xcite . to understand the nature of these systems , we seek to infer the spatial distribution of absorbing material , and the relationship to star forming objects . all three single cloud weak absorbers have two phases that produce a narrower ( @xmath18@xmath176 ) and a broader ( @xmath19@xmath201 ) absorption component centered at the same velocity . the narrower component is of lower ionization than the broader component . for a fixed haardt madau spectrum intensity , this implies a higher density ( @xmath207@xmath194 ) for the narrower component , and also a smaller size , @xmath93@xmath139 pc . the broader component would arise in a higher ionization / lower density phase ( @xmath208 ) with a larger size ( @xmath93@xmath18 kpc ) . two simple scenarios could be consistent with the inferred properties of the narrow and broad components : 1 ) the first would have the lower ionization region embedded within the higher ionization region , and the higher ionization region would present a larger cross section . we would then expect many systems to be observed for lines of sight that pass through only the higher ionization region of such structures . these `` only systems '' might typically have lower column densities than two phase weak systems . they would be produced at large impact parameter in the structure , at which the pathlength would be shorter and the gas densities would be lower . 2 ) in the second scenario the lower ionization components could be produced by parts of a shell structure surrounding a lower density , higher ionization region . the covering factors for the low ionization shell fragments would be limited ( @xmath209 ) by the lack of observation of many two cloud weak systems . however , with such a small covering factor , we would again expect a large incidence of `` only systems '' which in this scenario would tend to have similar phases to those of two phase weak systems . we searched the pg @xmath0 spectra and found no `` only systems '' with comparable in strength even to that of the @xmath36 weak absorber . however , we have identified several weaker candidate doublets at @xmath210 , confirmed by a line detected at the expected position of . many other lines of sight need to be systematically surveyed , but there clearly will be limits on the geometry and covering factors of the two phases of gas . an alternative to the idea of embedded phases is to have separate clouds along the line of sight with different densities and sizes . this is consistent with the presence of an offset cloud in the @xmath28 system , and the spread of three clouds over @xmath180 in the @xmath35 system . these separate clouds could exist as condensations in larger structures with velocity dispersions of tens of , e.g. , dwarf galaxies . however , if the phases were all completely separate from each other then it would be hard to explain the close alignment in velocity of the lower ionization cloud with one of the higher ionization clouds . the metallicities of these absorbing structures present a clue as to their place of origin . we do not expect that they are in the vicinity of luminous galaxies ( @xmath211 galaxies are not found within impact parameters of @xmath12 kpc from the quasar ) . although no useful image is available for the pg @xmath0 field in particular , other single cloud weak absorbers are rarely found near such luminous galaxies ( c. steidel , private communication ; @xcite ) . dwarf galaxies have metallicities significantly lower than solar , which would appear inconsistent with solar metallicities for single cloud weak absorbers . however , it is possible that the weak absorbers are concentrations of higher metallicity within lower metallicity structures . most strong absorbers ( @xmath212 ) also require a phase in addition to the clouds in order to produce the observed absorption . these absorbers are associated with @xmath213 galaxies @xcite , and the high ionization phase is inferred to have an `` effective doppler parameter '' or velocity spread of @xmath145 @xcite . at high resolution some of the profiles will separate into multiple components , while others may be due to a more uniform distribution of gas @xcite . the phase of some strong absorbers is reminiscent of what would be expected for a corona such as that observed in and absorption around the milky way disk . the single cloud weak absorbers , although they do have a second phase , do not appear to be related to such a corona . the combination of stis / hst and hires / keck high resolution spectra , covering multiple chemical transitions , provided the first opportunity to collect direct information on the metallicities and phase structure of weak absorbers . there are three weak absorbers , at @xmath36 , @xmath28 , and @xmath35 along the line of sight toward the quasar pg @xmath0 . all three of these absorbers have a second , higher ionization phase , giving rise to the absorption . the broad phase in one case is consistent with a single cloud , and in the other two cases requires one or two additional clouds separated in velocity space from the one aligned with the absorption . two of the weak absorbers are constrained to have solar or greater than solar metallicity , and the other one to have a metallicity greater than @xmath214th solar . thus , in general , weak absorbers are not weak because of a low metallicity ( confirming the result of @xcite ) . as introduced in [ sec : intro ] , it is also possible that they have weak absorption because they are more highly ionized than their strong counterparts , or because their total column densities are smaller . the likely answer is that there is some combination of these two effects , perhaps leading to different populations of weak absorbers arising in different environments . the ionization parameters of the absorbers studied here are higher than inferred for many of the clouds in strong absorbers for which is also detected . [ is a tracer of low ionization conditions @xcite . ] however , there is also a population of weak absorbers with strong lines , amounting to about one third of the weak absorber population @xcite . detailed study of the phase structure of this sub group awaits spectra of additional quasar lines of sight with , and other transitions covered at high resolution . weak absorbers are potentially of general importance because they provide a sensitive probe of particular types of star forming environments . in principle , this population of objects can be used to track the chemical and ionization history of the universe in regions that are not in luminous galaxies which can be studied by other methods . for example , do they exist with solar metallicity to high redshifts , i.e. are there selected environments with extreme enrichment even at early times ? answering this question will require a large systematic study of many weak systems over a range of redshifts . such a study could address whether there is always absorption centered in velocity on a cloud . it could tabulate the distribution of clouds in velocity space , and address whether a large equivalent width is typically due to the presence of multiple clouds along the line of sight . spectral coverage of and will also provide better constraints on the ionization conditions of the high ionization phase . finally , it is highly desirable to search the quasar fields for faint galaxies that could be related to these weak absorbers . support for this work was provided by the nsf ( ast9617185 ) and by nasa ( nag 56399 and hst go08672.01a ) , the latter from the space telescope science institute , which is operated by aura , inc . , under nasa contract nas526555 . s. zonak , j. rigby , and n. bond were supported by an nsf reu supplement . we thank c. howk and k. sembach for their invaluable guidance on the analysis of high resolution stis spectra . & @xmath215 & @xmath216 & @xmath217 + & & @xmath218 & + 2853 & @xmath219 & @xmath220 & @xmath221 + 2796 & @xmath222 & @xmath223 & @xmath224 + 2803 & @xmath225 & @xmath226 & @xmath227 + 2600 & @xmath228 & @xmath229 & @xmath230 + 1193 & & @xmath231 & @xmath232 + 1260 & @xmath233 & @xmath234 & @xmath235 + 1335 & @xmath236 & @xmath237 & @xmath238 + 1084 & @xmath232 & @xmath239 & + 1207 & @xmath240 & @xmath241 & + 1394 & @xmath242 & & @xmath243 + 1403 & @xmath242 & @xmath244 & @xmath245 + 989 & & @xmath246 & + 1548 & @xmath247 & @xmath248 & @xmath249 + 1551 & @xmath250 & @xmath251 & @xmath252 + 1239 & @xmath219 & @xmath253 & @xmath254 + 1243 & & @xmath255 & @xmath256 + 1032 & & @xmath257 & + -0.05 in + & @xmath258 & @xmath259 & @xmath149 + & @xmath260 & @xmath261 & @xmath149 + & @xmath262 & @xmath263 & @xmath149 + & @xmath264 & @xmath265 & @xmath149 + & @xmath266 & @xmath267 & @xmath149 + + & @xmath268 & @xmath269 & @xmath149 + & @xmath270 & @xmath271 & @xmath149 + & @xmath272 & @xmath273 & @xmath149 + & @xmath274 & @xmath275 & @xmath149 + & @xmath276 & @xmath277 & @xmath149 + & @xmath278 & @xmath279 & @xmath16 + + & @xmath280 & @xmath281 & @xmath149 + & @xmath282 & @xmath283 & @xmath149 + & @xmath284 & @xmath285 & @xmath202 + & @xmath286 & @xmath287 & @xmath203 + -0.05 in 0.8182 & @xmath149 & @xmath288 to @xmath289 & @xmath290 & @xmath291 to @xmath292 & @xmath18 & @xmath293 to @xmath294 + & @xmath149 & @xmath295 to @xmath296 & @xmath297 & @xmath298 to @xmath299 & @xmath300 & @xmath301 to @xmath302 + 0.9056 & @xmath149 & @xmath303 to @xmath304 & @xmath297 & @xmath305 to @xmath194 & @xmath40 & @xmath306 + & @xmath149 & @xmath296 to @xmath197 & @xmath297 & @xmath143 to @xmath144 & @xmath19 & @xmath307 to @xmath301 + & @xmath16 & @xmath308 to @xmath296 & @xmath297 & @xmath143 to @xmath196 & @xmath309 & @xmath301 to @xmath310 + 0.6534 & @xmath149 & @xmath289 to @xmath303 & @xmath305 to @xmath144 & @xmath311 to @xmath144 & @xmath176 & @xmath312 to @xmath313 + & @xmath149 & @xmath182 to @xmath314 & @xmath194 to @xmath144 & @xmath18 to @xmath176 & @xmath201 & @xmath315 to @xmath316 + & @xmath202 & @xmath295 to @xmath183 & @xmath194 to @xmath144 & @xmath18 to @xmath176 & @xmath317 & @xmath301 to @xmath316 + & @xmath203 & @xmath295 to @xmath318 & @xmath194 to @xmath144 & @xmath18 to @xmath176 & @xmath309 & @xmath319 to @xmath320 + -0.05 in + @xmath321 & @xmath149 & @xmath322 & @xmath289 & @xmath323 & @xmath194 & @xmath324 & @xmath325 & @xmath326 & @xmath327 & @xmath328 & @xmath329 & @xmath317 & @xmath18 & @xmath40 + @xmath321 & @xmath149 & @xmath322 & @xmath318 & @xmath330 & @xmath331 & @xmath332 & @xmath333 & @xmath307 & @xmath334 & @xmath335 & @xmath336 & @xmath16 & @xmath300 & @xmath300 + + @xmath321 & @xmath149 & @xmath337 & @xmath304 & @xmath338 & @xmath339 & @xmath340 & @xmath341 & @xmath342 & @xmath343 & @xmath344 & @xmath345 & @xmath346 & @xmath40 & @xmath176 + @xmath321 & @xmath149 & @xmath337 & @xmath197 & @xmath311 & @xmath347 & @xmath348 & @xmath349 & @xmath350 & @xmath351 & @xmath352 & @xmath353 & @xmath354 & @xmath176 & @xmath19 + @xmath355 & @xmath16 & @xmath337 & @xmath296 & @xmath356 & @xmath357 & @xmath348 & @xmath358 & @xmath301 & @xmath359 & @xmath360 & @xmath353 & @xmath361 & @xmath309 & @xmath309 + + @xmath321 & @xmath149 & @xmath194 & @xmath289 & @xmath323 & @xmath18 & @xmath362 & @xmath363 & @xmath325 & @xmath364 & @xmath365 & @xmath366 & @xmath309 & @xmath176 & @xmath367 + @xmath321 & @xmath149 & @xmath194 & @xmath182 & @xmath368 & @xmath347 & @xmath369 & @xmath370 & @xmath371 & @xmath364 & @xmath372 & @xmath373 & @xmath374 & @xmath346 & @xmath201 + @xmath355 & @xmath202 & @xmath194 & @xmath295 & @xmath375 & @xmath376 & @xmath377 & @xmath370 & @xmath342 & @xmath378 & @xmath379 & @xmath380 & @xmath361 & @xmath381 & @xmath317 + @xmath382 & @xmath203 & @xmath194 & @xmath295 & @xmath375 & @xmath383 & @xmath377 & @xmath384 & @xmath306 & @xmath385 & @xmath343 & @xmath386 & @xmath387 & @xmath309 & @xmath309 + -0.05 in	high resolution optical ( hires / keck ) and uv ( stis / hst ) spectra , covering a large range of chemical transitions , are analyzed for three single cloud weak absorption systems along the line of sight toward the quasar pg @xmath0 . weak absorption lines in quasar spectra trace metal enriched environments that are rarely closely associated with the most luminous galaxies ( @xmath1 ) . the two weak systems at @xmath2 and @xmath3 are constrained to have @xmath4 solar metallicity , while the metallicity of the @xmath5 system is not as well constrained , but is consistent with @xmath6th solar . these weak clouds are likely to be local pockets of high metallicity in a lower metallicity environment . all three systems have two phases of gas , a higher density region that produces narrower absorption lines for low ionization transitions , such as , and a lower density region that produces broader absorption lines for high ionization transitions , such as . the profile for one system ( at @xmath2 ) can be fit with a single broad component ( @xmath7 ) , but those for the other two systems require one or two additional offset high ionization clouds . two possible physical pictures for the phase structure are discussed : one with a low ionization , denser phase embedded in a lower density surrounding medium , and the other with the denser clumps surrounding more highly ionized gas .
in very recent years , several stars have been proposed to have planetary companions on the basis of measured precise radial velocity variations . this field of research is experiencing rapid development , and updated reviews of the present situation can be found in the proceedings of the workshop on _ brown dwarfs and extrasolar planets _ edited by rebolo et al . ( 1998 ) and in the extrasolar planets encyclopaedia by j. schneider . once a solar - type star has been suggested to harbour a planetary - mass companion , it is interesting to investigate any similarities with the sun , as well as to find possible differences with respect to other single stars . chemical abundances are among the most important parameters to be compared and , in particular , precise abundances of light elements such as lithium and beryllium ( easy to destroy by @xmath6 nuclear reactions when the temperature reaches @xmath7 and @xmath8 k , respectively ) combined with the abundances of other elements which are not so readily destroyed in stellar interiors , should help to understand how the presence of planets may affect the chemical composition of their parent stars . gonzalez ( 1997 , 1998 ) has derived the overall metallicities as well as abundances of different elements ( including lithium ) for a wide sample of proposed parent stars , finding that four of the known systems show a metallicity significantly higher than the solar value . a peculiar system such as 16 cyg a & b , formed by twin solar - type stars of which only one has an orbiting planet ( cochran et al . 1997 ) , is an especially suitable candidate to perform a detailed abundance study . gonzalez ( 1998 ) found that both stars have a similar metallicity with a value slightly larger than solar , and confirmed independently a previous result of king et al . ( 1997a ) that 16 cyg b ( the star with a suspected planet ) is strongly depleted in lithium with respect to 16 cyg a. the knowledge of their beryllium abundances is of potential value in quantifying the possible influence of a planetary companion on the mixing mechanisms operating in the stellar interior . @xmath9 cnc is a star with spectral type g8v , and is the coolest known object which is a candidate to have a planetary companion . following gonzalez ( 1998 ) , this star falls into the group having roughly jupiter - mass companions with small circular orbits and very metal - rich parent stars . dominik et al . ( 1998 ) have shown recently that the planetary system of @xmath9 cnc also hosts a vega - like disk of dust , evidenced by an infrared excess at 60 @xmath10 m . the star is very depleted in lithium and its beryllium abundance could be compared with existing upper limits measured in younger stars with similar effective temperatures ( garca lpez et al . 1995a ) . in this paper we derive the beryllium abundances of the 16 cyg system and of @xmath9 cnc by comparing observations with spectral syntheses of the @xmath0beii @xmath1 3131 doublet . we use those , together with their published lithium values , as well as with available abundances for other stars ( with and without suggested planetary companions ) , and discuss briefly possible effects of planets on processes taking place in their structure and evolution . the observations were carried out in several runs conducted at the 2.6 m nordic optical ( not ) and 4.2 m william herschel ( wht ) telescopes of the observatorio del roque de los muchachos ( la palma ) , using the iacub ( mckeith et al . 1993 ) and utrecht echelle ( ues ) spectrographs , respectively . table 1 lists the stars observed , dates , and telescopes . most of the observations were performed at the not and only two exposures of 1200 s each were devoted to 16 cyg b at the wht . @xmath9 cnc and 16 cyg a were observed with iacub using a slit width of @xmath11 arc sec which provided spectra with a resolution @xmath12 recorded using a @xmath13 ( 19 @xmath10 m ) thomson blue - coated ccd , while the slit width used to observe 16 cyg b at the not was @xmath14 arc sec and the resolution @xmath15 . iacub is an un - crossed echelle spectrograph and the order corresponding to the beii doublet was isolated using an interference filter centered at @xmath1 3131 , with a fwhm of 45 , and a maximum transmission of 64 % . the ues observations of 16 cyg b were performed using the e79 grating , a slit width of @xmath16 arc sec , @xmath17 , and recorded with a @xmath18 ( 24 @xmath10 m ) site ccd . eleven spectral orders were available on the detector , the bluest one with useful signal containing the beii doublet . lccccccc star & name & @xmath19 & @xmath20 & telescope & date & exp . time ( s ) & s / n + hr 3522 & @xmath2 cnc & 5.95 & 0.87 & not & 19/11/1996 & 3600 & 35 + hr 7503 & 16 cyg a & 5.96 & 0.64 & not & 19/07/1997 & 1800 & 35 + hr 7504 & 16 cyg b & 6.20 & 0.66 & not & 22/07/1997 & 1800 & 30 + & & & & wht & 13/11/1997 & 2400 & 15 + [ table1 ] data reductions were performed by standard procedures using routines included in the iraf suite of programs . final signal - to - noise ratios ( s / n ) achieved in the pseudo - continuum surrounding the beii doublet are listed in table 1 . technical problems with the ues and the low elevation of the object during the observations ( airmass @xmath21 ) prevented the achievement of a better spectrum of 16 cyg b with the wht . due to the absence of thar lines available for these observations , wavelength calibrations were carried out using photospheric lines present in the stellar spectra whose values were taken from the moore et al . ( 1966 ) solar atlas . second - order polynomial fits where applied , using 1118 lines , to give calibrations with rms scatter between 0.004 and 0.011 . dispersions of 0.017 and 0.035 pixel@xmath22 were obtained for @xmath9 cnc and 16 cyg a & b , respectively , with iacub ( with a binning of two pixels in the spectral direction for the latter cases ) , while a dispersion of 0.033 pixel@xmath22 corresponded to the ues spectrum of 16 cyg b. stellar parameters were taken from the detailed abundance analysis carried out by gonzalez ( 1998 ) . adopted effective temperatures ( @xmath23 ) , surface gravities ( @xmath24 ) , and metallicities ( [ fe / h ] ) are listed in table 2 . typical errors associated with these quantities are @xmath25 k for @xmath23 and @xmath26 dex for both @xmath24 and [ fe / h ] . a solar oxygen abundance log n(o)@xmath27 ( anders & grevesse 1989 ; where log n(x)@xmath28log ( x / h)@xmath2912 ) was taken , in all cases . small changes of this abundance ( like the updated solar value log n(o)@xmath30 by grevesse & noels 1996 ) , which affect the oh lines located in the beii region , do not alter significantly the comparison between observed and synthetic spectra . the beryllium abundance analysis was performed by spectral synthesis fitting to the @xmath0beii @xmath1 3131.065 line , which is weaker but more isolated than its companion at @xmath1 3130.421 . synthetic spectra were computed using the code wita2 , a unix - based version of code abel6 ( pavlenko 1991 ) , which computes lte atomic and molecular line profiles , and the model photospheres were interpolated for the adopted stellar parameters within a grid of atlas9 models ( computed without overshooting ) provided by kurucz ( 1992 ) . details of the line list employed ( which is adjusted to reproduce the solar spectrum and uses accurate oscillator strengths for the beii lines ) , the absence of significant nlte effects on the derived abundance ( @xmath31 dex ) , and the low sensitivity of the observed feature to changes in the beryllium abundance for low @xmath23 values were shown in garca lpez et al . ( 1995a , b ) . the values of the parameters adopted for 16 cyg a & b are very close to the solar values . figure 1 shows the comparison between observed and synthetic spectra for the sun and 16 cyg a & b. the solar spectrum , with a resolution @xmath17 , was obtained by observing the moon with the combination wht@xmath29ues in a previous campaign during april 1995 . synthetic spectra have been convolved with gaussians with the appropriate fwhms to reproduce the different instrumental profiles . a beryllium abundance log n(be ) @xmath32 ( chmielewski et al . 1975 ; anders & grevesse 1989 ) reproduces very well the observed solar @xmath0beii @xmath1 3131.065 line , while this is not the case for other surrounding lines ( including the @xmath0beii @xmath1 3130.420 line ) , and provides a fiducial comparison for the quality of the best fit which can be achieved for solar - type stars . garca lpez et al . ( 1995b ) illustrate the high sensitivity of solar synthetic spectra to changes in the beryllium abundance . the best fit to the observed weak beii line for 16 cyg a indicates an abundance of log n(be ) @xmath33 , i.e. a beryllium abundance not significantly different from solar . changes of @xmath34 k in @xmath35 , @xmath36 dex in @xmath24 , and @xmath37 dex in [ fe / h ] imply variations of @xmath38 , @xmath37 , and @xmath38 dex , respectively , in the solar be abundance ( garca lpez et al . 1995b ) . the corresponding maximum uncertainties for the parameters adopted for 16 cyg a are @xmath38 , @xmath39 , and @xmath38 dex , respectively . including an additional 0.05 dex abundance error induced by the uncertainty in locating the pseudo - continuum around the beii doublet , and another 0.1 dex induced by the uncertainty associated with the available s / n , the final error resulting from combining these errors in quadrature amounts @xmath40 dex . cccccc star & @xmath23 ( k ) & @xmath24 & [ fe / h ] & log n(li ) & log n(be ) + @xmath9 cnc & 5150@xmath41 & 4.15@xmath41 & 0.29@xmath41 & @xmath420.04@xmath41 & @xmath43 + 16 cyg a & 5750@xmath41 & 4.20@xmath41 & 0.11@xmath41 & 1.24@xmath41 & @xmath44 + 16 cyg b & 5700@xmath41 & 4.35@xmath41 & 0.06@xmath41 & @xmath450.46@xmath41 & @xmath46 + 70 vir & 5538@xmath47 & 4.02@xmath47 & @xmath480.04@xmath47 & 1.79@xmath47 & @xmath49 + hd 114762 & 5865@xmath47 & 4.31@xmath47 & @xmath480.66@xmath47 & 1.94@xmath47 & @xmath50 + @xmath51 and & 6050@xmath52 & 4.0 @xmath52 & 0.06@xmath52 & 2.15@xmath52 & 0.90@xmath52 + @xmath53 boo & 6390@xmath52 & 3.8 @xmath52 & 0.30@xmath52 & @xmath45 0.6@xmath52 & @xmath54 + data taken from gonzalez ( 1998 ; @xmath55 ) , this work ( @xmath56 ) , stephens et al . ( 1997 ; @xmath57 ) , and boesgaard & lavery ( 1986 ; d ) . [ table2 ] the beryllium abundance derived from a similar analysis applied to the spectrum of 16 cyg b observed with not@xmath29iacub is log n(be ) @xmath58 , slightly larger than that derived for 16 cyg a but compatible with it and the solar value within the error bars . this shows that while there is a difference of a factor 6 ( at least ) in the lithium abundances of both stars ( listed in table 2 , and taken from gonzalez 1998 ) , there is no indication of beryllium depletion among them . a service observation of 16 cyg a & b was requested at the wht aimed at improving the s / n obtained with the not and better constraining the slight difference in be abundance between the two stars . however , technical problems with the ues and the restriction imposed at the time of the observation by the coordinates of the objects allowed us to obtain only one final low - s / n spectrum for 16 cyg b. the ues spectrum shown in fig . 1 has been smoothed slightly , and the synthetic spectrum overplotted was computed with an abundance of log n(be ) @xmath59 . the low effective temperature of this star makes its beryllium abundance analysis more difficult and uncertain . garca lpez et al . ( 1995a ) studied the sensitivity of the observed beii @xmath1 3131.065 feature to the beryllium abundance in late - type stars belonging to the hyades open cluster and the ursa major group ( umag ) , and found that this sensitivity decreases with decreasing @xmath23 . a line of another element which is blended with the be line becomes very important for cool stars , clearly dominating the feature when the temperature drops below @xmath60 k. they tentatively identified the perturbing line as mni @xmath1 3131.037 , and primas et al . ( 1997 ) and king et al . ( 1997b ) also found evidence of such a blend in their analyses of @xmath61 cen a & b , but suggesting different blending features . as a result of this limitation , garca lpez et al . derived reliable beryllium abundances only for three hyades stars with @xmath62 k , and established upper limits for the cooler stars in their sample . on the other hand , garca lpez ( 1996 ) investigated the potential use of @xmath0bei lines to overcome the uncertainties associated with the beii line at low effective temperatures , finding that the best candidate ( @xmath1 2348.609 , not observable from the ground ) would not provide reliable abundances . figure 2 shows the comparison between the observed spectrum of @xmath9 cnc and several synthetic spectra computed with different be abundances and the stellar parameters listed in table 2 . as seen in the upper panel , the fit of the synthetic spectrum to the observed one is not as good as for solar - like stars ( fig . 1 ) ; a value log n(be ) @xmath63 reproduces the observed feature . the lower panel is a zoom of the region surrounding the beii @xmath1 3131.065 line , and the observed spectrum is represented by photon statistics error bars to better demonstrate the sensitivity of the observed feature to changes in beryllium abundance . five synthetic spectra , computed without beryllium and with log n(be ) @xmath64 , 0.3 , 0.5 , and 0.7 , respectively , are also shown . although the spectrum computed with log n(be ) @xmath64 reproduces the observations , the spectrum without beryllium is well included within the error bars suggesting that an upper limit instead of a measurement is a more prudent result here . this upper limit could be as high as log n(be ) @xmath65 given the s / n in the points defining the line itself . an additional 0.15 dex uncertainty is induced by the errors in the adopted stellar parameters ( mainly @xmath24 ) , 0.1 dex is associated with the uncertainty in locating the pseudo - continuum , and , finally , we consider an error of 0.1 dex related to an uncertainty of @xmath37 dex in the estimated @xmath66 value for the mni line which blends strongly the beii line at this effective temperature . combining these uncertainties in quadrature , our conservative upper limit for the beryllium abundance of @xmath9 cnc is log n(be ) @xmath67 . the lithium and beryllium abundances of 16 cyg a are very similar to those of the sun and @xmath61 cen a , another well known solar analog . adopting @xmath68 k , @xmath69 , and [ fe / h]@xmath70 for @xmath61 cen a ( chmielewski et al . 1992 ) , king et al . ( 1997a ) found log n(li ) @xmath71 and king et al . ( 1997b ) derived log n(be ) @xmath72 for this star . primas et al . ( 1997 ) obtained log n(be ) @xmath73 using the same effective temperature and slightly different gravity and metallicity ( @xmath74 and [ fe / h]@xmath75 ) . stephens et al . ( 1997 ) derived li and be abundances for a large sample of f- and g - type main - sequence ( ms ) stars . two of the stars studied by stephens et al . have also been proposed to have very - massive planetary companions : 70 vir ( hr 5072 ) and hd 114762 , with @xmath76 @xmath4 and @xmath77 @xmath4 , respectively , following gonzalez ( 1998 ) . the effective temperatures and li & be abundances assigned by stephens et al . to these stars are listed in table 2 , and it can be seen that they do not show the same dramatic li depletion as does 16 cyg b. there are two additional parent stars with available be abundances : @xmath51 and and @xmath53 boo , studied by boesgaard & lavery ( 1986 ) using photographic spectra . these stars are the hottest parent stars found up to now . @xmath53 boo , with an effective temperature of 6390 k assigned by boesgaard & lavery , is located well into the so called `` li gap '' of the f - type stars ( boesgaard & tripicco 1986 ) and shows significant li and be depletions , while @xmath51 and ( with @xmath78 k and close to the red edge of the gap ) has kept large amounts of both elements ( see table 2 ) . most of the stars studied by stephens et al . in the interval @xmath79 k show lithium abundances in the range log n(li ) @xmath80 to 2.4 and beryllium abundances between 1.0 and 1.3 ; two stars show li abundances at 1.41 and 1.71 , and some be depletion at @xmath81 and @xmath82 , respectively ; and there is also one star , hd 160693 , which appears to be depleted in both li ( @xmath83 ) and be ( @xmath84 ) . the only two objects in that sample which show somewhat similar behaviour to 16 cyg b are hr 483 and hr 1729 . these have solar beryllium abundance but are significantly hotter ( @xmath85 and 5930 k , respectively ) and have very - high upper limits for lithium ( log n(li ) @xmath86 and @xmath87 ) , which make them also compatible with an object like 16 cyg a. this distribution of abundances can be seen in figure 3 , where we have plotted be against li abundances for stars cooler than @xmath88 k ( to avoid the large scatter observed for the abundances of the hotter stars -stephens et al . 1997 ; garca lpez et al . 1998- ) , including @xmath61 cen a & b as well as the hyades and umag stars studied by garca lpez et al . ( 1995a ) . tables 2 and 3 list the data used to produce fig . 3 . note that there is a difference of about three orders of magnitude in the lithium abundances plotted in the figure , while the corresponding beryllium range is @xmath16 dex . although the stellar parameters employed to derive the abundances for the different samples are not homogeneous and detailed star to star comparisons can not be made , fig . 3 shows that there are not significant differences between stars with and without massive planetary companions , with the exception of 16 cyg b. lcccc star & @xmath23 ( k ) & log n(li ) & log n(be ) & reference + hr 219 & 5883 & 1.90 & 1.19 & 1 + hr 483 & 5862 & @xmath451.85 & 1.16 & 1 + hd 30649 & 5716 & 1.71 & 0.86 & 1 + hr 1729 & 5931 & @xmath451.76 & 1.12 & 1 + hr 3064 & 5941 & 2.04 & 1.47 & 1 + hd 65583 & 5328 & @xmath450.94 & 0.80 & 1 + hr 4845 & 5830 & 1.87 & 1.06 & 1 + hr 5914 & 5801 & 2.36 & 1.27 & 1 + hr 6060 & 5809 & 1.58 & 1.16 & 1 + hd 148816 & 5833 & 1.84 & 0.90 & 1 + hr 6189 & 6058 & 2.41 & 0.72 & 1 + hr 6349 & 6028 & 2.74 & 1.13 & 1 + hd 157089 & 5739 & 1.87 & 1.03 & 1 + hd 160693 & 5701 & @xmath451.14 & 0.42 & 1 + hr 6775 & 5955 & 2.31 & 0.93 & 1 + hd 18499 & 5725 & 1.41 & 0.68 & 1 + hd 195633 & 5864 & 2.29 & 0.71 & 1 + hd 208906 & 5940 & 2.39 & 0.81 & 1 + hr 9088 & 5377 & @xmath450.68 & 0.98 & 1 + hr 9107 & 5574 & 1.86 & 1.34 & 1 + @xmath61 cen a & 5800 & 1.37 & 1.32 & 2,3 + @xmath61 cen b & 5350 & @xmath450.40 & 0.61 & 4,5 + vb 17 & 5635 & 1.99 & 1.10 & 6 + vb 21 & 5250 & 0.35 & 0.90 & 6 + vb 26 & 5465 & 1.22 & 1.00 & 6 + vb 46 & 5065 & @xmath450.29 & @xmath450.90 & 6 + hd 41593 & 5140 & 0.86 & @xmath450.90 & 6 + hd 109011a & 4760 & 0.71 & @xmath450.90 & 6 + hd 110463 & 4800 & 0.71 & @xmath450.90 & 6 + references : ( 1 ) stephens et al . ( 1997 ) ; ( 2 ) king et al . ( 1997a ) ; ( 3 ) king et al . ( 1997b ) ; ( 4 ) chmielewski et al . ( 1992 ) ; ( 5 ) primas et al . ( 1997 ) ; ( 6 ) garca lpez et al . ( 1995a ) . [ table3 ] gonzalez ( 1998 ) estimated an age of @xmath89 gyr for 16 cyg a & b using stellar evolution grids . if the difference in lithium abundances between the two stars ( which have kept a similar be abundance ) is caused by a slow mixing mechanism , which transports the material from the base of the outer convection zone to the li burning layer , associated with the angular momentum loss ( as suggested by king et al . 1997a ) , the stars must have had a different angular momentum history ( i.e. different initial angular momenta or different loss rates ) . this would also be the case if the differences in lithium abundance were related to changes in the stellar structure induced by different rotation rates during the pre - ms phase ( martn & claret 1996 ) . cochran et al . ( 1997 ) suggest that the difference in rotation rate or angular momentum loss between the two stars could be associated with the existence of a massive proto - planetary disk around 16 cyg b while the star was in its pre - ms phase ( giving rise to the present @xmath3 @xmath4 orbiting planet ) , which would have served to brake the stellar rotation and induced a more rapid li depletion compared to 16 cyg a , which would have lacked a similar disk and kept a faster rotation rate . apart from explaining how a system such as 16 cyg a & b , with only one star surrounded by a massive proto - planetary disk can be formed , one must also show why the li depletion is not so severe in 70 vir and hd 114762 with more massive planetary companions and greater ages ( @xmath90 and @xmath91 gyr , respectively ; gonzalez 1998 ) . a possible explanation could lie in the large eccentric orbits of the planetary companions to these stars , with proto - planetary disks which could have then had a smaller effect on the pre - ms stellar rotation rates . although the companion to 16 cyg b also presents a very - high orbital eccentricity , it is possible that it has been altered from a more circular one by gravitational perturbation due to 16 cyg a ( mazeh et al . in addition , a check has to be made on whether or not the mixing mechanisms proposed predict correctly the observed abundances . from figure 11 of stephens et al . ( 1997 ) , it seems that the rotationally - induced model of deliyannis & pinsonneault ( 1993 ) predicts significant li differences and similar ( solar ) beryllium abundances for stars with @xmath92 k , an age of 4 gyr , and a difference of 20 km s@xmath22 in their initial equatorial velocity . however , it is not clear from the figure what would be the predictions for stars of considerably greater ages ( in the range @xmath93 gyr ) . furthermore , long - period orbiting planets of small mass , like the planets of the solar system , could also be able to induce some lithium depletion but are not easy to detect with the currently available observing facilities , making so very difficult to complete separate a priori the stars with and without `` li - depleting '' planets . alternatively , the fact that we do not find stars with lithium and beryllium abundances comparable to those of 16 cyg b ( while it is common to find other stars with orbiting planets behaving like 16 cyg a ) could be telling us something about the chemical enrichment of stars with planetary companions . orbital migration of planets formed at large radii and transported inwards by tidal interactions with the proto - planetary disk could occur ( lin et al . 1996 ) , and would result in adding metal - rich material to the parent star . the degree of chemical enrichment would depend on the fraction of the star over which the accreted material is distributed , which is linked to the age and evolutionary stage of the parent star . laughlin & adams ( 1997 ) predict that solar - type stars with maximum disk lifetimes of @xmath94 myr should have virtually no metallicity enhancement , while more massive stars ( early f- to late a - type , with shallower convection zones ) could experience more significant chemical enrichments . gonzalez ( 1998 ) discusses the possibility that 16 cyg a could have had in the past a planetary companion like that of 16 cyg b , but which was engulfed by its parent star after the time needed to disturb the companion of 16 cyg b from its original circular orbit . this could have increased its observed lithium abundance ( which survives in the outer fraction of the stellar interior where the temperature is lower than @xmath7 k ) without changing significantly ( @xmath31 dex ) the overall stellar metallicity , but what would be its effect on the beryllium abundance ? it will be interesting to perform detailed computations simulating the accretion of planets of different sizes onto parent stars with different masses at different evolutionary stages , to obtain better estimates of the possible enrichment of light elements . however , if this were the explanation of the abundances observed for 16 cyg a & b , what would be the peculiarities of this system which cause it to differ from parent stars with similar effective temperatures , such as 51 peg , 47 uma , hd 114762 , the sun , or even the cooler 70 vir and hotter @xmath51 and , which do not show such dramatic lithium depletion ? new interesting challenges are open for detailed stellar structure and evolution studies . although only an upper limit , to the best of our knowledge this value represents the first evidence of significant beryllium depletion in the coolest ms stars measured . king et al . ( 1997b ) and primas et al . ( 1997 ) derived log n(be ) @xmath95 and log n(be ) @xmath96 for @xmath61 cen b adopting , respectively , @xmath97 and 5350 k , while the upper limit established by garca lpez et al . ( 1995a ) for one hyades and three umag late - type stars with @xmath98 k was log n(be ) @xmath99 , close to the solar abundance . the effective temperature adopted by garca lpez et al . for the umag star hd 41593 is the same as that adopted here for @xmath9 cnc . the analysis of hd 41593 shows that a similar low beryllium abundance is compatible with the observed spectrum , but its low s / n ( @xmath100 ) makes it also compatible with a larger abundance . the stellar parameters derived by gonzalez ( 1998 ) for @xmath9 cnc place the star in the subgiant region of the hr diagram , confirming previous claims in this direction . however , the age derived for the star using theoretical isochrones turns to be @xmath101 gyr , i.e. much greater than the accepted age of the universe . if @xmath9 cnc were indeed a subgiant star , the observed li and be depletions could be due to a dilution effect taking place once the star has evolved off the ms and the convection zone deepens and mixes li- and be - rich material with li- and be - free material from the inner region . strong li and be depletions have been found by boesgaard & chesley ( 1976 ) among several late g- and early k - type subgiants , which appear to agree with theoretical predictions for dilution . a possible alternative explanation suggested by gonzalez ( 1998 ) is that @xmath102 cnc is an unresolved stellar binary viewed nearly pole - on , and this could be tested by searching for variations in the line profile shapes . the very - low lithium abundance of @xmath9 cnc is compatible with strong depletion experienced by an old late g - type ms star ( @xmath103 gyr , baliunas et al . 1997 based on caii h & k chromospheric activity ) , as predicted by extrapolating the 4 gyr li isochrones of the rotating models of deliyannis & pinsonneault ( 1993 ) presented in fig . 11 of stephens et al . extrapolating the corresponding be isochrones in the figure may not give an adequate prediction of the be abundance for @xmath9 cnc , because strong be depletion may occur for effective temperatures much cooler than those considered in the plot . indeed , older rotating models computed by pinsonneault et al . ( 1990 ; case a ) , which include angular momentum loss , predict beryllium depletions of 0.23 to 0.45 dex for 0.80.9 @xmath104 stars ( @xmath105 k ) at 1.7 gyr ( the greatest age considered ) , with initial angular momenta in the range @xmath106 gr @xmath107 s@xmath22 . other angular momentum loss and internal redistribution properties ( cases labeled as b to f ) predict similar be depletions at that age . the strong be depletion observed in @xmath9 cnc would therefore set a very significant constraint on the theoretical models . other old and cool stars which are not suspected to have a companion planet should be observed to distinguish any possible planetary influence on the be abundance of @xmath9 cnc . beryllium abundances have been derived for the solar - like stars 16 cyg a & b and the cooler object @xmath9 cnc , for which there are published values of their lithium abundances . 16 cyg b and @xmath9 cnc are candidates to be parents of extrasolar planets , and by measuring their be abundances we aim at studying the potential dependence on the presence of planetary companions of detailed processes operating in their structure and evolution . 16 cyg a & b show very similar be abundances , which are compatible with the solar value , while the lithium abundance of 16 cyg b is at least a factor 6 smaller than that of 16 cyg a. different rates of mixing of material in their interiors associated with different angular momentum histories , as well as the hypothetical ingestion of a planetary companion by 16 cyg a are discussed as potential explanations . the existence of two other solar - like parent stars , whose li ( and be ) does not show strong depletion , i.e. whose behaviour is like 16 cyg a , the sun and the majority of similar stars with li and be abundances available , implies that the 16 cyg system requires special observational and theoretical attention . a low upper limit has been derived for the beryllium abundance of @xmath9 cnc . this is the first time a precise limit has been set and that such strong be depletion has been observed in a late g-/early k - type ms star . this measurement clearly constrains the depletion predictions of the available mixing mechanisms , but requires observation of planet - free stars with similar age and spectral type to discard the potential effects of the planetary companion on the li and be depletions . claims have also been made indicating that @xmath9 cnc appears to be a subgiant . if this were the case , its strong li and be depletions could be explained by a dilution process taking place during its post - ms evolution . we are indebted to r. corradi , s. kemp , b. bates , and c. r. benn for obtaining part of the observations included in this work . we also thank r. rebolo , j. e. beckman , and g. gonzalez for wide discussions and critical remarks on this topic . the comments of the referee , dr . f. spite , have been of value in improving the content of this article . garca lpez , r. j. , domnguez herrera , m. c. , prez de taoro , m. r. , rasilla , j. l. , rebolo , r. , allende prieto , c. 1998 , in donahue , r. a. , bookbinder , j. a. ( eds . ) 10@xmath108 cambridge workshop on cool stars , stellar systems , and the sun ( in press )	the @xmath0beii @xmath1 3131 doublet has been observed in the solar - type stars 16 cyg a & b and in the late g - type star @xmath2 cnc , to derive their beryllium abundances . 16 cyg a & b show similar ( solar ) beryllium abundances while 16 cyg b , which has been proposed to have a planetary companion of @xmath3 @xmath4 , is known to be depleted in lithium by a factor larger than 6 with respect to 16 cyg a. differences in their rotational histories which could induce different rates of internal mixing of material , and the ingestion of a similar planet by 16 cyg a are discussed as potential explanations . the existence of two other solar - type stars which are candidates to harbour planetary - mass companions and which show lithium and beryllium abundances close to those of 16 cyg a , requires a more detailed inspection of the peculiarities of the 16 cyg system . for @xmath2 cnc , which is the coolest known object candidate to harbour a planetary - mass companion ( @xmath5 @xmath4 ) , we establish a precise upper limit for its beryllium abundance , showing a strong be depletion which constrains the available mixing mechanisms . observations of similar stars without companions are required to asses the potential effects of the planetary companion on the observed depletion . it has been recently claimed that @xmath2 cnc appears to be a subgiant . if this were the case , the observed strong li and be depletions could be explained by a dilution process taking place during its post - main sequence evolution .
the rapidity - gap events recently observed at hera @xcite revived the community s interest in regge - pole description of high energy scattering . the word ` pomeron ' seldom heard in recent years is once again found in the lexicon of experimentalists . it is perhaps then a good time to have a new look at the connection between perturbative qcd and regge pole . non - perturbative effect may be important for momentum transfer of the order of @xmath0 or smaller , but we shall avoid it by going to a larger momentum transfer if necessary . it was proposed by low and nussinov @xcite some years ago that the pomeron may simply reflect a two - gluon exchange in qcd . diagrammatic calculations have been carried out @xcite in the leading log approximation to substantiate this proposal and to study other aspects of the high energy near forward scattering amplitudes . we shall follow the classical approach of implementing physical ( @xmath1-channel ) unitarity by summing up feynman diagrams in the feynman gauge @xcite . however , other approaches are available . the use of physical gauge and dispersion relation @xcite can sometimes make things more transparent . one can also emphasize on the complex angular momentum aspect by concentrating on @xmath2-channel towers and their interactions as guided by @xmath2-channel unitarity @xcite . we shall speak no further of these alternate approaches except to remark that our cut diagram method , to be discussed below , can be thought of an intermediate link between the classical feynman diagram sum and the dispersion relation approaches . due to the gauge and non - abelian nature of the theory , perturbative calculations are lengthy and complicated . even by ignoring self energy and vertex corrections , as well as renormalization effects on the grounds that they will not alter the qualitative nature of high energy scattering we are trying to learn , even by ignoring quark pair productions as a first step , a complete qcd calculation can be carried out only up to the 6th order . in the much simpler case of qed , an 8th order calculation @xcite was reported to have taken sixteen months and two thousand pages to complete . in the case of qcd , it is much worse , a complete calculation of the 8th order is not available , though partial calculations have been carried out @xcite . what makes the calculation so complicated is the large number of diagrams that has to be tackled , and the inevitable cancellations between them . the computation of each diagram is fairly straight forward , though somewhat lengthy at higher orders . after the individual diagrams are calculated , care must also be exercised to add them up because of the presence of many delicate cancellations . leading - log dependences on energy get subtracted away , complicated functions of momentum transfer also disappear . what emerges at the end is a product that is surprisingly simple . to the extent that it has been verified , high energy near forward scattering is described by multiple reggeized gluon exchange , supplemented by elementary gluon production off the reggeons and by @xmath1-channel unitarity @xcite . this outcome has been verified up to the 6th order @xcite . as mentioned above , many delicate cancellations take place in the sum to enable this simple picture to emerge . a complete calculation in the 8th order is not available , but if one assumes these cancellations which take place up to the 6th order also occur in the 8th and higher orders , then this reggeized picture has been verified up to the 10th order@xcite . this gives a strong support to the conjecture that it is true to all perturbative orders . while the conjecture is attractive , it is impossible to verify or refute without a new method to simplify the complicated calculations . we report in this article such a new method , in which high energy scatterings are computed via _ cut diagrams _ rather than the normal feynman diagrams . these cut diagrams are _ not _ the cutkosky cut diagrams . the sum of all cut diagrams here is equal to the sum of all feynman diagrams , and not just their discontinuities . these cut diagrams can be regarded as a resummation of the feynman diagrams in which many of the delicate cancellations encountered in the latter have been built in and explicitly avoided . consequently , the simplicity of the final sum is revealed already in individual cut diagrams , and not masked by terms to be cancelled as is the case with feynman diagrams . besides , individual cut diagrams are easier to calculate than individual feynman diagrams . in this paper we shall introduce the formalism of cut diagrams , as well as an explict calculation to the 6th order to demonstrate its effectiveness . we shall discuss quark - quark scattering throughout but allow the quarks to carry any @xmath3 color charge . the result should be equally valid for gluon - gluon and other scatterings because high energy processes are insensitive to the spins of the colliding particles . we will study pure @xmath3 qcd and ignore the production of quark pairs . if one is bothered by considering quark - quark scattering while ignoring quark productions , one can consider instead gluon - gluon scattering , and as mentioned above , the result would be the same . in a subsequent paper@xcite we shall use the cut diagrams to discuss the reggeization conjecture in higher orders . in sec . 2 the result of feynman diagram calculations to @xmath4 is reviewed . how these diagrams combine to give a sum much simpler than the individuals will be discussed in some detail . in sec . 3 , we begin to examine more carefully these cancellations by using sum rules . in sec , 4 , the method of cut diagrams is explained , and in sec . 5 , this method is applied to 4th and 6th order calculations to demonstrate the savings effected by this method . leading - log calculations in qcd , at high c.m . energy @xmath5 and comparatively small momentum transfer @xmath6 , are discussed , among others@xcite , in the book of cheng and wu @xcite . citations to the original literature can also be located there . we review in this section the calculation of quark - quark scattering to the 6th order done in this standard way . -0 cm 3 truein -2 cm.1 cm the relevant feynman diagrams are given in fig . the 2nd order diagram is labelled a , the 4th order diagrams labelled b1 and b2 , and the 6th order diagrams c1 to c21 . this last labelling is identical to the ones used in fig . 12.7 of ref . @xcite@xcite . under the interchange of the mandelstam variables @xmath7 and @xmath8 , the spacetime parts of these diagrams are either self conjugate , or one goes into another . for example , a@xmath9a , b2@xmath9b1 , c2@xmath9c1 , c15@xmath9c16 , c17@xmath9c19 , and c18@xmath9c20 , under @xmath10 . normalizing the dirac spinors to @xmath11 , and designating the fermion mass and gauge coupling constant as @xmath12 and @xmath13 , with @xmath14 , the t - matrix element @xmath15 receives the following contributions from the individual diagrams to @xmath16 [ 3,4]@xcite : @xmath17 \!\cdot\!{\bf g}_3\nonumber\\ \overline{c2}&=&-\beta^2({\ln^2s})\left [ { 1\over 2 } \delta^2i_2 ^ 2-j_2i_2\right ] \!\cdot\!({\bf g}_3+c^2{\bf g}_1)\nonumber\\ c3&=&+\beta^2\{\ln^2\left(se^{-\pi i}\right)-{\ln^2s}\ } { 1\over 4}j_2i_2\!\cdot\!({\bf g}_3-c{\bf g}_2)\nonumber\\ & = & c4=c5=c6\nonumber\\ c7&=&-\beta^2\ln^2\left(se^{-\pi i}\right){1\over 4}j_2i_2\!\cdot\!(-c{\bf g}_2)\nonumber\\ & = & c8=c9=c10\nonumber\\ c11&=&+\beta^2({\ln^2s}){1\over 4}j_2i_2\!\cdot\!(-c{\bf g}_2-c^2{\bf g}_1)\nonumber\\ & = & c12=c13=c14\nonumber\\ c15&=&-\beta^2(\ln s ) 2j_3\!\cdot\!{\bf g}_4\nonumber\\ c16&=&-\beta^2(\ln s ) 2j_3\!\cdot\!({\bf g}_4-{\bf g}_3 + 3c{\bf g}_2+c^2{\bf g}_1)\nonumber\\ c17&=&+\beta^2(\ln s)(j_3+\pi ii_3)\!\cdot\!({\bf g}_4+c{\bf g}_2)=c18\nonumber\\ c19&=&+\beta^2(\ln s)(j_3-\pi ii_3)\!\cdot\!({\bf g}_4-{\bf g}_3 + 2c{\bf g}_2)=c20\ .\end{aligned}\ ] ] part of @xmath18 has been combined with @xmath19 to form @xmath20 , and the remaining part of @xmath18 has been combined with @xmath21 to form @xmath22 . note that the result for different diagrams is consistent with their @xmath10 character , under whose exchange the @xmath23 amplitude responds with the swap @xmath24 . see app . a and b for a brief discussion of how these amplitudes are computed . the two - dimensional vector @xmath25 in the transverse direction is the momentum transfer . the functions @xmath26 and @xmath27 are defined as follows : @xmath28 the functions @xmath29 are denoted respectively by @xmath30 in ref . the infrared divergences of these integrals can be regulated by a mass , either put in by hand or via the higgs mechanism . this regulation discussed in the literature [ 39 ] does not affect the following discussions so we shall ignore it . there is however also an ultraviolet divergence in the integral defining @xmath31 , but it turns out that this function disappears in the _ sum _ of the sixth order diagrams so it causes no trouble either . the factors @xmath32 in ( 2.1 ) are the color factors , with @xmath33 for @xmath3 colors . as defined in @xcite , @xmath34 are respectively the color factors for the diagrams a , b1 , c1 , and c15 . they can be computed by an elegant graphical method @xcite from the @xmath3 commutation relation and identities @xmath35=if_{abc}t_c\ , \quad f_{abc}f_{abd}=2c\delta_{cd}\ , \quad i^3f_{adg}f_{bed}f_{cge}=cif_{abc}\ .\end{aligned}\ ] ] hence the combination of color factors @xmath32 given in ( 2.1 ) for the various diagrams remain valid whatever the color of the quark is , although @xmath32 themselves would be different for different color of the quark . see app . a for a brief discussion on the computation of the color factors . the sum of all the terms in ( 2.1 ) , from @xmath36 to @xmath37 , is @xmath38 \!\cdot\!{\bf g}_1 + { 1\over 2 } g^2i(i_2 - 2\beta ci_3\ln s)\!\cdot\!{\bf g}_2\nonumber\\ & + & g^2i\beta\ln s\left [ i_3-{1\over 2}\delta^2i_2 ^ 2\right ] \!\cdot\ ! { \bf g}_3 + g^4{1\over 6}i_3\!\cdot\ ! { \bf g}_4\ , \end{aligned}\ ] ] with @xmath39 it is important to note the various cancellations taking place to make the sum ( 2.4 ) vastly simpler than the individual terms appearing in ( 2.1 ) . for example , 1 . in the fourth order , the leading term proportional to @xmath40 is cancelled out between @xmath41 and @xmath42 in the color amplitude proportional to @xmath43 , though not in @xmath44 . 2 . in the sixth order , the leading @xmath40 contributions to @xmath45 from @xmath46 to @xmath37 also add up to zero . the expressions given in ( 2.1 ) are not accurate enough to deal with the subleading terms . the term in ( 2.4 ) proportional to @xmath45 is obtained separately from the eikonal formula . 3 . as a result of these cancellations , the energy dependence and the @xmath3 ( or @xmath47 ) dependence of the @xmath44 amplitude is @xmath48 , and those of @xmath49 are respectively @xmath50 , @xmath51 , @xmath52 . these dependences can be summarized all at once by introducing a different notation for the color factors . we shall use the notation @xmath53 to denote a color factor with @xmath54 parallel vertical lines connecting the two fermions , and @xmath55 parallel horizontal lines joining any two of the vertical gluon lines . we shall also write @xmath56 simply as @xmath57 . the relations with the color factors @xmath32 are @xmath58 , and @xmath59 . the @xmath60 and @xmath40 dependences of @xmath53 in ( 2.4 ) are then given by @xmath61 for a diagram of order @xmath62 . we shall refer to such dependences as _ regge - like _ , for the reggeization of the scattering amplitude to be discussed later relies critically on this feature of the scattering amplitude . note from ( 2.1 ) that contributions from individual diagrams are not regge - like . only the sum is . simplification in transverse - momentum dependences also occur in the sum . the simple integrals @xmath63 survive , but the complicated integral @xmath64 and the divergent integral @xmath65 do not appear in the sum . this cancellation is highly nontrivial because both of them contribute different amounts to different color amplitudes . more specifically , 5 . the function @xmath66 appears in all the color amplitudes @xmath67 and @xmath45 in diagrams c15 to c20 . those in @xmath68 actually get cancelled out in the sum , but its presence in the @xmath44 amplitude survives . however , since this term is of order @xmath69 , it is negligible compared to terms of order @xmath70 appearing in the @xmath44 amplitudes of @xmath22 and @xmath71 to @xmath72 , it can be ignored in the leading - log result displayed in ( 2.4 ) . @xmath31 appears in the color amplitudes @xmath73 and @xmath74 in individual diagrams c1 to c14 and all these appearances get cancelled out . as a result of these cancellations , @xmath16 acquires a very simple interpretaion in terms of reggeized gluon exchanges . these exchanges are constructed in such a way to ensure @xmath1-channel unitarity [ 3,7 ] . let us denote the reggeon propagator by @xmath75 this reduces to the ( transverse part of the ) ordinary propagator @xmath76 for small @xmath77 . similarly , let us denote the reggeized version of @xmath27 by @xmath78 indicating the exchange of @xmath79 reggeons . then to order @xmath80 in @xmath23 , we can write @xmath81 \nonumber\\ & + & { g^4\over 3!}r_3(\delta , s)\!\cdot\!{\bf f}_3\ .\end{aligned}\ ] ] in other words , the @xmath82 , and @xmath83 components looked precisely like diagrams a , b1 , c1 , and c15 respectively , but with the vertical gluons replaced by their reggeized version whose propagators are given in ( 2.6 ) , _ and _ with all longitudinal - momentum integrations omitted . to interpret it this way for @xmath84 we need to know the lipatov - dickinson vertex @xcite describing how elementary gluons are produced and absorbed from the reggeized gluons . this remarkable simplicity and regularity led to the conjecture @xcite that the reggeized formula ( 2.8 ) , suitably generalized , is the correct high energy limit to all perturbative orders . this conjecture is very difficult to verify on account of the shear complexity in higher order calculations . for qcd in the 8th order it is simply not manageable without simplifying assumptions . if one assumes all cancellations occured up to @xmath4 will also occur in higher orders , the final result can be extracted from a _ relatively _ small set of diagrams , then it is reported that this reggeization conjecture is true to 8th and 10th orders @xcite . even so these heroic calculations are so lengthy and complicated that to our knowledge the full details have never been published . we shall discuss in the next two sections a method of using _ cut diagrams _ to sum up the feynman diagrams , a method in which most of the cancellations discussed in the last section are automatically built in . this simplification shortens the computations and makes it possible to study higher order diagrams . in this paper we shall discuss how it simplifies the calculations up to @xmath4 . in a subsequent paper @xcite we shall discuss how it helps to verify part of the reggeization conjecture to all orders . to prepare for the grounds for both , we discuss here the relevant kinematical features of high energy scattering which enables this new method to work . we will assume the colliding beams in their c.m . system to be directed along the z direction . in lightcone coordinates , @xmath85 , the components of a four - vector are labelled in the order @xmath86 , with the two - dimensional vector @xmath87 lying in the transverse x y plane . in this notation , the incoming fermion momenta are @xmath88 and @xmath89 , in which their mass @xmath12 has been neglected . the outgoing fermion momenta are approximately given by @xmath90 and @xmath91 . see fig . 1 . suppose @xmath79 gluons are hooked up to the upper fermion line as shown in fig . the initial and final fermions are on - shell but the gluons can be off - shell , though with an amount of energy far less than @xmath5 . at high energy , the numerator of the propagator can be approximated by @xmath92 provided the dirac spinors are normalized to @xmath93 . with that , the dominant current @xmath94 at high energy is just its translational part @xmath95 . this shows that the spin content at high energy is unimportant . all that it does is to enforce helicity conservation of the fermion , and to produce a factor @xmath96 at each vertex together with an overall normalization factor of @xmath97 . for most of the discussions below we shall ignore this qed factor and concentrate on the contribution from the denominators , _ i.e. , _ the corresponding scalar theory . -.2 cm 3.2 truein 0 cm.1 cm the denominator of the @xmath54th inverse propagator is @xmath98 where @xmath99 . the scalar amplitude in fig . 2 is given by @xmath100 times @xmath101\equiv -2\pi i\delta\left ( \sum_{j=1}^nx_j\right)\prod_{i=1}^{n-1}{1\over\sum_{j=1}^ix_j+i \epsilon}\ .\end{aligned}\ ] ] note that a momentum conservation @xmath102-function for the negative components ( together with an explicit factor @xmath103 ) has been incorporated . in eq . ( 3.3 ) the ordering of the gluon lines from left to right is @xmath104 $ ] . if they are ordered differently , say @xmath105\equiv v$ ] , then the corresponding amplitude is @xmath106\equiv a[v ] \equiv -2\pi i\delta\left ( \sum_{j=1}^nx_{v_j}\right)\cdot\prod_{i=1}^{n-1}{1\over\sum_{j=1}^ix _ { v_j } + i \epsilon}\ .\end{aligned}\ ] ] it is possible to compute sums of feynman diagrams without much of the delicate cancellations discussed in sec . 2 , by using the _ cut diagrams _ which we shall describe in this section and the next . the derivation of cut diagrams relies on two exact combinatorial formulas for the quantity @xmath107 $ ] in ( 3.4 ) : the _ factorization formula _ and the _ multiple commutator formula _ derived in ref . we shall discuss the former in this section , and the latter in the next section . consider an ordering of @xmath108 gluon lines : @xmath109\equiv v_i$ ] . we shall use the notation @xmath110 to denote the _ set _ of _ all _ orderings of the @xmath111 gluon lines , _ provided _ the relative orderings of lines within each @xmath112 are maintained . the number of orderings in this set is given by the multinominal coefficient @xmath113 . for example , if @xmath114 , v_2=[24]$ ] , then @xmath115 consists of the @xmath116 orderings [ 13524 ] , [ 13254 ] , [ 13245 ] , [ 12354 ] , [ 12345 ] , [ 12435 ] , [ 21354 ] , [ 21345 ] , [ 21435 ] , and @xmath117 $ ] . we shall use the notation @xmath118\end{aligned}\ ] ] to denote the sum of all amplitudes for the gluon orderings in the set . the factorization formula @xcite then states that @xmath119\ \end{aligned}\ ] ] in particular , if each set @xmath120 $ ] consists of only one gluon line labelled by @xmath121 , then @xmath122 is the set of _ all _ orderings of the @xmath12 gluon lines . in that case the factorization formula reduces to the well - known _ eikonal formula _ other special cases of this formula have also been discovered before@xcite . it is useful to adopt an alternative notation for the right hand side of ( 4.2 ) to denote @xmath123 $ ] simply as @xmath124 $ ] . this notation is suggestive because the vertical bar can be interpreted graphically as a cut in the fermion propagator between the last gluon line of @xmath112 and the first gluon line of @xmath125 . for a cut propagator , instead of the usual factor @xmath126 , we have @xmath127 . this notation is also convenient because it makes ( 4.2 ) deceptively simple . it now reads @xmath128 $ ] ; we simply have to change the semicolons to vertical bars . cut propagators are not limited to tree diagrams like fig . 2 . the offshell gluons can be connected to other diagrams to form a composite diagram that inherits the original cuts . the cut diagrams so formed are similar to but different from the cutkosky cut diagrams . similar because we have the same factors for the cut propagators . different because the cuts here occur only on fermion lines whereas in a cutkosky diagram they can occur on any line . moreover , via ( 4.2 ) , our cut diagram represents a sum of @xmath113 ( uncut ) feynman diagrams , with their real and imaginary parts fully included , which is unlike the cutkosky diagrams in which only the imaginary part or the discontiuity is represented . it is clear from ( 4.2 ) that the factorization formula can be thought of as a sum rule , to represent sums of feynman diagrams as cut diagrams . as will be discussed in appendix b , a cut diagram is easier to compute than an uncut diagram . in this way not only it is unnecessary to compute the individual diagrams first , the cut diagram representing the sum is actually easier to compute than just one single feynman diagram . we shall now apply the factorization formula to compute sums of amplitudes quoted in sec . 2 . we shall see that the @xmath40 factor and the @xmath129 functions that get cancelled out in the sum never once appear in the cut diagrams . in what follows we shall use the notation @xmath130 to denote the spacetime part of @xmath41 ( without the color factor @xmath43 ) . similar notation will be used for the spacetime part of other diagrams as well . -.3 cm 3.0 truein -0 cm.2 cm in the 4th order , the @xmath43 coefficient in @xmath16 is given by @xmath131 . using the eikonal formula @xmath132+a[21]=a[1\|2]$ ] on the upper fermion line ( see fig . 3 ) , @xmath131 is reduced to a cut diagram that can be easily calculated ( see appendix b for all the calculations ) , yielding the correct result @xmath133 given in ( 2.1 ) and ( 2.4 ) . similary , the sum rule @xmath134 $ ] applied to the six horizontal ladder diagrams in the 6th order , yields a cut diagram ( see fig . 4 ) which gives the sum of @xmath135 to @xmath136 to be @xmath137 . this is the correct coefficient of @xmath45 given in ( 2.4 ) . note that neither @xmath40 nor @xmath64 ever appears . -0 cm 3 truein 0 cm.1 cm -0 cm 2.7 truein 0 cm.1 cm -0 cm 3.7 truein .1 cm.3 cm let us now look at the cancellation of @xmath66 in the @xmath43 and @xmath74 color factors , as discussed in point ( 5 ) of sec . 2 . according to ( 2.1 ) , the coefficient of @xmath138 from diagrams c15 to c20 is given by @xmath139 . using @xmath140 $ ] on the upper fermion line as in fig . 5 , this sum is given by a cut diagram which can be evaluated to be @xmath141 . again @xmath64 does not appear in the cut diagram . similarly , the coefficient of @xmath142 from diagrams c15 to c20 is given by ( 2.1 ) to be @xmath143 . this is equal to twice the previous sum , plus @xmath144 . to use the sum rule , this requires an expression ( see fig . 6 ) for @xmath145+a[213]+a[132]$ ] . it is easy to see that this is equal to @xmath146 , and by the factorization formula , also equal to @xmath147-a[23\|1]+a[13\|2]$ ] . there may seem to be very little gained by replacing the sum of three terms by the sum of another three terms , but this is not so . we are replacing the sum of three uncut diagrams by the sum of three cut diagrams . the computation of cut diagrams is much easier than the computation of uncut diagrams . for one thing the complicated function @xmath64 that appears on all the three uncut diagrams but disappears from their sum never appears in any of the three cut diagrams . evaluating the cut diagrams , the coefficient of @xmath142 from c15 to c20 is @xmath148 , agreeing with the answer given in ( 2.4 ) . the method introduced in the last section has a serious shortcoming . it does not tell us which cut diagrams to compute without detailed considerations of the kind carried out there . in this section we discuss a remedy for this shortcoming with the help of the _ multiple commutator formula _ [ 12 ] . so far we have been treating fig . 2 mostly as a qed or a scalar amplitude . for qcd the non - abelian color matrices @xmath149 have to be incorporated . instead of ( 3.4 ) the amplitude is now @xmath150=a[v_1v_2\cdots v_n]t[v_1v_2\cdots v_n ] \equiv a[v]t[v]\equiv a[v]\ , \end{aligned}\ ] ] where @xmath151=t_{v_1}t_{v_2}\cdots t_{v_n}$ ] . what we want is a formula for the sum of the @xmath152 permuted gluon orderings , @xmath153 $ ] . for qed , where we can take all @xmath154 , this is simply the eikonal formula , so what we need is the non - abelian generalization of it . this is the multiple commutator formula @xmath155t[v]=\sum_{v\in s_n}a[v_c]t[v_c']\ .\end{aligned}\ ] ] it expresses the sum of @xmath107t[v]$ ] in terms of sums over the corresponding cut amplitude @xmath156t[v_c']$ ] . compared to the eikonal formula this looks complicated ; instead of a single term on the right hand side we have now a sum over @xmath152 terms . the complication is inevitable because we are attempting to sum up amplitudes for _ every _ color . however , we shall see that many of these terms are actually zero , and moreover , the cut diagrams on the right are considerably simpler to evaluate than the uncut diagrams on the left . again delicate cancellations will largely be avoided as before . what remains to be described is what the cut diagram @xmath157 that corresponds to the feynman diagram @xmath158 is , as well as what the amplitudes @xmath156 $ ] and @xmath159 $ ] are . given a @xmath160 $ ] , start from the rightmost number @xmath161 and proceed leftward until one comes to the first number less than @xmath161 . put a cut just to the right of this number . then start from this number and proceed leftward again until one comes to the first number that is less than this number , and another cut is put just to the right of this new minimum number . continue this way until the end and we have constructed the cut diagram @xmath157 . for example , for @xmath162 , the 2 cut diagrams are @xmath163_c=[1\|2]$ ] and @xmath164_c=[21]$ ] . for @xmath165 , the six cut diagrams are @xmath166_c=[1\|2\|3]$ ] , @xmath167_c=[21\|3]$ ] , @xmath168_c=[31\|2]$ ] , @xmath169_c=[1\|32]$ ] , @xmath170_c=[231]$ ] , and @xmath171_c=[321]$ ] . to each cut diagram we associate a spacetime cut amplitude @xmath156 $ ] as described in the last section . namely , it is given by ( 3.4 ) except the propagator at a cut is replaced by @xmath172 . the complementary diagram @xmath173 of a cut diagram @xmath157 is obtained as follows . if a cut appears between two numbers in @xmath157 , then there will be no cut between the same two numbers in @xmath173 , and vice versa . for @xmath162 , the complementary cut diagrams are @xmath174'=[12]$ ] and @xmath164'=[2\|1]$ ] . for @xmath165 , the complementary cut diagrams are @xmath175'=[123]$ ] , @xmath176'=[2\|13]$ ] , @xmath177'=[3\|12]$ ] , @xmath178'=[13\|2]$ ] , @xmath170'=[2\|3\|1]$ ] , and @xmath171'=[3\|2\|1]$ ] . when no cut appears in @xmath173 the color factor @xmath159 $ ] is simply @xmath179=t_{v_1}t_{v_2}\cdots t_{v_n}$ ] . if a cut appears between @xmath121 and @xmath180 , then the product @xmath181 is replaced by their commutators @xmath182 $ ] . if two or more consecutive cuts appears , then the corresponding product of @xmath2 s is replaced by multiple commutators . for example , @xmath183=[t_2,t_1]t_3 $ ] , @xmath184=[t_2,[t_3,t_1]]$ ] , and @xmath185=[t_4,[t_3,[t_2,t_1]]]t_5 $ ] . we shall compute the cut diagrams to @xmath4 to demonstrate the simplifications obtained therefrom . -0 cm 3.7 truein 0 cm.1 cm using ( 5.2 ) on the upper fermion , we obtain a set of cut diagrams as shown in fig . 7 . this set is not unique as others can be obtained from a different labelling of the gluon lines . in the horizontal ladder diagrams b1 , b2 , c15c20 , we choose the planar diagrams b1 and c15 to be the ones whose upper lines are completely cut . this fixes the labelling [ 12 ] for b1 and [ 123 ] for c15 as shown , and it in turn determines how diagrams b2 and c16 to c20 are to be cut . using computational methods discussed in app . a and b , these diagrams yield @xmath186 similarly , we choose to cut the line of the planar diagram c1 to obtain @xmath187 \!\cdot\!{\bf g}_3&\nonumber\\ \overline{c2}_c&=-\beta^2({\ln^2s})\left [ { 1\over 2 } \delta^2i_2 ^ 2-j_2i_2\right ] \!\cdot\!c^2{\bf g}_1&\end{aligned}\ ] ] the twelve diagrams c3 to c20 can be divided into four groups of three , each giving identical contributions , so it is necessary to consider only one of these groups . the group of c3 , c7 , c12 have been chosen for that purpose . there is a symmetry between gluon lines 2 and 3 so we may double this group and consider it as a sum of six feynman diagrams . by applying the multiple commutator formula , the six cut diagrams shown in fig . 7 are obtained . their values are @xmath188 the expressions in ( 6.1)(6.3 ) should be compared with those of the uncut diagrams , eq . it should also be compared with the sum found in eq . ( several points can be noted from these comparisons : 1 . @xmath40 factors that get cancelled in the sum of the feynman amplitude ( see points ( 1 ) to ( 3 ) in sec . 2 ) never even appear in ( 6.1)(6.3 ) . cancellation of this kind is automatically built into the cut diagram formalism . 2 . the tranverse function @xmath64 appears only in @xmath189 in eq . this expression survives the sum but can be ignored compared to the contribution from @xmath190 . in other words , as opposed to the feynman amplitude ( 2.1 ) where @xmath64 appears in many places but most of them are cancelled out at the end ( see points ( 4 ) and ( 5 ) in sec . 2 ) , in the cut amplitude @xmath64 does not appear except when it survives the sum . the cut amplitude is not as successful in cancelling the transverse function @xmath65 ( point ( 6 ) of sec . 2 ) , although there is still an imporvement here over ( 2.1 ) in that @xmath65 appears in fewer places . in fact , it appears in @xmath191 to @xmath192 only when absolutely needed to cancel its previous appearance in @xmath193 and @xmath190 . in order for @xmath65 to disappear completely it is necessary to combine diagrams with triple and four gluon vertices together using the lipatov - dickinson vertex @xcite . the technique of cut diagrams by itself , which deals mainly with the fermion lines , is not sufficient for that purpose . 4 . other than the @xmath65 complication mentioned above , the summands of the final answer ( 2.4 ) appear directly in the cut amplitudes . in that sense the cut amplitudes are as economical and as simple as they can ever be . in particular , the regge - like feature mentioned in point ( 3 ) of sec . 2 is present already in individual cut diagrams . this research is supported in part by the by the natural science and engineering research council of canada , and the fonds pour la formation de chercheurs et laide la recherche of qubec . yjf acknowledges the support of the carl reinhart foundation , and ohr thanks the ministry of culture and higher education of iran for financial support . the method to compute the color factors graphically @xcite is briefly reviewed here . it is suitable for both cut and uncut diagrams . -0 cm 2.7 truein -0.5 cm.3 cm the basic tool is the graphical identities depicted in figs . 8 and 9 , which are nothing but the commutation relation and identities in ( 2.3 ) . the commutation relation is valid for all representations of the color matrix @xmath149 , hence fig . 8 remains true when the quark line is replaced by a gluon line . -0.3 cm 3.2 truein -.5 cm.1 cm -1 cm 3.2 truein -.8 cm.1 cm as an example , the computation of the color factor for the uncut diagram c19 in eq . ( 2.1 ) is carried out in fig . 10 , and the computation of the color factor for the cut diagram c19@xmath194 in eq . ( 6.1 ) is carried out in fig . 11 . for the latter , we have to remember to use the _ complementary _ cut diagram for the color factor , and not the cut diagram itself shown in fig . 0 cm 3.5 truein 0 cm.2 cm the color factors in eqs . ( 2.1 ) , ( 6.1)(6.3 ) can all be obtained in this way . high energy limit of a feynman diagram can be computed either in the feynman - parameter representation@xcite , or directly in momentum space using lightcone coordinates@xcite . we follow the latter approach and review briefly the main points . for a more detailed discussion see ref . it serves well to remember at this point that the bottom fermion carries mostly the ` @xmath195 ' momentum and the top fermion line carries mostly the ` @xmath196 ' momentum . the procedure to follow for the computation is simple : ( i ) carry out the ` @xmath196 ' momentum integration using residue calculus ; ( ii ) then carry out the ` @xmath195 ' momentum integration to produce the @xmath40 dependence . the transverse momentum integrations are never explicitly carried out , which is why the answers are all expressed in terms of functions like @xmath27 and @xmath26 given in eq . ( 2.2 ) . step ( i ) requires a knowledge of the location of poles for each propagator , which in turn depends on the choice of loop momenta @xmath197 . suppose @xmath198 is the momentum flowing through a propagator , with @xmath199 being some combination of external momenta and @xmath200 some integer coefficients . the propagator @xmath201 has a pole in the integration variable @xmath202 if @xmath203 . this pole is located in the upper half plane if @xmath204 is negative , and in the lower half plane if it is positive . a way to keep track of the sign is to draw a _ flow diagram _ for the ` @xmath195 ' momentum , with arrows indicating the direction of the ` @xmath195 ' flow . see fig . 12 for some illustrative examples . the arrows around a loop pointing one way ( clockwise or counter clockwise ) have their poles in one half plane , and those pointing the other way have poles in the opposite half plane . the arrows in the flow diagrams must obey momentum conservation , and one may also assume they do not go around in closed loops . otherwise the poles will all be in the same half plane , and the integral is zero if the contour is closed in the opposite half plane . with these constraints we see in fig . 12 that a one - loop diagram allows only one flow path , but for two and more loops there is bound to be more than one flow diagrams because these constraints simply can not fix the direction of the ` @xmath195 ' flow on a boundary line of two loops . we shall always close integration contours in the lower half planes , and indicate the poles so enclosed by a cross ( x ) in the flow diagram . for a _ diagram , the @xmath23-matrix element is equal to the product of @xmath206 propagators @xmath207 , integrated over the @xmath208 loop momenta @xmath209 , with an extra numerical factor @xmath210^\ell$ ] . each ` @xmath196 ' integration produces a factor @xmath103 , the @xmath23-matrix is equal to @xmath211 where @xmath212 is the scaled ` @xmath195 ' momenta , @xmath213 is the measure for transverse momentum integration , and @xmath214 is either the propagator evaluated at the x poles or the residue of the x pole divided by @xmath5 . each summand in ( b1 ) corresponds to a flow diagram with a pole taken from the lower half plane ( an x pole ) of each loop . a flow diagram may have more than one set of x poles , in which case the sum is taken over all possible sets . for example , the one - loop diagram b1 in fig . 1 has only one flow path ( fig . 12(a ) ) and one set of x poles . the two - loop diagram c17 of fig . 1 has two flow paths , figs . 12(b ) and 12(c ) . the first flow diagram has two sets of x poles , and the latter flow diagram has one set of x poles . we shall see later that in the leading log approximation , we may discard the @xmath215 dependence on the upper fermion line and the accompanied poles , in which case 12(b ) is also left with one set of x poles . let @xmath217 . every @xmath218 is a linear combination of @xmath219 , and in the case of a propagator along the bottom fermion line , @xmath220 is also involved . the sign of each @xmath218 is fixed by the direction of arrows in the flow diagram . if there are @xmath221 cut lines , then the last @xmath221 @xmath219 will be chosen to be equal to @xmath222 of these cut lines . the @xmath206 indices @xmath54 will now be divided into three sets , @xmath223 for indices from 1 to @xmath208 labelling the internal lines with an x pole , @xmath224 for indices labelling propagators on the top fermion line , and @xmath12 for the rest . then @xmath225 , and @xmath226 can be expressed as linear combinations of @xmath227 , _ @xmath228 within the leading log approximation , one has @xmath229 the ` @xmath195 ' momentum flows mainly along the bottom fermion line , with very little seeping out to avoid a substantial mixing with the ` @xmath196 ' momentum coming from the top fermion line , for a finite mixture of these two at a propagator would make it proportional to @xmath1 and therefore negligible . this means the dominant contribution to @xmath23 comes from regions where all @xmath219 are small . since @xmath230 for @xmath231 , on account of the @xmath102-functions of the cut lines , the region of ` @xmath195 ' integration can be roughly divided into regions where the remaining @xmath232 @xmath219 s are strongly ordered , and regions where two or more of these @xmath219 s are of the same order of magnitude . we shall label any one of the latter regions by @xmath233 , and the former regions by @xmath234=\{1\gg x_1\gg x_2\gg\cdots \gg x_\ell'\ge a / s\}$ ] and its permutations . the @xmath235-dependence of @xmath236 in @xmath234 $ ] is of the form @xmath237 , so the @xmath235-integral encountered in ( b1 ) is @xmath238 where @xmath239 and @xmath240 is determined by how many times the sum @xmath241 reaches zero by varying @xmath242 from 1 to @xmath243 . clearly @xmath244 , and the only way for @xmath245 is to have all @xmath246 , in which case we will call the @xmath40 dependence of @xmath23 _ saturated_. otherwise it is said to be _ unsaturated_. for the uncut diagrams , @xmath247 and @xmath248 . we see in ( 2.1 ) that all the diagrams except for the 6th order horizontal ladder diagrams c15 to c20 are saturated . for the cut diagrams in ( 6.1 ) to ( 6.3 ) , only @xmath189 is unsaturated , but this diagram does not contribute to the final sum because of its subdominance . the integral in an @xmath233 region is like one in an @xmath249 region with @xmath243 reduced . for example , suppose @xmath250 to @xmath251 are roughly equal but @xmath252 to @xmath253 ( @xmath254 ) are strongly ordered as in @xmath255=r''$ ] . then the volume element in the last three variables in spherical coordinates is @xmath256 , so the integration region is effectively @xmath257 but with @xmath253 replaced by @xmath258 . since @xmath259 , integrals in @xmath233 will never lead to saturation . cut diagrams are easier to compute than uncut feynman diagrams for three reasons . first and most trivial , a cut line contains a @xmath260 which makes the corresponding ` @xmath195 ' integration simpler to carry out . second and more importantly , the @xmath102-function demands the absence of the ` @xmath195 ' momentum across this line , so the cut line is cut also in the sense of being an open electric circuit . this generally reduces the number of possible flow diagrams and makes the corresponding integral easier to saturate . for example , each of the two - loop diagrams in fig . 13 has only one flow path . thirdly , the flow pattern often leads to vanishing cut diagrams . for example , both diagrams 13(c ) and 13(d ) are zero from the ` @xmath196 ' integration around the loop ( 12345 ) because the ` @xmath195 ' flows around that loop are all in the same direction . this accounts for the equality @xmath261 in eq . ( 6.1 ) . the cost of this simplicity is the presence of _ tricky diagrams _ , which are cut diagrams that are apparently logarithmic divergent at large @xmath202 . this happens whenever there is a loop with only one arrow present . arrows on the top fermion line should not be counted for this purpose because their approximate propagators @xmath262 no longer carry any @xmath202 . 13(a ) and 13(b ) are examples of such diagrams . this apparent divergence is produced by the approximation of replacing the inverse propagators of an upper line by @xmath263 , thus losing some @xmath215 factors needed for convergence . to regulate it we must replace the @xmath102-function of the cut line by a smeared @xmath102-function , thereby allowing a small amount of ` @xmath195 ' momentum to flow through , and in the process restoring the lost @xmath215 factor . if we do so to 13(a ) , the @xmath264 integration is no longer divergent , it will have exactly the same flow path as in 12(a ) and this is indicated in 13(a ) by the dotted arrows , so a pole can be taken at the bottom fermion line as shown . the result is given by ( b1 ) to be @xmath265 , the extra @xmath266 is there because the @xmath235 integration is bounded by the flow path in 12(a ) to be between 0 and 1 , so only half of the @xmath267 is integrated . when multiplied by the qed vertex and normalization factor @xmath268 , one obtains @xmath269 in ( 6.1 ) . there is another way to compute fig . this is to recognize the fact that the cut makes it symmetric in lines 1 and 2 , so we may replace the diagram by half the sum of it and its crossed diagram . using the sum rule ( 3.4 ) , @xmath132+a[21]=a[1\|2]$ ] , a cut can be produced at the bottom line , and the resulting factor @xmath270 makes the @xmath271 integration convergent . see fig . moreover , the result of this ` @xmath196 ' integration is @xmath103 , exactly the same as if we were to do it by residue calculus . the extra factor of @xmath266 obtained in the last paragrpah now emerges because the double - cut diagram on the right of fig . 14 is the sum of two diagrams . now we come to the tricky diagram 13(b ) , which unlike the one - loop case is much more difficult to compute by regulating the @xmath102-functions . the reason is that there are now two cut lines , the relative magnitude of the small ` @xmath195 ' flows matters , and that produces once again two flow diagrams like 12(b ) and 12(c ) . the poles are no longer at the bottom fermion line and the computation is no simpler than the uncut diagrams . we must then compute it by the second method . with the double cuts in 13(b ) , it is symmetric in all the @xmath272 variables . insisting on this symmetry , 13(b ) is equal to @xmath273 times the sum of 6 diagrams , obtained by permuting the bottom gluon lines in all possible ways . from the factorization formula @xmath134 $ ] used on the bottom line , the result is equal to a diagram with all its fermion propagators cut . so 13(b ) is given by ( b1 ) to be @xmath274 . incorporating the extra qed factor @xmath275 for the @xmath23 matrix , we obtain the contribution of @xmath276 to @xmath277 shown in ( 6.1 ) . we conclude this appendix by discussing the remaining expressions in eqs . ( 6.1)(6.3 ) . we shall use the notation @xmath278 to denote the spacetime contribution to @xmath279 , etc . then @xmath280 , @xmath281 , @xmath282 , and @xmath283 can be obtained from ( 2.1 ) . the zero of @xmath284 in ( 6.1 ) is due to the vanishing of the color factor as shown in fig . next , we compute @xmath285 . the flow path of this is shown in fig . 13(f ) , with the dotted arrows indicating the small regulating current which is allowed to flow only in the direction shown . the calculation is identical to the uncut diagram @xcite , except that the factor @xmath286 is replaced by @xmath287 ( @xmath103 from the cut , @xmath40 because only one uncut line is left on top , and @xmath266 because only half of the @xmath102-function is integrated ) . the result can then be read off from ( 2.1 ) to be @xmath288 which gives rise to the expression in ( 6.3 ) . finally , we must show that @xmath289 and @xmath290 . the former can be seen from the flow path in fig . 13(g ) , where the arrows around the small triangle goes around in a closed loop . the latter is so because the scalar diagram in fig . 7 for @xmath291 is symmetrical in lines 2 and 3 but the triple gluon vertex is antisymmetrical in these two lines . nieh and y.p . yao , phys . 32 ( 1974 ) 1074 ; phys . d13 ( 1976 ) 1082 ; b. mccoy and t.t . wu , phys . 35 ( 1975 ) 604 ; phys . d12 ( 1975 ) 3257 ; d13 ( 1976 ) 1076 ; l. tyburski , _ ibid . _ 13 ( 1976 ) 1107 ; l.l . frankfurt and v.e . sherman , yad . 23 ( 1976 ) 1099 [ sov . j. nucl . ( 1976 ) 581 ] ; a.l . mason , nucl . phys . b117 ( 1976 ) 493 . j. bartels , phys . lett . 68b ( 1977 ) 258 ; nucl . b151 ( 1979 ) 293 ; j.b . bronzan and r.l . sugar , phys . d17 ( 1978 ) 585 ; v.n . gribov , zh . eksp . 53 ( 1967 ) 654 ; l.v . gribov , e.m . levin , and m.g . ryskin , phys . 100 ( 1983 ) 1 . note however that for later convenience the femion arrows in fig . 1 and all subsequent figures are drawn from right to left , reflecting the order one writes matrices as well as initial and final states on a piece of paper , whereas in ref . @xcite they are drawn in a more conventional way from left to right . lee and r.f . sawyer , phys . 127 ( 1962 ) 2266 ; j.c . polkinghorne , j. math . phys . 4 ( 1963 ) 503 ; p.g . federbush and m.t . grisaru , annals of physics 22 ( 1963 ) 263 , 299 ; j.d . bjorken and t.t . wu , phys . 130 ( 1963 ) 2566 ; g. tiktopoulos , phys . 131 ( 1963 ) 480 , 2372 ; r.j . eden , p.v . landshoff , d.i . olive , and j.c . polkinghorne , _ ` the analytic s - matrix ' _ , ( cambridge university press , 1966 ) ; j.v . greenman , j. math . ( 1967 ) 1782 ; 8 ( 1967 ) 26 ; c.s . lam , nuovo cimento 62a ( 1969 ) 97 .	a new approach is introduced to study qcd amplitudes at high energy and comparatively small momentum transfer . novel cut diagrams , representing resummation of feynman diagrams , are used to simplify calculation and to avoid delicate cancellations encountered in the usual approach . explicit calculation to the 6th order is carried out to demonstrate the advantage of cut diagrams over feynman diagrams . = -1.5 cm = -1 cm epsf.tex * cut diagrams for high energy scatterings * + +
via imaging , spectroscopy and time - series analysis , the microwave part of the spectrum provides vital information on the properties of flare - accelerated particles and the plasma and the magnetic field in which their emission is formed . although there are considerable complexities in modeling and interpreting the data , microwaves are uniquely rich in diagnostic information and are crucial for flare studies . however , flares are characterised in part by the fact that - for the few minutes of the impulsive phase at least - emission is generated across the entire electromagnetic spectrum . therefore we have the ability to set our microwave observations in context , though in practice the number of flares with excellent multi - wavelength coverage including microwaves remains small . this highlights the ongoing need for continued operation of facilities such as the nobeyama radioheliograph ( norh ) and radio polarimeters ( norp ) in the current era of multi - wavelength observations . in this short paper we will review some multi - wavelength flare observations , focusing on recent results relevant to the flare impulsive phase . these include hints of fine structuring in chromospheric footpoints , very compact footpoint sources , rapid changes in the tilt of the magnetic field during the flare impulsive phase , and hot chromospheric footpoints . in the light of these results we will speculate on what the combination of multi - wavelength and microwave flare data can potentially bring . the ` natural ' partner for microwave emission is hard x - rays , and an extensive review of the strengths of combining radio and hard x - ray data can be found in @xcite . the most directly interpretable signature of non - thermal electrons in solar flares are the non - thermal hard x - rays ( hxrs ) emitted in bremsstrahlung interactions , particularly in the dense plasma of the solar chromosphere . hxr emission from the chromosphere is usually interpreted in terms of the collisional thick target model ( cttm ) , which means that the emission is generated as electrons slow down , under the influence of collisions only , and stop within the target . to interpret a given observation in this way also requires that the slowing - down timescale ( fractions of a second ) is less than the integration time used to make an image or spectrum , which is generally the case . under the assumptions of the collisional thick target model , the total number of electrons that must be injected into the thick target to produce the observed spectrum can be inferred , in a way that does not depend on the precise density structure of the atmosphere . however , with imaging from the ramaty high energy solar spectroscopic imager ( rhessi , * ? ? ? * ) the density structure can be probed in more detail . the systematic offset of source position ( first moment of the source intensity distribution ) as a function of energy when interpreted in the cttm gives a value for the target density scale - height @xcite . the source full - width at half - maximum intensity ( 2nd moment ) as a function of energy can also be measured @xcite , and compared with the predictions of the cttm ; doing so is rather interesting as it is not possible to easily reconcile the modelled and observed behaviour of this quantity . the sources have a much larger observed fwhm than straightforward models of an electron distribution in a ` monolithic ' loop would predict @xcite . this may point to much finer sub - structure in the chromosphere composed of multiple sub - resolution structures of different density scale - heights , as shown in figure [ fig : strands ] , such that the mean source position as a function of energy as expected from the cttm is preserved , but the ` variance ' is increased @xcite . the implication is that chromospheric sub - structures on scales of a tenth of an arcsecond or less could exist . the fine - structuring would presumably have a coronal counterpart . however , the result may also point to problems with the applicability of the cttm . ( 80mm,40mm)kontaretal2010.eps another important result to emerge from rhessi is the inference , using the photosphere s albedo to hxr photons , that the electron angular distribution where the hxr footpoint emission is produced is consistent with one in which as many electrons are traveling ` upwards ' as ` downwards ' @xcite . again this presents a challenge to the cttm . the possibilities of developing such an electron distribution from a combination of scattering ( collisional and non - collisional ) and magnetic mirroring in the chromosphere are still being investigated , but already it is clear that considerable fine - tuning of the electron , field and density parameters is needed to recover the observational results . the ongoing investigations into the causes of these observationally deduced properties of the electron spatial and energy distributions have yet to be concluded , but it is clear that a model of a downward - beamed ` monolithic ' electron beam entering a simple , uniform , collisional chromosphere , is not correct . the electron distribution looks likely to be finely structured in space , and probably also time ( though some average properties can be recovered ) , and may have a complex angular distribution . co - ordinated observations with flares in the optical also make clear that electrons arrive at the chromosphere over small patches . it is not clear that optical footpoints are resolved on a scale of 1@xmath0 @xcite , so the typical optical footpoint is more like @xmath1 , rather than the @xmath2 often used as a ` canonical ' footpoint size ( see section [ sect : optical ] ) . this may have significant implications for the usual ` trap - plus - precipitation ' models used in microwave modeling . [ sect : optical ] optical , or ` white light ' flares , previously thought of primarily as accompanying large flares , are in fact common phenomena , but ill - understood . the flare optical and uv emission contains the majority of the flare s radiation output , and a wealth of spectral lines are available to probe the conditions in the flare chromosphere . for those reasons one would expect this part of the spectrum to have received more attention . however , optical flares are difficult to observe , as they have a low contrast compared to the bright photosphere , and spectroscopic observations require the good fortune to have a spectrometer slit on a flaring kernel at exactly the right time , which has in the past been rare . nonetheless broad - band optical emission is observed in flares from c - class to x - class . the key is to have stable , high - cadence imaging or photometric observations which can be used to perform reliable differencing observations to pick up the faint flare signatures against the bright photospheric background . doing this has revealed that optical footpoints are strongly correlated in space and time with hxr footpoints , and thus with the presence of large numbers of fast electrons @xcite and that optical footpoints are very compact , with areas of around @xmath3 or perhaps less . the energy contained in the white - light continuum is around 70% of the total flare energy , independent of the flare class . the emission mechanism of this broad - band optical radiation is not known . it seems unlikely that it is due to direct heating of the photosphere by electrons accelerated in the corona , as the typical electron energy required to reach the photosphere is around 2 mev , assuming a column mass to the photosphere of 1 gm @xmath4 . an interesting recent analysis by @xcite of a limb flare observed by rhessi and also by one of the stereo spacecraft uses careful triangulation to place both the flare optical emission and the hxr emission at 30 - 50 kev at only 300 km above the photosphere . this corresponds to an electron stopping energy of around 1 mev . ( this single observation has yet to be repeated ) . for the optical luminosity to be produced at or near the photosphere by electrons arriving in a beam from the corona , a large fraction of the electron energy in the injected spectrum would have to be above @xmath5 1 mev this is inconsistent with parameters for the electron distribution derived from hxr measurements . microwave and millimeter observations are informative here , as the emission is generated by electrons in the 100 kev - plus energy range , and it is interesting that these observations suggest that the spectrum may be substantially harder at energies above a few 100 kev ( e.g. * ? ? ? * ) than would be implied by the continuation of the hxr - emitting spectrum . however , this is still not adequate . for example , application of gyrosynchrotron models to microwave emission by @xcite for the large flare sol2002 - 07 - 23t00:35 ( x4.8 ) results in a non - thermal electron density distribution of @xmath6 , corresponding to an electron energy density above @xmath71 mev of @xmath8 if we assume @xmath9 g . this is far too small compared to the energy density of the photospheric plasma ( around @xmath10 ) to produce an observable optical intensity perturbation . for now the observation of @xcite remains a puzzle . another mechanism for producing optical emission is the free - bound continuum that results from the ionisation and recombination in a flare - heated chromosphere ( note , the heating can be , but does not need to be , provided by non - thermal electrons ) . the uv ( balmer and lyman ) continua may also penetrate downwards and backwarm the photosphere , effectively recycling this radiation as optical emission . by indirect means , optically thin emission has been deduced in one flare with an extended white - light ribbon @xcite . this would be consistent with free - bound emission from a hot plasma . the temperature of the emitting plasma should be high enough that the hydrogen is substantially ionised , i.e. above @xmath11k . the brightness temperature of the surrounding non - flaring chromosphere is also in this range , so associated microwave emission would not be visible unless the free - bound emitting plasma were hotter . moreover , in the free - bound model the electron density in the region emitting in the optical may be around @xmath12 @xcite , so any microwaves generated here with frequencies below the corresponding plasma frequency ( @xmath5 28ghz ) will not propagate . although we do not expect to make direct microwave observations of the plasma that radiates the optical emission , the observed optical source properties , coupled with those inferred from hard x - ray spectra , have implications for beam parameters which should be recognised in future microwave modeling . the optical footpoint area is small , and in the event presented by @xcite it is consistent with an unresolved hxr footpoint size . interpreted in the cttm the _ non - thermal _ electron beam density at the location of hxr emission in this event is at least @xmath13 electrons @xmath14 above 20 kev . this is consistent with the value inferred in another x - flare by @xcite from the microwave emission in another large flare , but this time in the corona , where the 17 ghz emission is located . if magnetic mirroring occurs , due to field convergence between the corona and chromosphere ( as one might expect given the increasing evidence for very small footpoints ) then the hxr - generating electrons in the chromosphere would represent only the component that can escape the magnetic trap , giving a lower limit to the overall non - thermal coronal electron density in the coronal loop . on the other hand if the electrons were beamed exactly along the magnetic field then there would be no trapped component and , in a field that diverges into the corona , the coronal beam density requirement would be reduced . but a highly - beamed distribution is at present inconsistent with the angular distributions inferred using photospheric hxr albedo ( section [ sect : hxr ] . ) such high beam densities challenge electron transport models , but may have some bearing on electron spectra relevant to microwave and x - ray comparisons , because beam - return current instabilities can substantially distort the electron spectrum . unless the background plasma through which the beam passes is much denser than the beam , such that the return current speed is low , the beam - return current system will be subject to plasma instabilities causing the majority of its energy to be dissipated as electron and ion heating , via wave generation . analytical considerations suggest that a beam with even a fraction of the density implied by the combination of x - ray and optical observations should , together with its return current system become unstable to the ion - acoustic instability ( when its return - current speed is greater than the ion - acoustic speed ) , with the beam energy redistributed in heating , unless the ambient density is much larger than the beam density @xcite . numerical simulations of the beam - return current system are now very elaborate , including vlasov and pic simulations , in magnetised and non - magnetised scenarios . for example , work by @xcite and indicates that the majority of the beam energy - around 70% - is redistributed as plasma heating , but that the remaining fraction may be available to re - accelerate higher energy electrons . this is interesting for the comparison between hxr and microwave radiation , since it suggests that the lower energy x - ray generating electrons and the higher energy microwave - emitting electrons need not be described by one spectral index . a number of joint studies of hxr and microwave flare conclude that there is a substantial difference in the electron spectral indices at low energy and high energy ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , where microwave emission interpreted as optically - thin gyrosynchrotron radiation implies that electron spectra at high energies ( above @xmath15kev ) are substantially flatter than at low energies . ( 80mm,80mm)petriesudolfig.eps the impulsive phase of a solar flare has detectable consequences for the low solar atmosphere , i.e. the photosphere , apart from the possible photospheric origin for optical flares . it is now well established that significant abrupt ( step - like ) , non - reversing change in the line - of - sight photospheric field occurs for major ( x - class and m - class ) flares , co - temporal with the flare impulsive phase ( e.g. * ? ? ? * ; * ? ? ? an example of such a change is shown in figure [ fig : petriesudol ] . the location of this change can be in the umbra , the penumbra or elsewhere in the active region , and is co - spatial with increases in uv footpoint intensity @xcite . the onset of the goes soft x - ray emission leads the field changes @xcite as do peaks in the uv intensity - by on average 4 minutes @xcite . as remarked on by these authors , the timing pattern is consistent with the flare causing the photospheric field changes , and not vice versa . the hxr footpoints are associated with some , but not all , locations ( e.g. * ? ? ? * ; * ? ? ? * ) , though no systematic study of this has yet been carried out . the change of the line - of - sight field is taken as a sign that the coronal magnetic field is reconfigured by the flare , as shear relaxes or the field shrinks , and that this reconfiguration propagates through the atmosphere to the field s anchor points in the photosphere . the tendency is for the field to become more horizontal @xcite . ( 80mm,80mm)tilt_angle_newcolour.eps the change in the field direction over the duration of the flare impulsive phase corresponds of course to a variation in the viewing angle . since gyrosynchrotron emission from the high - energy non - thermal electrons which are also present during the flare impulsive phase is anisotropic this variation in the viewing angle will influence the microwave emission observed . we do not know of any comparisons yet being made between field changes and variations in the microwave , and indeed it might be difficult to disentangle variations due to changes in the magnetic field direction from those due to variations in the intrinsic properties of the population of emitting electrons . however , we can anticipate the effects . in figure [ fig : tilt ] we show the variation in the microwave spectrum as the viewing angle changes while all other parameters of the source stay the same . vertical lines on this plot show the norh observing frequencies . the emission is calculated for a non - thermal electron density of @xmath16 and electron spectral index of 3 in a compact source of diameter 5 " , thickness 5000 km , and temperature 5mk in a magnetic field of 500 g and ambient density of @xmath17 . this could correspond to a gyrosynchrotron source near the footpoints of a coronal loop , or flare - heated upper chromosphere ( see section [ sect : hotfp ] ) . the three curves correspond to the same field strength , viewed at an angle to the field direction of 20@xmath18 , 40@xmath18 and 80@xmath18 . a decrease in the line - of - sight component of the field strength caused by an increasing field tilt leads to higher microwave intensity and an increase in the peak frequency . it would be interesting to search for a systematic effect such as this in the data , but in any such effort the many other parameters affecting the microwave spectrum must also be accounted for . [ sect : hotfp ] ( 90mm,80mm)dems.eps the chromosphere in solar flares is strongly heated . this is readily seen in e.g. euv images of flare ribbons which indicate plasma of at least a million degrees . however , it has been known at least since the days of _ yohkoh _ , though not widely appreciated , that more extreme plasmas exist in the chromospheric footpoints during the flare impulsive phase . impulsive soft x - ray footpoints observed via their bremsstrahlung emission by @xcite and , in a large sample by show temperatures close to 10 mk , and densities of at least a few times @xmath19 ( depending on assumptions about their size ) . using euv spectroscopy of flare footpoints from _ @xcite have determined the emission measure distribution for impulsive phase footpoints in a number of small ( b- and c - class ) events , and these typically also peak at 10 mk . an example of a footpoint emission measure distribution for a c1.1 flare is shown in figure [ fig : fpems ] , compared with the loop emission measure distribution from the same time in this event , and non - flaring active region emission measure distribution . independently , density diagnostics of this event return values of around @xmath20 at a temperature of 1.8 mk @xcite . direct density diagnostics for higher temperatures were not available . the gradient of @xmath21 for the footpoint is 1 in all cases studied ; we note that this is consistent with conductive heating balanced by radiative cooling @xcite . the consequences of these hot , dense compact footpoints for microwave footpoint emission have not been explored . in the usual microwave flare modelling , the characteristics of the _ coronal _ flaring source are carefully studied , e.g. the inhomogeneities of the coronal field or the effects of pitch - angle distribution of the electrons on the emission ( e.g. * ? ? ? but the modelling assumes that footpoints are rather cool as well as dense , and microwave emission will thus be free - free absorbed . the hot footpoint plasmas that we find are essentially at ` coronal ' temperatures but located at chromospheric altitudes and with density and magnetic field strength higher than typically found in the corona . they would be expected to produce intense compact sources , with spectral properties similar to those computed for coronal loops , and dominating any coronal emission in their optically - thin ranges due to higher densities and fields . the high plasma densities might however lead to razin suppression at low frequencies . in figure [ fig : hotfps ] we show calculations of footpoint emission for different temperatures and densities in the ranges deduced from the euv observations . of course , the observations also suggest that the hot footpoint plasma would be very inhomogeneous , with temperature varying by a factor 10 over a distance of probably 1000 km or so . so these spectra are for the moment only indicative . the parameters used in these calculations are : field strength of 100 , 500 or 1000 g ( magenta , green and blue curves , respectively ) and viewing angle of 45@xmath18 , isotropic electron distribution with flux spectral index @xmath22 , having a minimum electron energy of 200 kev and a maximum energy of 5 mev . the source angular diameter is 5@xmath0 and depth along the line of sight is 2@xmath0 , comparable to the depth of the chromosphere . the non - thermal electron density in the footpoint is around one part in @xmath23 of the background thermal density , or @xmath24 . ( 80mm,80mm)mw2_v2.eps for the temperatures @xmath25k and @xmath26k ( shown in figure [ fig : hotfps ] by continuous and dashed curves ) and densities @xmath27 and @xmath20 ( shown by thin and thick lines ) found from euv and soft x - rays , observable footpoint microwave sources in the norp frequency ranges are predicted . the emission is mostly non - thermal gyrosynchrotron , and source intensity is determined mainly by the magnetic field strength , where stronger fields shift the spectrum peak towards higher frequencies with stronger emission ( e.g. * ? ? ? * ) . the contribution of the non - thermals to the microwave spectrum as shown in figure [ fig : hotfps ] is greatly in excess of the thermal gyrosynchrotron which would be expected from these same plasma parameters . for a given field strength the intensity in the norh range , and norp above 3.8 ghz , is rather insensitive to the different values of density and temperature used here . the exception is in the weak - field case where the emission is affected by free - free absorption and razin suppression . the razin suppression ( e.g. * ? ? ? * ) is significant for microwave frequencies below the razin frequency @xmath28 , which is below 5 ghz for all cases considered , except for @xmath29 g and @xmath30 , where @xmath31 ghz . the resulting spectra are thus strongly suppressed , and show slighly different contributions from that produced by the free - free mechanism , for @xmath32k or @xmath26k . the steep emission decrease towards lower frequencies is caused by absorption below the plasma frequency ( around 1 ghz and 3 ghz for densities of @xmath27 and @xmath20 , respectively ) . imaging spectropic analysis of future observations with e - ovsa , coupled with euv , uv and optical observations , will provide an excellent diagnostic tool for deriving the magnetic field , plasma density and temperature in chromospheric flaring regions . we should also note that the recently discovered sub - thz spectral component above 100 ghz @xcite is likely to be generated in the chromosphere , when considering the proposed radiation mechanisms @xcite . in terms of gyrosynchrotron emission , it can be shown that the second spectral peak can be produced in hot and dense footpoints with strong magnetic fields and strong razin suppression , self - consistently with the typical microwave spectrum generated in the coronal source @xcite . observational understanding of the energetically dominant processes in the flare lower atmosphere during the impulsive phase , drawing on the many space- and ground - based instruments currently observing the sun , is developing rapidly and in some unexpected directions . the new results presented here on the plasma and field parameters in the chromosphere during the flare impulsive phase are important for future microwave modelling , and the multi - wavelength data that we now have at our disposal must also be confronted with ongoing microwave imaging and spectra observations , which provides unique diagnostics of both plasma and field . to this end , the continued operation of the nobeyama radioheliograph and radio polarimeters remain crucial for our exploration and understanding of flares . l.f . is very grateful to the conference organisers for the financial support which allowed her participation in this most simulating meeting . this work was supported by stfc grant st / i001808/1 and by ec - funded fp7 project hespe ( fp7 - 2010-space-1 - 263086 ) .	this short paper reviews several recent key observations of the processes occurring in the lower atmosphere ( chromosphere and photosphere ) during flares . these are : evidence for compact and fragmentary structure in the flare chromosphere , the conditions in optical flare footpoints , step - like variations in the magnetic field during the flare impulsive phase , and hot , dense ` chromospheric ' footpoints . the implications of these observations for microwaves are also discussed .
the calar alto observatory is located at 2168 m height above the sea level , in the sierra de los filabres ( almeria - spain ) at 45 km from the mediterranean sea . it is the second largest european astronomical site in the northern hemisphere just behind the observatorio del roque de los muchachos ( located in the island of la palma ) , and the most important in the continental europe ( with excellent communications , making logistics easy , unexpensive and reliable ) . currently there are six telescopes located in the complex , three of them directly operated by the centro astronmico hispano alemn a.i.e . , a partnership between the spanish national research council ( csic ) and the german max - plank society ( mpg ) . these telescopes include the zeiss 3.5 m , the largest telescope in the continental western europe . the observatory is under operations since 1975 , when its 1.23 m zeiss reflector saw first light . the observatory operates a very large array of optical and near - infrared astronomical instrumentation , including imagers and spectrographs with different field - of - view and resolutions . there has been different attempts to characterize some of the main astronomical properties during its 35 years of operations : ( i ) leinert et al . ( 1995 ) determined the sky brightness corresponding to the year 1990 ; ( ii ) hopp & fernandez ( 2002 ) studied the extinction curve corresponding to the years 1986 - 2000 ; ( iii ) ziad et al . ( 2005 ) estimated the median seeing in the observatory from a single campaign in may 2002 ; and more recently ( iv ) snchez et al . ( 2007 ) where the optical night sky spectrum was presented , including an analysis of the light pollution , together with a more accurate estimation of the night - sky extinction , the typical seeing , the night - sky brightness and the fraction of useful time ; and ( v ) snchez et al . ( 2008 ) , where the night sky brightness in the near - infrared and the fraction of useful time was presented . several of these features are discussed below . the comprehensive database for the weather is public and it can also be obtained upon request . * left.- * extinction due to dust at calar alto . the values are comparatively smaller than representative values of observatories closer to the equator , both in the northern and the southern hemispheres . the behaviour is highly seasonal . * right.- * seeing distribution ( june 2001 sept . 2005 ) at the calar alto observatory . the median value is about 0.9 arcsec ( snchez et al . , title="fig:",width=257 ] * left.- * extinction due to dust at calar alto . the values are comparatively smaller than representative values of observatories closer to the equator , both in the northern and the southern hemispheres . the behaviour is highly seasonal . * right.- * seeing distribution ( june 2001 sept . 2005 ) at the calar alto observatory . the median value is about 0.9 arcsec ( snchez et al . , title="fig:",width=283 ] the optical spectrum at the calar alto observatory shows a strong contamination produced by different chemicals , in particular from mercury lines , which contribution to the sky - brightness in the different bands is of the order of 0.09 mag , 0.16 mag and 0.10 mag in b , v and r respectively . regarding the strength of the sodium pollution line in comparison with the air - glow emission , it is very strong , a problem which we expect to address in the near future in collaboration with the regioanl governament and the nearby towns . in any case , the observatory complies with the iau recommendations of a dark site . as a matter of fact , caha is classified as a class c site `` major observatory sites with operating telescopes having apertures @xmath02.5 m and zenith light pollution levels less than the natural variation in night - sky brightness associated with the 11-year solar cycle '' ] . a light - pollution regulation was been recently approved by the andalusian regional government , and the observatory is involved in its development , in order to reduce and revert the effect of human light pollution . the effect of such laws have been strong in other astronomical sites , as it was demonstrated for the kitt - peak observatory . we expect a similar influence in the next years . therefore , the darkness of the observatory should improve considerably in the near future , making it even more attractive for new instrumentation . * left.- * wind - gust at calar alto observatory during the last 13 years . note the logarithmic scale in the frequency . the absolute velocity maximum is below 47 m / s . * right.- * the pie chart shows the distribution of time spent at different velocities . , title="fig:",width=298 ] * left.- * wind - gust at calar alto observatory during the last 13 years . note the logarithmic scale in the frequency . the absolute velocity maximum is below 47 m / s . * right.- * the pie chart shows the distribution of time spent at different velocities . , title="fig:",width=257 ] the zenith - corrected values of the moonless night - sky surface brightness , for a typical dark night , are 22.39 , 22.86 , 22.01 , 21.36 , 19.25 , 15.95 , 13.99 , and 12.39 mag arcsec2 in u , b , v , r , i , j , h and k bands , which indicates that calar alto is a particularly dark site for optical and near - ir bands . these values are similar to those at other astronomical sites , including paranal , la silla , la palma and mauna kea . only the last one is clearly darker in the near - ir bands than any of the others . the typical extinction in the observatory in the v - band is 0.15 mag in the winter season , with little dispersion . in summer the extinction has a wider range of values , due to an increase of the aerosol extinction ( dust grains ) , although it does not reach the extreme peaks observed at other sites ( figure 1a ) . in particular , the extinction in the summer season is much lower than those reported for the observatories in the canary islands ( benn & ellison 1998 ) , where the influence of the sahara desert in that archipelago makes the extinction to rise above 0.25 mag in the v - band for a 20% of the time in this season . the presence of ozone is negligible at the calar alto observatory . the derived extinction curve shows that the typical extinction in the u - band is 0.4 - 0.5 mag , a remarkable good value for observatories at this height . the distribution of seeing was derived using the information gathered by the seeing monitor installed in the observatory for more than 4 years ( 300.000 individual measurements ) , and the science data from the alhambra ( moles et al . 2008 ) survey , from @xmath15000 individual frames . in both cases it was obtained that the median seeing was about 0.9 " , with a sub - arcsecond seeing in @xmath170% of the time ( figure 1b ) . the historical record in the observatory shows the maximum registered wind speed was 47 m / s in the last 28 years . however , values above 25 m / s are very rare , and these events are smooth and last for short periods of time . figure 2 displays the frequency of the gust - wind , and the duration for each type of event . as can be seen , there is no sudden variations . this fact makes the observatory a excellent site for large infrastructures . the fraction of astronomical useful time at the observatory was derived independently using the information obtained using the time when the extinction monitor is under operation , and the time when the telescopes were open for a long term project ( alhambra , 358 nights distributed along 6 years , moles et al . a total of 70% of total time was useful to perform astronomical observations under the following criteria : no obvious clouds or cirrus , extinction under 0.2 mag in the v - band , relative humidity under 95% . the fraction of complete clear nights ( more than six continuous hours of useful time ) , drops to a 50% . finally , a 30% of the nights were photometric , defined as nights for which the extinction was stable within the a 20% of the average value ( ie . , it was possible to perform direct photometric calibration with an accuracy of 0.2 mags ) . * left.- * clear nights at calar alto observatory during the period 1976 - 2009 . clear nights are defined as those having at least 6 observing hours of clear or mostly clear sky ( clouds cover not more than 25% ) . * right.- * total number of hours per year when astronomical observations were carried out , for the period 1976 - 2009 . , title="fig:",width=298 ] * left.- * clear nights at calar alto observatory during the period 1976 - 2009 . clear nights are defined as those having at least 6 observing hours of clear or mostly clear sky ( clouds cover not more than 25% ) . * right.- * total number of hours per year when astronomical observations were carried out , for the period 1976 - 2009 . , title="fig:",width=268 ] figure 3ab contains two histograms with the historical record for the observatory ( for the period 1976 - 2009 ) . the first one represents the number of clear nights ( defined as those having at least 6 observing hours of clear or mostly clear sky , with clouds cover below 25% ) . the data collected during these 30 years ( about three solar activity cycles ) clearly indicate that the number of clear nights are around 170 nights per year . see second histogram displays the number of useful hours per year , a more representative quantity since it is closely connected with the observations and the amount of data which can be produced . the histogram is highly variable , but during the last years an average of 2000 hours has been achieved . this fact might be related with the extensively use of service mode , much more effective than programs executed under a conventional visitor mode . assuming 9 useful hours per night , 2000 hours are equivalent to 222 full nights per year . * left.- * average for the number of clear nights per month during the period 2000 - 2009 . the bars provide an idea of the range of variability for each month . * right.- * monthly average for the number of useful hours for the period 2000 - 2009.,title="fig:",width=260 ] * left.- * average for the number of clear nights per month during the period 2000 - 2009 . the bars provide an idea of the range of variability for each month . * right.- * monthly average for the number of useful hours for the period 2000 - 2009.,title="fig:",width=264 ] as in the case of the extinction ( subsect . [ subsection : extinction ] ) , the effectiveness of the observatory , in terms of nights or hours per month , is highly seasonal . figure 4 illustrates this fact . in the first panel , we represent the number of clear nights per month . there is a conspicuous peak during the summer and a minimum in december , but january shows a very large dispersion , with years containing 20 clear nights and other with less than 10 clear nights . when expressed in terms of useful hours ( figure 4b ) the situation is more balanced . the summer peak is still present , but modulated by the night length . therefore , the minimum is located during early spring . surprisingly , january can be either very poor in useful hours or outstanding . .current and future instrumentation [ cols="^,^,^,^,^ " , ] + @xmath2 commissioning 2011 . + @xmath3 commissioning 2012 . + @xmath4 commissioning 2014 . + additional information at : + www.caha.es/telescopes-overview-and-instruments-manuals.html [ tab1 ] the regional government of andalusia ( junta de andaluca ) approved in 2007 an environmental law that includes a chapter on light pollution . this law becomes fully effective after the publication of its ancillary regulation ( august 2010 ) . calar alto has been advising junta de andaluca from the beginning of this legal project , with one representative in the advisory committee established to define the general terms of the law and its regulation . the law foresees creating an office for the protection of night sky against light pollution , and calar alto has secured its participation in this official institution . a protocol is under preparation and its implementation will allow caha to perform specific and individual negotiations with all the relevant cities and towns in the neighbourhood of the observatory . therefore , the quality as an outstanding astronomical site will be secured for many years to come . the observatory operates three telescopes at the present time , with primary mirrors of 3.5 m 2.2 and 1.23 meters . the instrumental suite is very diverse in the first two cases , and include optical and near - infrared imagers and spectrographs ( six instruments for the 3.5 m and four for the 2.2 m ) . in order to optimize the observatory resources ( man - power , financial and observing time ) , we are introducing a number of changes in our operational mode , including the implementation of a public archive and a simplified suite of instruments , following the recommendations of the european telescopes strategic review committee ( etsrc ) , which was appointed by the astronet board in coordination with the opticon executive committee . the current and possible future suite of instruments are summarized in table 1 . an example of a query in the caha public archive , developed under vo standards.,title="fig:",width=574 ] in collaboration with the center of astrobiology ( cab ) , which hosts the spanish node of the virtual observatory ( solano 2006 ) , we have developed a web - based tool in order to provide access to the data acquired at calar alto observatory . all data taken after july 1st 2010 will be made public after one year of proprietary rights . for data collected prior that date ( starting january 2008 ) , we have requested permit from the principal investigators for each individual program . the final goal is to provide raw and reduced data , as well as meta - data and links to other vo - compliant archives and tools ( as an example , see vosa , bayo et al . 2009 ) , in order to maximize the scientific exploitation of the data . figure 5 includes an example of a caha public archive query . we expect that the utility will be opened in early 2011 . the observatory has recently approved its first legacy program , the the califa survey ( calar alto legacy integral field area survey , snchez et al . 2010 , 2011 ; marino et al . this program will provide the largest and most comprehensive wide - field ifu survey of galaxies carried out to date , addressing several fundamental issues in galactic structure and evolution . we will observe a statistically well - defined sample of @xmath1600 galaxies in the local universe using 210 observing nights with the pmas / ppak integral field spectrophotometer , mounted on the calar alto 3.5 m telescope . the defining science drivers for the project are : a ) star formation and chemical history of galaxies , b ) the physical state of the interstellar medium , c ) stellar and gas kinematics in galaxies , and d ) the influence of the agns on galaxy evolution . califa will provide a valuable bridge between large single - aperture surveys such as sdss and more detailed studies of individual galaxies with ppak and other instruments . additional legacy program might be implemented in the future , specially at the 2.2 m telescope . * left.- * scheme of the optical design of cafe , including its main components . * right.- * first assembling of cafe of its optical bench , in september 2010 . the laser is used to align the different components . , title="fig:",width=283 ] * left.- * scheme of the optical design of cafe , including its main components . * right.- * first assembling of cafe of its optical bench , in september 2010 . the laser is used to align the different components . , title="fig:",width=238 ] at the present time , we are developing three new instruments , namely a high - spectral resolution optical spectrograph for general use at the 2.2 m ( cafe ) , a wide field - of view ( 30x30 arcmin ) near - infrared camera for the 2.2 m and the 3.5 m ( panic ) , and a high - spectral near - infrared spectrograph for exoplanetary searches ( carmenes ) . more information can be found in these proceedings or below for cafe . the calar alto fiber - fed echelle spectrograph ( cafe ) is an instrument under construction at caha to replace foces , the high - resolution echelle spectrograph at the 2.2 m telescope of the observatory . foces is a property of the observatory of the munich university , and it was recalled from calar alto by this institution in 2009 . the use of this instrument was very extensive , and represented a substantial fraction of the 2.2 m telescope time during its operational life - time . due to this fact , the observatory decided in 2008 to to build an improved replacement . cafe share its basic characteristics with those of foces . however , significant improvements have been introduced in the original design ( adding new calibration units ) , the quality of the materials , and the overall stability of the system . it is expected that the overall efficiency and the quality of the data will be significantly improved with respect to its predecessor . in particular , cafe is design and built to achieve resolutions of r@xmath170000 , which will be kept in the final acquired data , allowing it to compete with current operational extrasolar planets hunters . lcc parameter & foces & cafe + design & echelle spectrograph & echelle spectrograph + telescope & calar alto 2.2 m & calar alto 2.2 m + operational live - time & 1997 - 2009 & starting in 2011 + resolution & 46000/64000 & 70000 + ccd pixel scale & 24/15@xmath5 m & 13.5@xmath5 m + ccd cooling system & liquid nitrogen & peltier + wavelength range & 3800 - 7450@xmath6@xmath2 & 3800 - 7450@xmath6 + stability system & un - stabilized & vibrations and temperature + moving parts & slit , grating and prisms & no moving parts + calibration system & th - ar lamps & th - ar lamps and iodine - cell + quality of optical components & @xmath7/10 & @xmath7/20 + expected s / n@xmath3 & @xmath1100 & @xmath0150 + + + [ tab1 ] it is expected that the final assembling of the instrument take place before the end of 2010 . telescope commissioning is expected for spring 2011 . we expect that panic will be operational in 2012 and carmenes , which has received green light very recently , in 2014 . the calar alto observatory has two main properties : a very well - characterized and excellent astronomical site ( specially for spectroscopy ) and outstanding logistics ( excellent communications and a location in continental europe ) . the current and future suite of instruments for the optical and near - ir telescopes , and the support of two of the main research institutions in europe ( the spanish national research council and the german max - plank society , channeled through the instituto de astrofsica de andaluca and the max - plank institut fr astronomie ) , provide the necessary conditions for its continuation as relevant astronomical facility . we thank the _ `` viabilidad , diseo , acceso y mejora '' _ funding programs , icts-2008 - 24 and icts-2009 - 10 , of the spanish ministry of science , and the _ proyecto de excelencia _ funding program fqm-08 - 00360 of the _ junta de andaluca _ , for the support given to this project . the observatory will not be possible without the superb professionalism of its staff .	the calar alto observatory , located at 2168 m height above the sea level in continental europe , holds a significant number of astronomical telescopes and experiments , covering a large range of the electromagnetic domain , from gamma - ray to near - infrared . it is a very well characterized site , with excellent logistics . its main telescopes includes a large suite of instruments . at the present time , new instruments , namely cafe , panic and carmenes , are under development . we are also planning a new operational scheme in order to optimize the observatory resources .
i shall discuss some work that i have done with simon dalley @xcite on induced operators in qcd . i shall begin with a very brief review of the schwinger model , the one model that includes induced operators that can be analyzed in complete detail @xcite . the reasons for reviewing the schwinger model are to remind ourselves that induced operators do exist and do have physical consequences , and to set some notation and ideas for the discussion of qcd , where the analysis can not be done in complete detail . i shall then derive two of the induced operators in qcd and illustrate aspects of their effects with simple models . i shall end with a brief discussion . when we solve the schwinger model in the light - cone representation @xcitewe must solve the light - cone constraint equation _ - _ - = - i _ + , as _ - = _ -^0(x^+ ) - i_+ dx^- , where _ -^0 = z_-()e^_-^(-)()_- e^_-^(+ ) ( ) . here , @xmath1 is a spurion and @xmath2 is the fusion field . the spurion , @xmath1 , is needed to construct the physical vacuum as : _ m=-^^im ; = ( _ + ^_-)^m . the field @xmath2 is entirely unphysical , but if we write the physical field @xmath3 in bosonized form as _ + = z_+ e^_+^(-)_+ e^_+^(+ ) , we find that the operator @xmath4 includes the term @xcite dx^- z_-z_+ _ -^(_-^_+ ( e^_+^(-)e^_+^(+ ) - 1 ) + c.c . ) dx^- . [ io ] here , we have not included the fields @xmath2 since they do not act in the physical subspace . also , @xmath5 acts as a constant in the physical subspace so the physical subspace can be taken to be the usual light - cone subspace created by operators from the @xmath3 field but with the addition of the operator ( [ io ] ) with @xmath5 replaced by the appropriate c - number . the operator ( [ io ] ) is an induced operator . it is important to understand that all the above results can be obtained by quantizing the model either on the light - cone or at equal - time , so the conclusions are independent of quantization surface . details , including the operator solution can be found in @xcite . before leaving the schwinger model there is one further point we should discuss since it will be important to what follows . if we consider the case where periodicity conditions are imposed on @xmath6 ( dlcq ) then most of the results obtained in the continuum still follow . but the coefficient of the induced operator goes to zero as the periodicity length , @xmath7 , goes to infinity . that is because the induced operator is proportional to the chiral condensate which in the continuum is given by _ 1 = - , 2 ^3/2 e^ , but in the periodic case is given by @xcite \| = - , l . the fix for the problem is to multiply the induced operator in the periodic case by a renormalization constant which is equivalent to taking z_- = , 2 ^ 3/2 . with this correction , the all the results in the periodic case go to those in the continuum case when @xmath8 . we do not know any way to determine the correct coefficient in the periodic case other than to compare with the continuum . if we had access to two dimensional experimentalists that could be comparing with experiment . we now turn our attention to qcd using the schwinger model as a guide @xcite . we quantize the fermi fields on @xmath6 in the standard way : ^(a)_+,s(0,x^-,x^ ) = 1_k_e^ik_x^ ^(a)_+,s(0,x^-,k^ ) , where ^(a)_+,s(0,x^-,k^ ) = _ n = 1^ b^(a)_s(n ,- k _ ) e^-ik_-(n)x^- + d^(a)_-s(n , k _ , ) e^ik_-(n)x^- . [ ff ] the object in ( [ ff ] ) is a standard one dimensional fermi field and so , can be bosonized in the standard way . we write ^(a)_+,s(0,x^-,k^ ) = e^-_s^(a)(-)(x^-,k _ ) _ + , s^(a)(x^-,k _ ) e^-_s^(a)(+)(x^-,k _ ) , where _ s^(a)(+)(x^-,k _ ) = -_n=1^c^(a)_s(n , k_)e^-i_-(n)x^- , _ s^(a)(-)(x^-,k _ ) = -_s^(a)(+)^ = _ n=1^c^(a)_s(n , k_)e^i_-(n)x^- , and c^(a)_s(n , k _ ) & = & ^n-1_= 0 d^(a)_-s(+ , 2,k _ ) b^(a)_s(n - - , 2,-k _ ) + + & & ^_= 0 b^(a)_s ( + , 2,-k _ ) b^(a)_s ( + n + , 2,-k _ ) - + & & ^_= 0d^(a)_-s ( + , 2,k _ ) d^(a)_-s ( + n + , 2,k _ ) in qcd , the field @xmath9 satisfies a constraint equation and the solution to that equation involves an integration constant , @xmath10 , ( a function of @xmath11 and @xmath12 ) . we specify that integration constant as @xmath13 as with the case of the schwinger model , the only physical operator from this field is the spurion , @xmath14 ( the other operators in @xmath10 are needed to maintain the canonical structure of the theory ) . we shall need only the transverse components of the gluon fields . we work in the helicity basis defined by a _ = ( a_1 -ia_2)/ , a _ = ( a_1 + ia_2)/ , then quantize as a^(c)_s(0,x^-,x^ ) = 1_k_e^ik_x^ ^(c)_s(0,x^-,k^ ) , where & & ^(c)_s(0,x^-,k^ ) = + & & _ n = 1^ , ( a^(c)_s(n ,- k _ ) e^-i_-(n)x^- + a^(c)_s^(n , k _ , ) e^i_-(n)x^- ) . so far everything has just been algebra . now we need an ansatz for the vacuum and we shall use the schwinger model for a guide . the schwinger model might suggest that we require _ -,-s^(a)(x^-,k_)_+,s^(a)(x^-,k _ ) = . [ vev ] the vector which satisfies that requirement is = _ s ; a ; k_(_n = -^ ( _ -,-s^(a)(0,k _ ) _ + , s^(a)(0,k_))^n ) . that vector is not a suitable choice for the vacuum since it is not gauge invariant . but we find that if we choose the gauge invariant state = _ s ; k_(_n = -^(_a _ -,-s^(a)(0,k _ ) _ + , s^(a)(0,k_))^n ) [ vac ] , to be the vacuum we can work out the action of the induced operators in the physical subspace assuming that ( [ vev ] ) holds even though ( [ vac ] ) does not satisfy that relation in the entire space . with all this machinery in place we can specify the induced operators . the first one , which we call @xmath15 , is a part of the term @xmath4 if we keep only the part of @xmath4 that depends on @xmath10 ( actually , on @xmath14 ) . we have _ 1 = dx^-d^2x^(^(a)0_-,^ ^(a)_+ , + ^(a)_+,^ * ^(a)0_- , + ^(a)0_-,^ * ^(a)_+ , + ^(a)_+,^ * ^(a)0_- , ) which , keeping only the spurion from @xmath10 , can be rewritten as _ 1 & = & g_1 dx^-d^2k^ ( _ -,^(a)(0,k_)_+,^(a)(0,k _ ) ) + & & ( e^-_^(a)(-)(x^-,k _ ) e^-_^(a)(+)(x^-,k _ ) -1 ) + c.c . + spin flip ) , where @xmath16 is a constant included to account for the fact that , as in the case of the schwinger model , the operator may scale in a nontrivial way with the periodicity length @xmath7 . since we do not have a continuum solution with which to compare , @xmath16 will have to be fit to a symmetry or to data . now , we can commute the combination , @xmath17 , or its adjoint , to the vacuum where , by ( [ vev ] ) , it will become a c - number , either @xmath18 or @xmath19 ( if we wish to have a c - invariant theory @xmath18 must be real ) . once we have eliminated @xmath17 in favor of @xmath18 , the remaining operators are operators that act in the usual dlcq subspace and we can perform the rest of the calculation in that space taking the vacuum to be the bare light - cone vacuum . as an example of the action of @xmath15 i shall give its action on a valence type state for a meson . we have & & i_1 ^_(n ,- k _ ) d^(a)^_(m , k _ ) = g_1 ( b^(a)^_(n-1,-k _ ) d^(a)^_(m + 1,k _ ) + & & - 2 b^(a)^_(n ,- k _ ) d^(a)^_(m , k _ ) + b^(a)^_(n + 1,-k _ ) d^(a)^_(m - 1,k _ ) ) . we find that if we restrict ourselves to meson valence states of the form _ 0 ^ 1 dx f(x ) b^(a)^_(x ) d^(a)^_(1-x ) , then : if @xmath16 is independent of @xmath7 , @xmath15 goes to zero as @xmath7 goes to infinity ; if @xmath16 scales linearly with @xmath7 , @xmath15 becomes a pure endpoint operator as @xmath7 goes to infinity ; if @xmath16 is proportional to @xmath20 , @xmath15 becomes the operator @xmath21 as @xmath7 goes to infinity . if @xmath15 does not go to zero , it will give the relation that the mass of the pion squared is proportional to the bare quark mass for small bare quark mass . now let us develop the other induced operator , @xmath22 , the induced operator that is part of the term @xmath23 . we write & & i_2 = i_2,1 + i_2,2 + i_2,3 + i_2,4 = g dx^-d^2x^_abc_a b^c + & & ( _ + , ^(a ) _ -,^0 ( b ) a_^(c ) -_-,^0(a ) * _ + , ^(b ) a_^(c ) + _ -,^0(a ) * _ + , ^(b ) a_^(c ) - _ + , ^(a ) * _ -,^0(b ) a_^(c ) ) where @xmath24 is the usual qcd coupling constant . consider the first of these operators i_2,1 & = & g g_2 d^2p_d^2k_dx^- _ abc _ a b^c + & & _ + , ^(a)(x^-,k _ ) _ -,^(b)(x^-,k_- p _ ) _ ^(c)(x^-,p _ ) where , as was the case with @xmath16 , @xmath25 is a constant incorporating the unknown dependence of the coefficient of @xmath22 on the periodicity length @xmath7 . we now insert the number 1 written as @xmath26 to get i_2,1 = -g g_2 d^2p_d^2k_dx^- _ abc&&_a b^c_^(c)(x^-,p _ ) _ + , ^(a)(x^-,k _ ) _ + , ^(b)(x^-,k_- p _ ) & & we can commute the operator in brackets through any other operators until it accts on the vacuum at which time it will just give a c - number . after performing that operation all the remaining operators will act in the usual dlcq subspace and we an perform the remainder of the calculations in that space taking the bare light - cone vacuum as the vacuum . working similarly , we find that the rest of @xmath22 is given by i_2,2 = g g_2 d^2p_d^2k_dx^- _ abc & & _ a b^c_^(c)(x^-,p _ ) _ + , ^(a)(x^-,k_+ p _ ) _ + , ^(b)(x^-,k _ ) + & & i_2,3 = - g g_2 d^2p_d^2k_dx^- _ abc & & _ a b^c_^(c)(x^-,p _ ) _ + , ^(a)(x^-,k_+ p _ ) _ + , ^(b)(x^-,k _ ) + & & i_2,4 = g g_2 d^2p_d^2k_dx^- _ abc & & _ a b^c _ ^(c)(x^-,p _ ) _ + , ^(a)(x^-,k _ ) _ + , ^(b)(x^-,k_- p _ ) + & & to illustrate the action of @xmath22 we work out its action on a valence type meson state . we find that if we define @xmath27 then & & i_2 = ( i_2,1 + i_2,3 ) , where & & i_2,1 = - g g_2 ^ d^2p _ _ abc^c_ab + & & b^(a)_(n ,- q_)d^(b)_(1,q_- p _ ) a^(c)_(k - n-1,p _ ) , and & & i_2,3 = - g g_2 d^2p _ _ abc^c_ab + & & b^(b)_(1,-q_- p_)d^(a)_(k - n , q _ ) a^(c)_(n-1,p _ ) the pattern seen here is general : although the form of @xmath22 may appear to be complicated , its action is easy to state . it acts successively on each quark just as a standard qcd three point vertex except that the quark spin is always flipped and the gluon absorbs all the available longitudinal momentum leaving the quark in the lowest longitudinal momentum state . color works as in the standard vertices and the transverse momentum is shared between the quark and the gluon in all possible ways just as in the standard vertices . the operator is nonzero at zero quark bare mass and will split the pi and rho masses at zero bare quark mass . let us further illustrate the action of @xmath15 and @xmath22 by solving a simple model . we shall keep only @xmath15 and @xmath22 and the quark kinetic energy as our @xmath28 . we shall take the quark bare mass to be @xmath29 , @xmath18 to be -1 and @xmath16 to be @xmath30 where @xmath31 is the dlcq harmonic resolution . we can find an eigenstate for the pion that has the form = & & c ( a_^(1)b^_(0 ) d^_(0 ) + a_^(1 ) b^_(0 ) d^_(0 ) ) + + & & _ 0 ^ 1 dx ( x ) 1(b^_(x ) d^_(1-x ) - b^_(1-x ) d^_(x ) ) . with the wave function in that form , @xmath15 has the form @xmath32 in the large @xmath31 limit . thus , in that limit , @xmath33 satisfies the equation - g_11 , dx^2 + , x(1-x ) ( x ) = m^2 ( x ) , where @xmath34 is the mass of the pion . to get a finite , nontrivial result we must choose @xmath25 to have the form g_2 = , g k^,2 - , 3 g g_11 - g_22 , where @xmath35 is a free parameter . with that choice we find that m^2_= , g_11 + 4 g g_22 , and ( x ) & = & + ( , g_11)^,2 x ( 1-x ) + + & & , g_11 ^ 2 x ( 1-x ) + ( , g_11)^,2 x^2 ( 1-x)^2 . we can find an eigenstate of the rho of the form = _ 0 ^ 1 dx ( x ) 1(b^_(x ) d^_(1-x ) + b^_(1-x ) d^_(x ) ) . for values of @xmath29 that are not too large we find that , to a good approximation ( x ) ; m^2_= ^2 g_11 + 4.88 ^2 . as an example , if we choose @xmath36 , @xmath37 and @xmath38 we get @xmath39 , @xmath40 and @xmath41 . while , given the mass dimensions of the various constants , these numbers are all consistent with @xmath42 being a few hundred mev , the model is not very realistic in that @xmath22 not only splits the pi and rho masses but gives the rho all of its mass ( except for a small contribution from the kinetic energy ) . presumably the rho gets most of its mass from other operators in the @xmath28 of qcd that are left out of this simple model . further samples of model calculations with different scalings of the coefficients of the induced operators may be found in ref . @xcite , including a dlcq calculation of dimensionally reduced gauge theory that includes the standard part of @xmath28 as well as the induced operator @xmath22 . in all the cases studied the pi and rho masses are split in the continuum limit . we have given two of the induced operators in qcd . these arise due to the fact that there are ( unphysical ) quarks in the vacuum state . there will probably be other induced operators due to the presence of ( unphysical ) gluons in the vacuum ; we do not yet have a form for these operators or for the glue in the vacuum . it is possible that the induced operators we have given here are the most important ones for determining the structure of low mass hadrons but we do not know that . to derive the induced operators we made an ansatz for the vacuum based on the schwinger model . in principal it is possible to check whether or not that ansatz is correct . but to do so we would need a nonperturbative regulator for the theory that preserves gauge invariance and lorentz invariance . we hope some day to have such a regulator but we do not yet have one that we can use to check the ansatz for the vacuum . as they stand , the induced operators have coefficients that may scale with the dlcq harmonic resolution and whose values are unknown . in principal these values , including the scaling , are determined by requiring that all the fields are canonical at all space - like separations in the continuum case . we are unable to carry out the analysis required to use that principle to determine the coefficients and must fit them either to a symmetry or to data . we have used the induced operators in calculations of simple models in two dimensions . we hope soon to extend these studies to four dimensions and to include the effects of additional flavors . on the basis of the results from the simple models it is clear that the operators can split the masses of the pion and the rho at small values of the quark mass , can give a linear growth of the mass squared of the pion with bare quark mass and can modify the endpoint behavior of the pion distribution function .	light - cone quantization always involves the solution of differential constraint equations . the solutions to these equations include integration constants ( fields independent of @xmath0 ) . these fields are unphysical but when they are consistently removed from the dynamics , additional operators ( induced operators ) , which would not be present if the integration constants were simply set to zero , are included in the dynamics . these induced operators can be taken to act in the usual light - cone subspace , for instance , the space used for dlcq . here , i shall give a derivation of two such operators . the operators are derived starting from the qcd lagrangian but the derivation involves some guesses . the operators will provide for the linear growth of the pion mass squared with the quark bare mass and for the splitting of the pi and the rho at zero quark mass .
the hadronization of quark gluon plasma ( qgp ) possibly produced in the early universe or expected to be formed in relativistic heavy - ion collisions has been the focus of much attention during the past few years . however , the mechanism of hadronization ( qcd phase transition ) remains an open question . the prediction of lattice qcd on the order of the transition is still unclear , if physical masses for quarks are used @xcite . quenched qcd ( no dynamical quarks ) shows a first - order phase transition , albeit a weak one , with small surface tension and latent heat @xcite . assuming the transition to be first - order , homogeneous nucleation theory @xcite has been invoked extensively to study the dynamics of the quark hadron phase transition both in the context of early universe as well as for the plasma produced during relativistic heavy - ion collisions @xcite . in this picture , the transition is initiated by the nucleation of critical - size hadron bubbles from a supercooled metastable qgp phase . these hadron bubbles can grow against surface tension , converting the qgp phase into the hadron phase as the temperature drops below the critical temperature , @xmath0 . this is indeed the case for a sufficiently strong first order transition , where the assumption of a homogeneous background of qgp is justified at the time when the nucleation begins . however , for a weak enough transition , the qgp phase may not remain in a pure homogeneous state even at @xmath1 , due to pre - transitional phenomena . for temperatures much above @xmath0 , matter is in a pure qgp phase with the effective potential exhibiting one minimum at @xmath2=0 . here @xmath2 is an effective scalar order parameter generally used to model the effective potential describing the dynamics of a phase transition . as the plasma expands and cools to some temperature @xmath3 , an inflection point is developed away from the origin which on further cooling separates into a maximum at @xmath2=@xmath4 and a local minimum at @xmath5 , corresponding to the hadron phase . at @xmath1 , the potential is degenerate with a barrier separating the two phases . `` pre - transitional phenomena '' refers to nonperturbative dynamical effects above @xmath0 in the range @xmath6 . such phenomena are known to occur in several areas of condensed matter physics , as in the case of isotropic to nematic phase transition in liquid crystals@xcite , and are also expected in the cosmological electroweak phase transition leading to large phase mixing at @xmath1@xcite . in such cases , the phase transition may proceed either through percolation @xcite or , if the phase mixing is below the percolation threshold , by the nucleation of critical bubbles in the background of isolated hadronic domains , which grow as @xmath7 drops below @xmath0 . in either case , the kinetics is quite different from what is expected on the basis of homogeneous nucleation @xcite . we will argue that , for a wide range of physical parameters , a large amount of thermal phase mixing at @xmath1 is expected to occur during the quark - hadron phase transition in the early universe , as well for the plasma produced in heavy - ion collisions @xcite . for high enough temperatures and low enough cooling , large - amplitude thermal fluctuations will populate the new minimum at @xmath5 in the range @xmath8 . although these fluctuations which are in the form of subcritical hadron bubbles will shrink and finally disappear , there will always be some non - zero number density of hadron bubbles at a given temperature @xmath7 . in this work , we study the equilibrium density distribution of subcritical hadron bubbles for a wide spectrum of very weak to very strong first order qcd phase transition , using the formalism developed in refs . it is found that the density of subcritical hadron bubbles builds up faster as the transition becomes weaker , leading , in some cases , to complete phase mixing at @xmath1 . further , using reasonable values for the surface tension and correlation length as obtained from lattice qcd calculations , we find that ( although large ) the amount of phase mixing remains below the percolation threshold . therefore , the quark - hadron phase transition will begin with the nucleation of critical - size hadron bubbles from a supercooled and inhomogeneous background of quark - gluon plasma . since the background contains subcritical hadron bubbles , the homogeneous theory of nucleation needs to be modified . in ref . @xcite , an approximate method was suggested to incorporate this inhomogeneity by modelling subcritical bubbles as gaussian fluctuations , resulting in a large reduction in the nucleation barrier . here , we will study inhomogeneous nucleation in the framework of homogeneous theory , but with a reduced nucleation barrier that accounts for the inhomogeneity of the medium . finally , we also briefly discuss possible implications of inhomogeneous nucleation to relativistic heavy - ion collisions and cosmology . the paper is organized as follows . in the next section , we begin with the discussion of a quartic double - well potential used to describe the dynamics of a first - order quark - hadron phase transition . the parameters of the potential are obtained in terms of relevant physical quantities such as critical temperature , surface tension and correlation length . in section iii , we estimate the equilibrium fraction of subcritical hadron bubbles from very weak to strong first - order phase transitions . we also estimate the reduction in the nucleation barrier by incorporating the presence of subcritical bubbles in the medium . using this reduced barrier , we study nucleation and supercooling in section iv . finally , we present our conclusions in section v. we consider a general form of the potential ( or equivalently , the homogeneous part of the helmholtz free energy density ) to study the quark - hadron phase transition in terms of a real scalar order parameter @xmath2 given by @xmath9 where @xmath10 and @xmath11 are positive constants . ignatius et . @xcite use this parameterization to describe the phase transition from a qgp ( symmetric phase ) to a hadron phase ( broken symmetry ) . the meaning of @xmath2 is obvious for a symmetry - breaking transition , but the same description can be used if no symmetry is involved . the order parameter could then be related to energy or entropy density . the parameters @xmath12 , @xmath10 and @xmath11 are determined in terms of surface tension @xmath13 , correlation length @xmath14 and critical temperature @xmath15 . the potential has two minima , one at @xmath16 and the other at @xmath17 , which in our case will represent quark and hadron phases respectively . these phases are separated by a maximum defined by @xmath18 . at @xmath1 , @xmath19 having the required degeneracy . the above condition yields , @xmath20 using these relations , the barrier height at @xmath0 can be written as @xmath21 therefore , if the parameter @xmath11 is kept fixed , @xmath10 can be varied to characterize a wide spectrum of very weak to very strong first - order phase transitions . the transition is strong enough for large @xmath22 and very weak or close to second order as @xmath23 . in the following , we relate the parameters @xmath10 and @xmath11 to the surface tension and the correlation length in the quark phase . the surface tension can be defined as the one dimensional action given by , @xmath24.\end{aligned}\ ] ] under the thin - wall limit , @xmath25 , the surface tension can be expressed as @xcite @xmath26 similarly , the correlation length around the quark phase is obtained using @xmath27 @xmath28 . at the critical temperature , using eq . ( [ relat ] ) , we get @xmath29 from eqs . ( [ sig ] ) and ( [ xi ] ) we get @xmath30 in terms of the values of @xmath31 and @xmath32 at @xmath0 . the barrier height @xmath22 can now be written as @xmath33 thus , the barrier height is proportional to the ratio @xmath34 . the transition becomes very weak as @xmath31 decreases and @xmath32 increases . here , we fix @xmath35 fm at @xmath1 and vary @xmath31 to investigate phase transitions with different strengths . the temperature dependence of @xmath12 is deduced by equating the depth of the second minimum with the the pressure difference @xmath36 between the two phases at all temperatures . this yields an equation @xmath37 which is solved to get the parameter @xmath38 , giving the temperature dependence of @xmath32 . the surface tension will also have small temperature dependence which we ignore , as we are not going too far from the critical temperature . thus , we have parameterized the free - energy density in terms of the surface tension , correlation length , critical temperature and equation of state , which can be obtained from lattice qcd calculations . the bag equation of state which is a good depiction of the lattice results is used to calculate the quark / hadron pressure @xmath39 as follows @xmath40 where @xmath41 is the bag constant . the quark phase is assumed to consist of a massless gas of @xmath42 and @xmath43 quarks and gluons , while the hadron phase contains massless pions . thus , the coefficients @xmath44 and @xmath45 are given by @xmath46 and @xmath47 . the critical temperature is taken as @xmath48 mev . 1 shows the plot of @xmath49 as a function of @xmath2 at three different temperatures for a typical value of @xmath31 = 30 mev/@xmath50 and @xmath51 fm . at @xmath1 , the potential is degenerate with a large barrier that separates the two phases . below @xmath0 , the phase @xmath5 has lower free - energy density , and the qgp phase becomes metastable . above @xmath0 , the potential has a metastable minima at @xmath5 ( hadron phase ) as long as @xmath7 remains below @xmath3 . the temperature @xmath3 [ at which @xmath52 and @xmath53 can be obtained analytically by solving eq . ( [ atemp ] ) as , @xmath54^{1/4 } \ , t_c.\end{aligned}\ ] ] it may be mentioned here that the dynamics of the phase transition has also been studied in ref . @xcite using a different form of the potential which has been parameterized as a fourth order polynomial in the energy density @xcite . this form is unsuitable over a wide range of temperatures due to the persistence of metastability at much above and below @xmath0 . we closely follow the work of refs . @xcite to estimate the equilibrium density distribution of subcritical hadron bubbles by modeling them as gaussian fluctuations with amplitude @xmath55 and radius @xmath56 @xmath57 the amplitude @xmath55 is the value of the field at the bubble s core away from the quark phase . for smooth interpolation between the two phases in the system , @xmath58 . the free energy of a given configuration can then be found by using the general formula @xcite , @xmath59.\end{aligned}\ ] ] using eq . ( [ gau ] ) and eq . ( [ vt ] ) in eq . ( [ fs ] ) we get @xmath60 where @xmath61 , @xmath62 , @xmath63 and @xmath64 are given by @xmath65 \pi^{3/2 } \phi_a^2\end{aligned}\ ] ] and @xmath66 it may be mentioned here that @xmath67 is positive and is much greater than @xmath68 . therefore , the free energy grows linearly for small values of @xmath56 . further , hadron bubbles of all configurations will be subcritical as long as @xmath68 is positive . at @xmath1 , both @xmath62 and @xmath64 are positive for all amplitudes . however , below @xmath0 , @xmath62 may become negative for some values of @xmath55 . for such configurations , the free energy has a maximum at @xmath69 and these bubbles are not strictly subcritical . the same is true for @xmath64 above @xmath0 . we thus restrict the amplitudes @xmath55 to the range where @xmath68 is positive . if not exactly the same , the limits of integration @xmath70 and @xmath71 for @xmath55 are found to be quite close to @xmath4 and @xmath72 respectively . there will be fluctuations from quark to hadron phase and back . to obtain the number density @xmath73 of subcritical bubbles , we define the distribution function @xmath74 where @xmath75 is the number density of bubbles with radius between @xmath56 and @xmath76 and amplitude between @xmath55 and @xmath77 at time @xmath78 . it satisfies the boltzmann equation @xcite @xmath79 the first term on the rhs is the shrinking term . here , @xmath80 is the shrinking velocity , which we assume to be given by the velocity of sound ( @xmath81 ) in a massless gas . the second term is the nucleation term where @xmath82 is the gibbs distribution function defined as @xmath83 . here @xmath84 is the nucleation rate per unit volume of subcritical bubbles from the quark phase to the hadron phase . similarly @xmath85 is the corresponding rate from the hadron phase to the quark phase . the factor @xmath86 is defined as the fraction of volume in the hadron phase and is obtained by summing over subcritical bubbles of all amplitudes and radii within this phase . the gibbs distribution function is defined as @xcite @xmath87 where @xmath88 is of @xmath89 @xcite . if the equilibration time scale is smaller than the expansion time scale of the system , we can obtain the equilibrium number density of subcritical bubbles by solving eq . ( [ boltz ] ) with @xmath90 . since the early universe expands at a much slower rate @xcite , the above assumption is quite reasonable in the context of the cosmological qcd phase transition . however , qgp produced during heavy ion collision may expand at a faster rate as compared to the early universe . in this case , it is possible that the density distribution of the subcritical bubbles will not attain full equilibrium . for simplicity , we will assume an equilibrium situation so that the present results on the fraction of subcritical bubbles and phase mixing can be considered as an upper limit . using the boundary condition @xmath91 , we get the equilibrium distribution given by @xmath92 where @xmath93 the equilibrium fraction @xmath86 of volume occupied by subcritical bubbles is given by , @xmath94 which is solved to get @xmath95 where @xmath96 here , @xmath70 and @xmath71 define the range within which both @xmath62 and @xmath64 are positive . @xmath97 is the smallest radius of the subcritical bubbles taken as @xmath32 , the correlation length of the fluctuations . the @xmath56 integration should be carried out over all bubbles with radii from @xmath98 to @xmath99 . for very weak transitions , both @xmath100 and @xmath101 are very small and the @xmath56 integration may not have good convergence . however , we found that the @xmath86 value is maximized when @xmath102 is about 3 to 4 fm . therefore , we use @xmath102=3.5 fm . this is a reasonable choice as bubbles with @xmath103 will be statistically dominant and larger fluctuations have larger free energy and are exponentially suppressed . 2 shows the plot of the subcritical hadron fraction @xmath86 as a function of @xmath31 at @xmath1 and at a fixed value of @xmath104 fm . the fraction @xmath86 has been estimated ( dashed curve ) assuming that , for a degenerate potential , @xmath105 , as in ref . this assumption is valid only for the configuration for which @xmath106 . however , when we include other configurations in the range @xmath70 to @xmath71 , the integral @xmath107 turns out to be always higher than @xmath108 at @xmath0 . therefore , @xmath86 obtained using @xmath109 is always lower than when the approximation @xmath110 is used . in both cases , the value of @xmath86 increases with decreasing @xmath31 i.e. when the transition becomes weak . it may be mentioned here that as per lattice qcd calculations without dynamical quarks @xcite , @xmath31 lies between 2 mev/@xmath50 and 10 mev/@xmath50 . there would be @xmath111 to @xmath112 phase mixing corresponding to these @xmath31 values , which is still below the percolation threshold ( @xmath113 ) . if @xmath114 , the two phases will mix completely , the mean - field approximation for the potential breaks down , and the phase transition may proceed through percolation @xcite . however , for a surface tension in the range 2 mev/@xmath115 10 mev/@xmath50 , the phase transition will proceed through the formation of critical - size hadron bubbles from a supercooled metastable qgp phase . since the qgp phase is no longer homogeneous , the dynamics of the phase transition will be quite different from what is expected on the basis of homogeneous nucleation theory @xcite . we refer to it as `` inhomogeneous nucleation . '' we would also like to mention here that the present results are in disagreement with the findings of ref . @xcite , where a large fraction of subcritical hadron phase was found at and above @xmath0 . this scenario is highly unrealistic and probably could be due to the choice of the potential parameterization , which shows a metastable hadron phase much above @xmath0 . therefore , the authors of ref . @xcite found a finite fraction of hadron phase at temperatures as high as twice @xmath0 . furthermore , the value of @xmath86 strongly depends on how the shrinking term is incorporated in the calculation . in our case , it is proportional to the gradient @xmath116 that appears in the kinetic equation ( [ boltz ] ) in a natural way , whereas in ref . @xcite , a specific assumption is made to take into account the shrinking of the hadronic volume . the nucleation rate in the standard theory @xcite which neglects phase mixing , is given by @xmath117 here @xmath118 is free energy needed to form a critical bubble in the homogeneous metastable background . for an arbitrary thin - wall spherical bubble of radius @xmath56 and amplitude @xmath119 , the free energy of the bubble takes the well - known form @xmath120 in the above , @xmath121 is defined as the difference in free - energy density between the background medium and the bubble s interior . for a homogeneous background ( metastable ) , we can write , @xmath122 if there is significant phase mixing in the background metastable state , its free energy is no longer @xmath123 . one must also account for the free energy density of the nonperturbative large amplitude fluctuations . following ref . @xcite , we write the free energy density of the metastable state as @xmath124 , where @xmath125 is the extra free energy density which can be estimated from the density distribution of subcritical bubbles as follows : @xmath126 once we know the hadronic fraction @xmath86 and the free energy @xmath127 for a bubble of a given radius @xmath56 and amplitude @xmath55 , we can estimate the free - energy density correction due to the presence of gaussian subcritical bubbles . since , for a critical size bubble , @xmath128 , we can use eq . ( [ fthin ] ) to obtain the free energy needed to form a thin - wall critical bubble in a background of subcritical bubbles , @xmath129 for a very strong first - order phase transition , the subcritical bubbles are suppressed @xmath130 , and both @xmath131 and @xmath132 approach the homogeneous background expression . however , in the presence of subcritical bubbles , extra free energy becomes available in the medium , reducing the nucleation barrier . in other words , the extra background energy enhances the nucleation of critical bubbles . to illustrate this , we have plotted @xmath133 and @xmath86 as a function of @xmath134 in figs . 3 to 5 with @xmath31 values of 50 mev/@xmath50 , 30 mev/@xmath50 and 10 mev/@xmath50 , respectively , which are widely used in the literature . as evident , with decreasing temperature , the nucleation barrier decreases and the subcritical hadron fraction @xmath86 increases . the reduction in barrier height due to @xmath125 ( or due to @xmath86 ) is more significant for lower values of @xmath31 , corresponding to a weaker transition . since the height of the nucletaion barrier decreases , the nucleation rate will also be enhanced , reducing the amount of supercooling further . the time evolution of the temperature and the supercooling are discussed in the next section . as mentioned before , the background metastable state is inhomogeneous due to subcritical hadron bubbles . it is now possible to study the kinetics of the nucleation of critical hadron bubbles using the corrected nucleation rate , as obtained in the previous section . in the present work , the prefactor in the nucleation rate is taken as @xmath135 [ see eq . ( [ rate ] ) ] . in our previous work , @xcite , we have used a prefactor derived by csernai and kapusta @xcite for a dissipative qgp . in ref . @xcite , ruggeri and friedman had derived a prefactor for a non - dissipative qgp . recently , using a more general formalism , we have also derived a prefactor @xcite which has both dissipative and non - dissipative components corresponding to ref . @xcite and ref . @xcite , respectively . however , for consistency with the subcritical formalism , we use a more generic form @xmath136 , with @xmath88 a constant of order unity , as used in many studies of quark - hadron phase transition ( see , for example , refs . the question of how to estimate the prefactor appearing in the nucleation rate of subcritical bubbles remais open . using the nucleation rate @xmath137 , the fraction @xmath138 of space which has been converted to hadron phase due to nucleation of critical bubbles and their growth can be calculated . if the system cools to @xmath0 at a proper time @xmath139 , then at some later time @xmath140 the fraction @xmath138 is given by @xcite , @xmath141 v(\tau',\tau).\end{aligned}\ ] ] here , @xmath142 is the volume of a critical bubble at time @xmath140 which was nucleated at an earlier time @xmath143 ; this takes into account the bubble growth . the factor @xmath144 $ ] accounts for the available space for new bubbles to nucleate . the model for bubble growth is simply taken as @xcite @xmath145 where @xmath146^{3/2}$ ] is the velocity of the bubble growth at temperature @xmath7 @xcite . the evolution of the energy density in 1 + 1 dimensions is given by @xmath147 the energy density @xmath148 , enthalpy density @xmath149 and the pressure @xmath150 in pure qgp and hadron phases are given by the bag model equation of state . in the transition region , the @xmath148 and @xmath149 at a time @xmath140 can be written in terms of the hadronic fraction as @xmath151 h(\tau ) , \nonumber \\ \omega(\tau ) = \omega_q(t ) + [ \omega_h(t)-\omega_q(t ) ] h(\tau).\end{aligned}\ ] ] equations ( [ frac ] ) , ( [ hydro ] ) , and ( [ mixed ] ) are solved to get the temperature as a function of time in the mixed phase @xcite with the initial conditions for temperature @xmath152 mev and proper time @xmath153 fm / c at @xmath0=160 mev . after getting @xmath7 and @xmath138 as a function of time , the density of nucleating bubbles at a time @xmath140 can be obtained in our model as @xmath154.\end{aligned}\ ] ] the density @xmath155 would increase as the temperature drops below @xmath156 and would ultimately saturate as @xmath138 increases . figure 6 shows the temperature variation as a function of proper time @xmath140 at @xmath157 mev/@xmath50 . as the system cools below @xmath0 , the nucleation barrier decreases and also @xmath86 increases . if only homogeneous nucleation ( dashed curve ) is considered , the system will supercool up to 0.945 @xmath0 . at this temperature , the hadronic fraction @xmath86 has reached 10 % ( see fig . 3 ) , which corrects the amount of supercooling ( solid curve ) by about @xmath15810 % ( up to 0.95 @xmath0 ) . figure 7 shows a similar study at @xmath31 = 30 mev/@xmath50 . since the nucleation barrier reduces with decreasing @xmath31 , the system supercools only up to 0.98 @xmath0 under homogeneous nucleation . the hadronic fraction @xmath86 corresponding to this value is @xmath159 % ( see fig . 4 ) which reduces the amount of supercooling by about @xmath160 ( up to 0.984 @xmath0 ) . for @xmath31 around 10 mev/@xmath50 , the supercooling will be reduced further ( up to @xmath161 ) . lattice qcd calculations predict a surface tension even smaller than 10 mev/@xmath50 , indicating a very weak first order transition . although supercooling will be reduced further with decreasing @xmath31 , we do not use very small @xmath31 due to increased numerical inaccuracy . further , it may be mentioned here that , although the fraction @xmath86 grows with decreasing @xmath31 , we never encountered @xmath86 greater than 0.3 : we remained within the sub - percolation regime throughout our analysis . apart from @xmath31 , the amount of supercooling also depends on @xmath139 , the time taken by the system to cool from @xmath162 to @xmath0 . in qgp phase , the solution of eq . ( [ hydro ] ) @xmath163 = constant ) predicts @xmath164 . the choice of @xmath165 = 1fm , @xmath162=250 mev and @xmath0=160 mev results in @xmath139=3.8 fm / c . however , formation of qgp with higher initial temperature ( as high as 3 to 4 times @xmath0 resulting in large @xmath139 ) can not be ruled out at rhic and lhc energies @xcite . therefore , we have also studied the effect of @xmath139 on supercooling , specifically , on the hadronization rate as well as on the density of nucleating bubbles . 8(a ) and 8(b ) show the plots of @xmath166 and @xmath167 as a function of @xmath140 for two typical values of @xmath139 ( 3.8 fm / c and 25 fm / c ) corresponding to @xmath31=10 mev/@xmath168 both with ( solid curve ) and without ( dashed curve ) inhomogeneity corrections . a general observation ( both with and without correction ) is that the amount of supercooling , the rates of hadronization and bubble nucleation are reduced when @xmath139 becomes larger . although supercooling reduces with increasing @xmath139 , the system will get reheated at an earlier temperature and also will encounter a larger nucleation barrier as compared to the case when @xmath139 is small . as a result , the rate of hadronization and also the rate of increase of density of the nucleating hadron bubbles will proceed at a slower rate when @xmath139 is large . however , the reverse happens when the inhomogeneity correction is applied . even though the medium gets heated up earlier , the reduction in the barrier height is quite significant as @xmath7 approaches @xmath0 . another parameter that affects both @xmath166 and @xmath167 is the expansion rate of the medium , _ i.e. _ , the rate of change of temperature between @xmath139 and @xmath140 , which also depends on @xmath139 . the overall effect is that both @xmath166 and @xmath167 rise faster as compared to their homogeneous counterparts ( see fig . 8 for @xmath139=25 fm / c ) , particularly when @xmath139 is very large . ( compare the left and right curves on fig . 8 . ) the increase in rates of @xmath166 and @xmath167 is also larger for small @xmath31 at large @xmath139 . for weak enough transition , the presence of inhomogeneity may also affect several observables which can be detected experimentally . for example , the faster rate of hadronization at large @xmath139 as compared to its homogeneous counterpart will lower the amount of entropy production , which , in turn , will affect the final hadron multiplicity distributions . although not studied here , the bubble size distribution will also be affected by the dynamics of nucleation @xcite . since the nucleating bubble will act as a source of pion emission , the effect of inhomogeneity can also be inferred through interferometry measurements . in a cosmological context , the value of @xmath139 is much larger than what was quoted here . since the presence of inhomogeneities weakens the transition , more critical bubbles will be nucleated per unit volume , decreasing the inter - bubble distance , @xmath169 ; the presence of subcritical bubbles can be thought as seeds for nucleation . as a consequence , the transition will produce smaller fluctuations in baryon number , protecting homogeneous nucleosynthesis . although the present study is indicative enough of the reduction in the mean inter - bubble separation as compared to homogeneous nucleation , a quantitative estimate would require a more detailed analysis , including expansion . however , since the cosmological expansion rate is typically much slower than the subcritical bubble nucleation rate , we believe our results for the inter - bubble distance will carry on in this case as well . we have estimated the amount of phase mixing due to subcritical hadron bubbles from very weak to very strong first - order phase transitions . with a reasonable set of values for the surface tension and correlation length ( as obtained from lattice qcd calculations ) , we found that phase mixing is small at @xmath1 , building up as the temperature drops further . we have shown that the system does not mix beyond the percolation threshold , allowing us to describe the dynamics of the phase transition on the basis of homogeneous nucleation theory with a reduced nucleation barrier . accordingly , we have found an enhancement in the nucleation rate which further reduces the amount of supercooling . although we have not included cosmological expansion in our analysis , we believe that our results indicate that the presence of an inhomogeneous background of subcritical bubbles will decrease the inter - bubble mean distance , and thus the fluctuations in baryon number which could damage homogeneous nucleosynthesis . we have assumed that the equilibration time - scale for subcritical fluctuations is much larger than the cooling time - scale of the system . this may be the case for a quark - hadron phase transition in the early universe , where the expansion rate is quite slow . in the case of qgp produced at rhic and lhc , the cooling rate is much faster than cosmological time - scales , and the subcritical bubbles density distribution may not attain full equilibration . we are presently investigating this issue in more detail . however , the present results should provide an upper bound on the fraction of subcritical hadron bubbles and their effect on supercooling and nucleation rates . 3 the nucleation barrier @xmath133 for critical bubbles with ( solid line ) and without subcritical bubble correction ( dashed curve ) as function of temperature for @xmath157 mev/@xmath50 is shown in upper panel . corresponding subcritical hadron fraction @xmath86 is shown in the lower panel . fig . 8 ( a ) density of nucleating bubbbles as a function of proper time with ( solid curve ) and without subcritical bubble correction ( dashed curve ) for @xmath31= 10 mev/@xmath50 ( b ) the hadronic fraction @xmath86 as a function of @xmath140 . the left curves are for @xmath172 fm / c and the right curves for @xmath173 fm / c .	the effect of subcritical hadron bubbles on a first - order quark - hadron phase transition is studied . these subcritical hadron bubbles are created due to thermal fluctuations , and can introduce a finite amount of phase mixing ( quark phase mixed with hadron phase ) even at and above the critical temperature . for reasonable choices of surface tension and correlation length , as obtained from the lattice qcd calculations , we show that the amount of phase mixing at the critical temperature remains below the percolation threshold . thus , as the system cools below the critical temperature , the transition proceeds through the nucleation of critical - size hadron bubbles from a metastable quark - gluon phase ( qgp ) , within an inhomogeneous background populated by an equilibrium distribution of subcritical hadron bubbles . the inhomogeneity of the medium results in a substantial reduction of the nucleation barrier for critical bubbles . using the corrected nucleation barrier , we estimate the amount of supercooling for different parameters controlling the phase transition , and briefly discuss its implications to cosmology and heavy - ion collisions . + pacs number(s ) : 12.38.mh , 64.60.qb , 05.70.fh , 25.75-q , 98.80.cq
in @xcite , kellerhals wrote the most basic objects in polyhedral geometry are orthoschemes " , and she gave a formula to calculate the volumes of complete orthoschemes in the three - dimensional hyperbolic space . what we discuss here is the existence and the uniqueness of the maximal volume of a family of complete orthoschemes parametrized by the height " . consider a family of pyramids in euclidean space with a fixed base polygon and the locus of apexes perpendicular to the base polygon . the volumes of pyramids strictly increases when the height increases , because pyramids strictly increases as a set . by the same reason , this phenomenon holds true for such a family of pyramids in hyperbolic space . in contrast to the euclidean case , the volume approaches to a finite value . furthermore , in hyperbolic space the apex can run out " the space . then we can still obtain finite volume hyperbolic polyhedron by _ truncation _ with respect to the apex . the volume converges to zero as the vertex goes away from the space . so it is an interesting question when the volume becomes maximum . as is mentioned above , one of the most fundamental one among all such pyramids is the orthoscheme . an orthoscheme is a kind of simplex which has particular orthogonality among its faces . let @xmath0 , @xmath1 , @xmath2 and @xmath3 be the vertices of a simplex @xmath4 in the three - dimensional hyperbolic space . we denote by @xmath5 the edge spanned by @xmath6 and @xmath7 , and by @xmath8 the face spanned by @xmath6 , @xmath7 and @xmath9 . such a simplex @xmath4 is called an _ orthoscheme _ ( in the ordinary sense ) if the edge @xmath10 is perpendicular to the face @xmath11 and the face @xmath12 is orthogonal to @xmath13 . in other words , an orthoscheme is a tetrahedron with a right - angled triangle @xmath12 as its base and an edge joining the apex and a non - right - angled vertex , say @xmath2 , is perpendicular to the base . vertices @xmath0 and @xmath3 are called the _ principal vertices _ of @xmath4 . its precise definition will be given in section [ sec : orthoschemes ] . though orthoschemes are also considered in euclidean or spherical spaces , in hyperbolic space the ordinary orthoschemes are extended to the so - called _ complete orthoschemes_. let @xmath14 be the open unit ball in the three - dimensional euclidean space @xmath15 centered at the origin . the set @xmath14 can be regarded as the so - called _ projective ball model _ of the three - dimensional hyperbolic space . any tetrahedron in hyperbolic space appears as a euclidean tetrahedron in @xmath14 . if one or both principal vertices of an orthoscheme @xmath4 lie in the boundary of @xmath14 , the set @xmath16 is called an _ ideal polyhedron _ , which is not bounded in hyperbolic space , while its volume is finite . take one step further and we allow principal vertices to be in the exterior of @xmath14 . the volume of @xmath16 is no longer finite , but there is a canonical way to delete ends of @xmath16 with infinite volume so that we obtain a polyhedron of finite volume , called a _ truncated polyhedron_. complete orthoschemes are , roughly speaking , either compact , ideal or truncated orthoschemes . the precise definitions of complete orthoschemes and truncation will also be given in section [ sec : orthoschemes ] . what we study in this paper is the maximal volume of a family of complete orthoschemes with one parameter . consider a family of complete orthoschemes that share the same base @xmath12 . we allow the vertex @xmath0 to be in the exterior of @xmath14 . in this case the base @xmath12 means the truncated polygon obtained from the triangle with vertices @xmath0 , @xmath1 and @xmath2 . such a family of complete orthoschemes is parametrized by the hyperbolic length of the edge @xmath13 when @xmath3 is in @xmath14 . when the hyperbolic length increases , the orthoscheme strictly increases as a set , which means the volume also increases with respect to the function of the hyperbolic length . this phenomenon holds until the vertex @xmath3 lies in the boundary @xmath17 of @xmath14 . the hyperbolic length of @xmath13 is beyond " the infinity when @xmath3 is in the exterior of @xmath14 , but we have a complete orthoscheme with finite volume by truncation . instead of the hyperbolic length , using the euclidean length of @xmath13 , which we mentioned as height " in the first paragraph , we can parametrize the family even if @xmath3 is in the complement of @xmath14 . the complete orthoscheme approaches the empty set when @xmath3 goes far away from @xmath14 . the family thus has maximal volume complete orthoschemes , which arise when @xmath3 lies in the complement of @xmath14 . as a toy model , let us consider the same phenomenon for the two - dimensional orthoschemes , namely hyperbolic triangle @xmath12 with right angle at @xmath1 . take a family of complete orthoschemes parametrized by the height " of @xmath18 . the area strictly increases when @xmath2 approaches to the boundary of @xmath19 , the projective disc model of the two - dimensional hyperbolic space . the area attains maximal when @xmath2 lies in @xmath20 . when @xmath2 is in the exterior of @xmath19 , the area decreases , but not necessarily monotonically . these facts are summarized as theorem [ thm : 2ortho ] in the appendix . one may expect that the same phenomenon happens for three - dimensional complete orthoschemes . is the volume attains maximal at least when @xmath3 is in @xmath17 ? does the volume decrease when @xmath3 goes far away from @xmath14 ? our main result , which is theorem [ thm : main ] in section [ sec : mainresult ] , answers both of the questions negatively . there are several models to introduce hyperbolic geometry . among them we use the _ hyperboloid model _ to calculate lengths and angles with respect to the hyperbolic metric , and use the _ projective ball model _ to define complete orthoschemes . definitions of these two models , together with formulae to calculate hyperbolic lengths and hyperbolic angles , are explained in this section . see @xcite for basic references on hyperbolic geometry . as a set , the hyperboloid model @xmath21 of the three - dimensional hyperbolic space is defined as a subset of the four - dimensional euclidean space @xmath22 by @xmath23 where @xmath24 , called the_lorentzian inner product _ , is defined as @xmath25 for any @xmath26 and @xmath27 in @xmath22 . the restriction of the quadratic form induced from the lorentzian inner product to the tangent spaces of @xmath21 is positive definite and gives a riemannian metric on @xmath21 , which is constant curvature of @xmath28 . the set @xmath21 together with this metric gives the _ hyperboloid model _ of the three - dimensional hyperbolic space . associated with @xmath21 , there are two important subsets of @xmath22 : @xmath29 every point @xmath30 in @xmath31 corresponds to a half - space @xmath32 bounded by a plane @xmath33 the intersection @xmath34 is a geodesic plane with respect to the hyperbolic metric . if @xmath30 is taken from @xmath35 , the set @xmath36 is defined as @xmath37 the intersection @xmath38 is called a _ horoball_. the intersection of the boundary @xmath39 of @xmath36 and @xmath21 is called a _ the lorentzian inner product is also used to calculate distances and angles with respect to the hyperbolic metric . the details of the following results are explained in 3.2 of @xcite . let @xmath30 be a point in @xmath21 and let @xmath40 be taken from @xmath31 with @xmath41 , then the hyperbolic distance @xmath42 between @xmath30 and the geodesic plane @xmath43 is calculated by @xmath44 suppose that @xmath30 is in @xmath35 and @xmath40 is in @xmath31 with @xmath41 . let @xmath42 be the signed hyperbolic distance between the horosphere @xmath34 and the geodesic plane @xmath45 . the sign is defined to be positive if the horosphere and the geodesic plane do not intersect , otherwise negative . then the signed distance @xmath42 is calculated by @xmath46 if both @xmath30 and @xmath40 are taken from @xmath31 with @xmath41 and @xmath47 , then there are three possibilities : @xmath48 intersects @xmath21 , intersects @xmath35 or does not intersect both @xmath21 and @xmath35 . the first case means that the geodesic planes @xmath34 and @xmath49 intersect and form a corner @xmath50 . the hyperbolic dihedral angle @xmath51 between these geodesic planes measured in this corner is calculated by @xmath52 the third case means that the geodesic planes @xmath34 and @xmath49 are _ ultraparallel _ , meaning that they do not intersect in @xmath21 and there is a unique geodesic line in @xmath21 which is perpendicular to these geodesic planes . the hyperbolic length @xmath42 of the segment between these planes is calculated by @xmath53 the second case is regarded as the first case with hyperbolic dihedral angle @xmath54 or the third case with the hyperbolic distance @xmath54 . geodesic planes in this case are called _ parallel _ in hyperbolic space . the _ projective ball model _ @xmath14 is another model of the three - dimensional hyperbolic space , which is induced from @xmath21 . let @xmath55 be the radial projection from @xmath56 to the affine hyperplane @xmath57 along the ray from the origin @xmath58 of @xmath22 . the projection @xmath55 is a homeomorphism on @xmath21 to the three - dimensional open unit ball @xmath14 in @xmath59 centered at @xmath60 . a metric is induced on @xmath14 from @xmath21 by the projection . with this metric @xmath14 is called the _ projective ball model _ of the three - dimensional hyperbolic space . the projection @xmath55 also induces the mapping from @xmath61 to the three - dimensional real projective space @xmath62 , which is defined to be the union of @xmath59 and the set of lines in the affine hyperplane @xmath63 through @xmath58 . in contrast to ordinary points in @xmath14 , points in the set @xmath17 of the boundary of @xmath14 are called _ ideal _ , and points in the exterior of @xmath14 are called _ ultraideal_. we often regard @xmath14 as the unit open ball centered at the origin in @xmath15 . we mention important properties of @xmath14 to be used in the definition of complete orthoschemes in the next section . first , every geodesic plane in @xmath14 is given as the intersection of a euclidean plane and @xmath14 . this is because every geodesic plane in @xmath21 is defined as the intersection of @xmath21 and a linear subspace of @xmath22 of dimension three , and a geodesic plane in @xmath14 is the image of that in @xmath21 by the radial projection . the projection @xmath55 thus gives a correspondence between points in the exterior of @xmath14 in @xmath62 and the geodesic planes of @xmath14 . we call @xmath64 for @xmath65 the _ pole _ of the plane @xmath66 or the geodesic plane @xmath67 . conversely , we call @xmath66 the _ polar plane _ of @xmath64 , and we call @xmath67 the _ polar geodesic planes _ of @xmath64 . if @xmath67 does not pass through the origin of @xmath14 , then its pole is given as the apex of a circular cone which is tangent to @xmath17 and has the base circle @xmath68 . the second important property is that , for a given geodesic plane , say @xmath69 , in @xmath14 , every plane or line which passes through the pole of @xmath69 is orthogonal to @xmath69 in @xmath14 . this is proved by using equation . following @xcite we introduce complete orthoschemes . as is mentioned in the introduction , an ( ordinary ) _ orthoscheme _ in the three - dimensional hyperbolic space is a tetrahedron with vertices @xmath0 , @xmath1 , @xmath2 and @xmath3 which satisfies that @xmath10 is perpendicular to @xmath11 and that @xmath12 is orthogonal to @xmath13 . the vertices @xmath0 and @xmath3 are called _ principal vertices_. _ complete orthoschemes _ are a generalization of ordinary orthoschemes by allowing one or both principal vertices to be ideal or ultraideal . take @xmath14 as our favorite model of the hyperbolic space in what follows . as a set , any orthoscheme in the ordinary sense are given as a euclidean tetrahedron in @xmath14 . when one or both principal vertices are ideal , the tetrahedron as a set in the hyperbolic space is no more bounded , but still has finite volume . we allow to call such tetrahedra ordinary orthoschemes . further generalization of orthoschemes is explained via _ truncation _ of ultraideal vertices . suppose a vertex @xmath70 of a tetrahedron @xmath4 is ultraideal . let @xmath71 be the half - space bounded by the polar plane of @xmath70 with @xmath72 . _ truncation _ of @xmath4 with respect to @xmath70 is defined as an operation to obtain a polyhedron @xmath73 . if @xmath70 is close enough to @xmath17 , then @xmath73 is non - empty . truncation is also explained by using the hyperboloid model . the inverse image of @xmath70 for @xmath55 on @xmath31 consists of two points . each of them gives a half - space in @xmath22 , and one of them corresponds to the inverse image of @xmath71 . in this sense there is a one - to - one correspondence between half - spaces in @xmath14 and points in @xmath31 . the point in @xmath31 corresponding to @xmath70 with respect to @xmath71 in the sense above is called the _ proper _ inverse image of @xmath70 for truncation of @xmath4 . this correspondence will be used to calculate hyperbolic lengths of edges and hyperbolic dihedral angles between faces of complete orthoschemes . when one of the principal vertices , say @xmath3 , is ultraideal and @xmath0 is not ( i.e , ordinal or ideal ) , we have a polyhedron with finite volume by truncation with respect to @xmath3 . such a polyhedron is called a _ simple frustum _ with ultraideal vertex @xmath3 . we remark that the vertices @xmath0 , @xmath1 and @xmath2 are simultaneously deleted by truncation when @xmath3 is far away from @xmath14 , since both the polar geodesic plane of @xmath3 and the triangle @xmath12 are orthogonal to @xmath13 in @xmath14 . suppose both @xmath0 and @xmath3 are ultraideal . there are three possibilities : the polar planes of @xmath0 and @xmath3 intersect in @xmath14 , they are parallel , or they are ultraparallel . in the first case , the polyhedron we obtain by truncation is well known as a _ lambert cube_. see ( * ? ? ? * figure 2 ) for example . the edge @xmath74 is deleted by truncation . in the third case , on the other hand , the polyhedron obtained by truncation still has the edge induced from @xmath74 . we call this polyhedron a _ double frustum_. the second case is the limiting situation of both the first and the third cases . we call polyhedra obtained in the second case _ double frustum with an ideal vertex_. as a summary , combinatorial types of complete orthoschemes are either * ordinary orthoschemes , whose principal vertices are either ordinarily points or ideal points , * simple frustums , * double frustums possibly with an ideal vertex , or * lambert cubes . kellerhals obtained formulae to calculate volumes of complete orthoschemes in @xcite ; the formula for lambert cubes is given in theorem iii , and the formula for other kinds of complete orthoschemes is given in theorem ii . in both formulae , they are parametrized by the three non - right hyperbolic dihedral angles . under the same setting used in section [ sec : orthoschemes ] , we denote by @xmath75 the hyperbolic dihedral angle between faces opposite to @xmath6 and @xmath7 . when a complete orthoscheme is a lambert cube , the geodesic planes containing faces opposite to @xmath1 and @xmath2 are ultraparallel . in this case @xmath76 is defined to be the hyperbolic dihedral angle between the polar geodesic planes of the vertex @xmath0 and @xmath3 . in this sense the formulae are parametrized by @xmath77 , @xmath76 and @xmath78 . kellerhals used the _ schlfli differential formula _ to obtain these volume formulae . the volume formulae are not used directly in our arguments ; what we will use is the fact that the formulae are parametrized by the three non - right hyperbolic dihedral angles . on the other hand , the schlfli differential formula itself plays an important role in our arguments . the schlfli differential formula gives an expression of the differential form of the volume function with respect to the hyperbolic lengths of edges and hyperbolic dihedral angles between faces . as is given in theorem i in @xcite , the differential form @xmath79 of the volume function @xmath80 of any complete orthoschemes is expressed as @xmath81 where @xmath82 is the hyperbolic length of the edge @xmath5 if both @xmath6 and @xmath7 are points in @xmath14 , @xmath82 is the hyperbolic distance between @xmath6 and the polar geodesic plane of @xmath7 if @xmath6 is a point in @xmath14 and @xmath7 lies in the exterior of @xmath14 , and @xmath82 is the hyperbolic distance between the polar geodesic planes of @xmath6 and @xmath7 if both @xmath6 and @xmath7 lie in the exterior of @xmath14 . if a complete orthoscheme is a lambert cube , then @xmath83 is taken as the hyperbolic length of the edge obtained as the intersection of the polar geodesic planes of @xmath0 and @xmath3 . if one of @xmath0 and @xmath3 is ideal , then the edges with the ideal vertex as an endpoint have infinite hyperbolic lengths . in this case we take any horosphere centered at the ideal vertex , and each infinite length is replaced by the signed hyperbolic distance between the other endpoint and the horosphere . as is mentioned in the concluding remarks in @xcite , the schlfli differential formula is still valid by this treatment . as a result , the schlfli differential formula is applicable to any kind of complete orthoschemes . we use the formula as the equation @xmath84 for @xmath85 . this equation plays a key role in the proof of theorem [ thm : main ] . suppose @xmath14 lies in the @xmath86-coordinate space of @xmath15 . by the action of an isometry , any ordinary orthoscheme can be put as the vertex @xmath0 is in the positive quadrant of the @xmath87-plane , the vertex @xmath1 is on the positive part of the @xmath88-axis , the vertex @xmath2 is the origin , and the vertex @xmath3 is on the positive part of the @xmath89-axis . such orthoschemes are parametrized by @xmath90 , where @xmath91 is the @xmath89-coordinate of @xmath3 , i.e. , the euclidean distance between @xmath2 and @xmath3 , @xmath92 is the euclidean distance between @xmath0 and @xmath2 , and @xmath51 is the euclidean angle between edges @xmath93 and @xmath18 . when we regard such an orthoscheme as a tetrahedron with base @xmath12 , the @xmath89-coordinate @xmath91 of @xmath3 is the height " of the tetrahedron . what we study in this paper is a family of complete orthoschemes parametrized by the height " . for fixed @xmath92 and @xmath51 , we have a one - parameter family @xmath94 of ordinary orthoschemes parametrized by @xmath91 . this family is extended even when @xmath95 and/or @xmath96 with @xmath97 , if we mean @xmath98 a complete orthoscheme . let @xmath99 be the hyperbolic volume of @xmath98 . by the volume formulae , the function @xmath100 is continuous on @xmath101 and piecewise differentiable on the intervals each of which corresponds to a combinatorial type of complete orthoschemes given at the end of section [ sec : orthoschemes ] . when @xmath91 increases in value approaching @xmath102 , the orthoscheme also increases as a set . so @xmath99 strictly increases in value approaching @xmath103 as @xmath91 approaches @xmath102 from below . when @xmath91 approaches positive infinity @xmath104 , the sequence @xmath98 of complete orthoschemes converges to the base @xmath12 ; the complete orthoschemes are always ordinary ones when @xmath105 , and the complete orthoschemes changes into lambert cubes from double frustums when @xmath106 . in any case @xmath99 converges to @xmath54 as @xmath91 approaches @xmath104 . based on these observations , we have set the following questions . for a given one - parameter family @xmath107 of complete orthoschemes , does the function @xmath100 attain maximal when @xmath3 is in @xmath17 ? is @xmath100 strictly decreasing on @xmath108 ? the next theorem , which is the main result of this paper , answers both of the questions negatively . [ thm : main ] for any @xmath109 and @xmath110 with @xmath97 , the volume @xmath99 of @xmath98 attains maximal for some @xmath111 . furthermore , the maximal volume is unique for any @xmath92 and @xmath51 , and it is given before @xmath98 becomes a lambert cube . the outline of the proof is as follows . using the schlfli differential formula , we can calculate @xmath112 for each combinatorial types of @xmath98 . since @xmath100 is a strictly increasing function on @xmath113 $ ] , proving @xmath114 tells us that the function @xmath100 attains maximal for some @xmath111 . the uniqueness of such @xmath91 is induced from the uniqueness of the solution of the equation @xmath115 on @xmath108 . our proof of theorem [ thm : main ] is organized as follows . after confirming the correspondence between combinatorial types of complete orthoschemes and conditions of parameters @xmath91 , @xmath92 and @xmath51 , we first obtain suitable inverse images of vertices of @xmath98 for @xmath55 . these are used to calculate hyperbolic lengths and hyperbolic dihedral angles appearing in the schlfli differential formula . under each of conditions of parameters , we prove that the volume function @xmath100 with respect to @xmath91 attains maximal on @xmath116 , and that such @xmath91 is unique . for @xmath106 , we also prove that @xmath100 does not attain maximal if @xmath98 is a lambert cube . by the definition of @xmath98 , the coordinates of the vertices are @xmath117 where @xmath118 . as is mentioned after theorem [ thm : main ] , it is enough to assume that @xmath119 in what follows . a complete orthoscheme @xmath98 is a simple frustum if @xmath120 , and a simple frustum with ideal vertex @xmath0 if @xmath121 . when @xmath106 , we always assume @xmath97 so that @xmath1 is in @xmath14 . under these assumptions , a complete orthoscheme @xmath98 with @xmath122 is either a double frustum , a double frustum with an ideal vertex , or a lambert cube . these are distinguished via the euclidean distance between the origin of @xmath15 and the edge @xmath74 ; @xmath98 is a double frustum , a double frustum with an ideal vertex , or a lambert cube if and only if the euclidean distance is less than , equal to , or greater than @xmath102 respectively . since the euclidean distance is @xmath123 , we have that these are equivalent to @xmath124 , @xmath125 , or @xmath126 respectively . the inequality @xmath127 is also equivalent to @xmath128 without the assumption that @xmath106 . we note that this inequality always holds for any @xmath129 and @xmath105 . as a summary , complete orthoschemes @xmath98 are parametrized by @xmath130 , and with @xmath119 and @xmath118 , and * when @xmath105 , complete orthoschemes @xmath98 are simple frustums with @xmath128 , * when @xmath106 with @xmath97 and @xmath131 , complete orthoschemes @xmath98 are double frustums ( possibly with an ideal vertex ) , and * when @xmath106 with @xmath97 and @xmath126 , complete orthoschemes @xmath98 are lambert cubes . we next give the proper inverse images of these vertices for @xmath55 . when a vertex is in @xmath14 , its inverse image for @xmath55 must be chosen in @xmath21 , which is uniquely determined . when a vertex is in the exterior of @xmath14 , its inverse image is chosen to be proper inverse image in the sense of truncation . finally , when a vertex is in @xmath17 , we choose its proper inverse image as any element in the inverse for @xmath55 , which is a subset in @xmath35 . let @xmath132 be the proper inverse image of @xmath6 in this sense . the coordinates of @xmath132 are then as follows : 1 . [ enu : case r<1 ] when @xmath120 , we have @xmath133 2 . [ enu : case r=1 ] when @xmath134 , the coordinates of @xmath135 , @xmath136 and @xmath137 are the same as in the first case and @xmath138 3 . [ enu : case r>1 ] when @xmath106 , the coordinates of @xmath135 , @xmath136 and @xmath137 are the same as in the first case , and @xmath139 the inverse image of the pole of a geodesic plane in @xmath14 consists of two points in @xmath31 . for each ( ordinary ) face of an orthoscheme @xmath98 , we choose the inverse image of the pole in @xmath31 so that the half - space defined by this inverse image contains @xmath98 . let @xmath140 be the inverse image of the pole of the face @xmath141 for @xmath142 in this sense . in other words , @xmath140 is a point in @xmath31 where @xmath143 contains @xmath98 and @xmath144 contains @xmath141 . for any @xmath92 , the coordinates of @xmath140 are as follows : @xmath145 we focus on the derivative @xmath112 to prove that @xmath100 attains maximal on @xmath146 , as well as its uniqueness . we first confirm that the function @xmath100 is piecewise differentiable with respect to @xmath91 in general . we first suppose that @xmath147 . by the schlfli differential formula , the function @xmath100 is differentiable with respect to the hyperbolic dihedral angles @xmath77 , @xmath76 and @xmath78 . by the expression of the coordinates of @xmath140 and @xmath132 for @xmath148 together with equation , these angles are given as smooth functions with respect to @xmath91 . by the chain rule , @xmath100 is thus differentiable with respect to @xmath91 . in particular @xmath100 is continuous on @xmath149 . if @xmath106 , then there are two combinatorial types of @xmath98 ; a double frustum or a lambert cube . the function @xmath100 is not only continuous but also piecewise differentiable on @xmath101 , for @xmath100 is differentiable on the intervals corresponding to each combinatorial types of @xmath98 by the same argument used for @xmath147 . recall that the function @xmath100 is continuous on @xmath149 , strictly increasing on @xmath113 $ ] and has its limit @xmath54 as @xmath91 approaches @xmath150 . so , to prove that @xmath100 attains maximal on @xmath116 , it is enough to prove that the limit of @xmath112 is positive as @xmath91 approaches to @xmath102 from above . the uniqueness of the maximal value of @xmath100 is induced from the fact that the solution of @xmath115 is at most one on @xmath116 . applying the chain rule and we have @xmath151 the parameter @xmath75 defined in section [ sec : schlafli ] is the hyperbolic dihedral angle between the polar geodesic planes of @xmath152 and @xmath153 . in other words , @xmath75 is the hyperbolic dihedral angle along the edge @xmath154 for @xmath142 . as is mentioned in the first paragraph of section [ sec : schlafli ] , if @xmath98 is a lambert cube , then @xmath76 is taken as the hyperbolic dihedral angle between the polar geodesic planes of @xmath0 and @xmath3 . the schlfli differential formula are used to calculate partial derivatives appeared in the equation above . by equation we have @xmath155 where @xmath82 is the hyperbolic length with respect to the edge @xmath5 defined in section [ sec : schlafli ] . furthermore , the hyperbolic dihedral angle @xmath77 , which coincides with the euclidean angle @xmath51 by the definition of @xmath98 , is constant with respect to @xmath91 , meaning that @xmath156 . we thus have @xmath157 we divide the remaining argument into three cases according to the value of @xmath92 . by equations we have @xmath158 and by equation we have @xmath159 derivatives of hyperbolic dihedral angles with respect to @xmath91 are obtained as follows : @xmath160 substitute equations , and to equation and we have @xmath161 where @xmath162 and @xmath163 . since @xmath164 we have @xmath165 which implies that @xmath100 attains maximal for some @xmath111 . this result together with @xmath166 implies that the uniqueness of the maximal value of the function @xmath100 with respect to @xmath91 is proved by showing that the equation @xmath115 has at most one solution on @xmath116 . since @xmath167 on @xmath116 by equation , we have @xmath168 since @xmath169 where @xmath170 , is negative on @xmath171 , the function @xmath172 is strictly monotonic with respect to @xmath91 . this implies that the number of elements in the right - hand side set of equation is at most one , so is the left - hand side . using equation , we have @xmath173 by equations and with @xmath134 and we have @xmath174 substitute these equations to equation and we have @xmath175 since @xmath176 we have @xmath177 in this case . the uniqueness of the maximal value of @xmath100 with respect to @xmath91 is obtained by the facts that @xmath178 is a strictly monotonic function and that @xmath167 on @xmath171 . since our strategy of proving that @xmath100 attains maximal on @xmath116 is to prove that the limit of @xmath112 is positive as @xmath91 approaches to @xmath102 from above , it is enough to consider the case that @xmath91 is close enough to @xmath102 , meaning that @xmath98 are double frustum , not lambert cubes . under this assumption , use equation and we have @xmath179 substitute this equation together with equations and to equation and we have @xmath180 where @xmath181 is the function defined in case . by the same reason explained in case , we have @xmath177 in this case as well . we next prove that @xmath100 does not attain maximal when @xmath98 is a lambert cube , i.e. , @xmath182 . what we actually prove is that @xmath100 is strictly decreasing , using equation . recall that @xmath83 is the hyperbolic distance between the polar geodesic plane of @xmath183 and @xmath184 , and @xmath76 is the hyperbolic dihedral angle between the polar geodesic planes of @xmath0 and @xmath3 , while @xmath185 and @xmath78 are the same as in other cases . using equations and , we have @xmath186 the value @xmath187 is positive on @xmath188 by this expression , so is @xmath189 by equation . the value @xmath185 is positive , for it is the hyperbolic length of an edge . by substituting these results to equation , if we can prove that @xmath190 , then we have @xmath191 , namely @xmath100 is strictly decreasing , on @xmath188 . the inequality @xmath190 is equivalent to @xmath192 calculating @xmath193 and we have an inequality @xmath194 which is equivalent to the previous one . this inequality holds on @xmath188 , for the right - hand side is negative while the left hand side is positive . we have thus proved that @xmath100 does not attain maximal when @xmath98 is a lambert cube . since @xmath100 does not attain maximal when @xmath98 is a lambert cube , for the proof of the uniqueness of the maximal value of @xmath100 , we can assume that @xmath195 $ ] . under this assumption together with the fact that @xmath167 on @xmath196 by equation , what we need to prove is that the number of elements in the set @xmath197 is at most one , where @xmath112 is calculated in equation and the function @xmath181 is given in equation . by equation , we have @xmath198 where we recall that @xmath199 . unlike case , the sign of the function @xmath200 is not expected to be constant on @xmath196 , for @xmath201 . since @xmath202 on @xmath196 , we have @xmath203 the function @xmath200 is quadratic with respect to @xmath91 , the coefficient of @xmath204 is negative and @xmath205 . these imply that the number of elements in the set of the right - hand side of the equation above is at most one , so is the set of the left - hand side of the equation . suppose that the number of elements in the set @xmath206 is more than @xmath102 . by the mean - value theorem together with the fact that the limit of @xmath207 is @xmath54 as @xmath91 approaches @xmath208 from below , the set @xmath209 must contain at least two elements , which contradicts the result obtained above . we have thus proved theorem [ thm : main ] . by the definition of orthoscheme , a triangle @xmath12 in the two - dimensional hyperbolic space is orthoscheme if the edge @xmath10 is perpendicular to the edge @xmath18 , namely @xmath12 is a right - angled triangle with the right angle at @xmath1 . without loss of generality , we suppose that @xmath12 lies in the projective disc model @xmath19 with the coordinates @xmath211 for a given @xmath212 , we consider a family @xmath213 of complete orthoschemes , where @xmath214 is a complete orthoscheme with vertices @xmath0 , @xmath1 and @xmath2 . what we discuss is the maximal area for this family . [ thm : 2ortho ] the maximal area for @xmath215 is obtained as follows : 1 . for any @xmath216 , the area of @xmath214 attains maximal just for @xmath217 . the maximal area is @xmath218 , where @xmath219 is the hyperbolic angle at @xmath220 of @xmath221 . 2 . the area of @xmath222 attains maximal for any @xmath223 . the maximal area is @xmath224 . 3 . for any @xmath106 , the area of @xmath214 attains maximal for any @xmath225 $ ] . the maximal area is @xmath224 . we start by recalling a formula to calculate the area @xmath226 of a hyperbolic convex @xmath227-gon with hyperbolic angles @xmath228 ; @xmath229 see theorem 3.5.5 of @xcite for the proof when @xmath230 . let @xmath231 be the area of @xmath214 . for any @xmath212 , a complete orthoscheme @xmath214 increases as a set when @xmath91 approaches @xmath102 from below , which implies that @xmath231 also increases . so , to prove the theorem , it is enough to assume that @xmath95 . using this formula , we obtain the area of @xmath214 for each case . 1 . suppose @xmath232 . let @xmath233 be the hyperbolic angle at @xmath0 of @xmath214 . + when @xmath217 , @xmath221 is a triangle with ideal vertex @xmath2 . since the hyperbolic angle at @xmath2 is @xmath54 , the area is @xmath234 + when @xmath119 , @xmath214 is a quadrilateral . the hyperbolic angles at the vertices constructed by truncation with respect to @xmath2 are right angles . the area is @xmath235 when @xmath91 approaches @xmath150 , the corner at @xmath0 increases as a set , so is the angle @xmath233 . this implies that @xmath231 is a strictly decrease function on @xmath236 . + as a result , @xmath231 attains maximal if and only if @xmath237 in this case . 2 . suppose @xmath134 . the hyperbolic angle at @xmath0 is @xmath54 in this case . use the argument in ( 1 ) with @xmath238 for any @xmath95 and we have the desired conclusion . 3 . suppose @xmath106 . + when @xmath237 , @xmath221 is a quadrilateral with angle @xmath54 at @xmath2 and three right angles . the area is @xmath239 + when @xmath240 , there are two kinds of @xmath214 , which correspond to double frustums and lambert cubes of three - dimensional complete orthoschemes . * if @xmath124 , then @xmath214 is a right - angled pentagon . the area is @xmath241 * if @xmath242 , then @xmath214 is a quadrilateral , whose edges consists of @xmath10 , @xmath18 and polar lines of @xmath0 and @xmath2 . let @xmath243 be the hyperbolic angle between these polar lines . then the area is @xmath244 the maximal area arrises when @xmath245 , which occurs if and only if the polar planes of @xmath0 and @xmath2 are parallel , namely @xmath125 . summarizing these results , we have completed the proof . the authors would like to thank the referee for his / her careful reading and useful suggestions .	a three - dimensional orthoscheme is defined as a tetrahedron whose base is a right - angled triangle and an edge joining the apex and a non - right - angled vertex is perpendicular to the base . a generalization , called complete orthoschemes , of orthoschemes is known in hyperbolic geometry . roughly speaking , complete orthoschemes consist of three kinds of polyhedra ; either compact , ideal or truncated . we consider a particular family of hyperbolic complete orthoschemes , which share the same base . they are parametrized by the height " , which represents how far the apex is from the base . we prove that the volume attains maximal when the apex is ultraideal in the sense of hyperbolic geometry , and that such a complete orthoscheme is unique in the family .

Subsets and Splits

Data Mining Abstracts

This query searches for and retrieves 2000 documents containing the phrase 'data mining' in the abstract, providing limited but specific document samples.